article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
mergers between clusters are exciting events , involving around @xmath10 ergs of kinetic energy . on cluster - wide scales the effects of the dissipation of this energy are evident in the intracluster medium ( icm ) .
x - ray observations of the icm in presumed cluster mergers often reveal non - spherical morphologies , sometimes with temperature variations and shocks . similarly , the presence of diffuse radio emission in the form of relics and halos seems to correlate with known cluster mergers . in such a dynamic environment , what is the effect on individual galaxies ?
any answers are likely complicated by numerous , and often competing , effects .
galaxies in groups which are being accreted by clusters may undergo bursts of star formation as time - varying tidal fields transfer gas to the galaxies centers @xcite .
similar `` pre - processing '' of galaxies is found in simulations where galaxy - galaxy interactions and tidal perturbations are enhanced in groups and substructures around clusters @xcite . on a larger scale ,
when clusters merge with each other they develop shocks in their intracluster gas and interactions between individual galaxies and these shocks may lead to starbursts @xcite . in the case of individual galaxies already within clusters , the successive influence of numerous high - speed passages may lead to starbursts ( harrassment ,
e.g. * ? ? ?
gas - rich galaxies entering a cluster would experience ram pressure through interaction with the intracluster gas @xcite , possibly compressing their molecular clouds and initiating bursts of star formation @xcite .
however , ram pressure is also a prime candidate for suppressing star formation in clusters , as it can efficiently remove the gas supply of the entering galaxies , thereby extinguishing any star formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . as a result ,
the issue of whether or not a starburst ever occurs is an open one , with some authors finding evidence for such starburst histories ( e.g. , * ? ? ?
* ) with others favoring `` strangulation '' of star formation ( e.g. , * ? ? ?
lastly , one should bear in mind that all of these arguments may be moot : the member galaxies of any single cluster may already be quiescent and gas poor , leaving little material available for transformation in a cluster merger .
recently , radio investigations have been used to study galaxy activity in cluster mergers .
radio emission is an excellent indicator of galaxy activity , as it can result from active galactic nuclei ( agn ) or star formation ( see * ? ? ? * for a review ) .
in fact , in the absence of any contamination due to an agn the radio luminosity of a galaxy ( typically measured at 1.4 ghz ) is directly proportional to its star formation rate ( e.g. , * ? ? ? * ) .
much of the interest in radio investigations of cluster mergers was inspired by a2125 , a merging cluster studied in tandem with the relaxed cluster a2645 in @xcite and @xcite .
the clusters are of the same richness and redshift , yet a2125 has about an order of magnitude more radio galaxies .
subsequent radio studies of presumed cluster mergers include the shapley supercluster ( a3556 , a3558 , and a3562 , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , a2255 @xcite , a2256 @xcite , a large sample of intermediate - redshift clusters @xcite , and a much more comprehensive analysis of a2125 @xcite
. the resulting picture may not be so simple , with the fraction of galaxies hosting radio sources increased in some mergers ( a2125 , a2255 ) , normal in others ( a2256 ) , and potentially even decreased in others ( a3558 ) .
these differences might be explained by noting the examined populations and considering the timing of the merger .
the former point is merely that the effects of a cluster merger are likely different on star - forming galaxies than they would be on brighter and more massive agn , whereas the latter notes that cluster mergers proceed on @xmath11gyr timescales while the radio emission is associated with events whose duration is about an order of magnitude less .
consequently , there are two clear directives for further observations : 1 ) to investigate previously - studied clusters to consistent and deep sensitivity limits , and 2 ) to investigate a larger number of merging clusters in order to better understand the role of merger timing .
the shapley supercluster represents an ideal testbed for further investigation of the effects of cluster mergers on the evolution of member galaxies .
the true extent of the supercluster is not determined , as redshift surveys covering increasing areas of the sky have continued to reveal its enormous scale through interconnected structures ( e.g. , * ? ? ?
. at the center of the supercluster lies a3558 ( or shapley 8) , a very rich cluster ( abell richness class 4 ) flanked by a3556 to the west and a3562 as well as a pair of poor clusters identified in both the optical and x - ray ( sc1327 - 312 and sc1329 - 313 , see e.g. , * ? ? ?
* ; * ? ? ?
* ) to the east .
this collection of clusters forms the core of the supercluster , and has been studied extensively at a variety of wavelengths .
furthermore , all of these clusters lie in the plane of the sky at @xmath12 making this a particularly attractive supercluster for a single observational campaign .
several research collaborations have committed large amounts of optical observing time to the collection of hundreds of galaxy spectra in the shapley supercluster , and the resulting dynamical assessments depict an active region rich in substructure . the supercluster has been studied on large scales ( tens of mpc ) in an attempt to estimate its total mass @xcite since it lies in the direction of the local group s motion relative to the cosmic microwave background .
@xcite presented a substructure analysis of the core of the supercluster , drawing on the velocities presented in a sequence of works @xcite .
they applied the dedica algorithm to identify up to 21 significant substructures within the supercluster core , and concluded that either of two hypotheses held : 1 ) the core of the shapley supercluster represents a cluster - cluster merger viewed shortly after one cluster core has passed the other for the first time , or 2 ) the entire region is a complex of lesser mergers ( such as groups merging with each other and with larger clusters ) .
similarly , x - ray observations indicate a dynamically rich system . the cluster known as sc1327 - 312 was first identified on the basis of _ einstein _ data @xcite , and subsequent studies with various x - ray observatories have continued to reveal interesting properties : for example , _ rosat _
@xcite , _ asca _
@xcite , _ beppo - sax _
@xcite , and _ xmm - newton _ @xcite .
all of the five clusters are linked by filamentary x - ray structure @xcite , and appear to be interacting .
@xcite suggested a dynamical sequence for the clusters in the chain including a relaxed poor cluster ( sc1327 - 312 ) , two rich `` developed '' clusters with substructure ( a3558 and a3562 ) , an ongoing merger ( sc1329 - 313 ) , and a candidate post - merger ( a3556 ) .
their interpretations were largely based on departures from expected relationships between @xmath13 , @xmath14 , and @xmath15 ( optical velocity dispersion ) .
the ongoing merger in sc1329 - 313 was deduced from its fitted fe k@xmath16 line , indicating the cluster was out of ionization - equilibrium ( the fitted energy of the iron line was consistent with the h - like iron k@xmath16 line instead of the usual he - like line expected for the measured temperature of the cluster ) .
@xcite were only able to confirm this result for sc1329 - 313 at the 2@xmath17 level , and instead argued that the whole system represented the aftermath of a major cluster - cluster merger seen after the first core passage , an interpretation also based on their optical and radio investigations . in this interpretation , the clusters other than a3558 are the remains of the merger .
@xcite argued that sc1329 - 313 had recently passed through the northern outskirts of a3562 . despite the varied interpretations of the merger history and partners ,
it is clear that the core of the shapley supercluster is a dynamically - rich system .
prior radio works targeting specific regions within the system also reveal a complex environment .
@xcite studied a3556 using the australia telescope compact array ( atca ) and molonglo observatory synthesis telescope ( most ) , identifying nine cluster radio galaxies including the unusual narrow - angle tail source j1324 - 3138 @xcite .
this appears to be the remnant of a radio galaxy , where the central engine has switched off possibly in response to the cluster merger event .
@xcite extended the same general radio survey of the shapley supercluster to include the cores of the other clusters in the chain , increasing the number of cluster radio galaxies to 28 . using these identifications to construct the radio luminosity function ( rlf ) , they found a deficit of powerful radio galaxies in the system relative to the cluster rlf determined in @xcite .
they argued that cluster mergers may switch off existing radio sources , as appeared to be the case for j1324 - 3138 .
@xcite continued the overall survey , in particular covering the region around a3562 .
they identified 33 cluster radio galaxies ( 26 of which were not in the preceding works ) , and found that the deficit of powerful radio galaxies noted in @xcite was restricted to a3558 . among these sources
were a number of lower radio luminosity sources , presumed to be starbursts . in this paper , a comprehensive radio and optical view of the entire core of the supercluster
is presented .
the radio data were obtained with the nrao very large array ( vla ) and consist of a 63-pointing mosaic which covers nearly 7 square degrees at a fairly uniform sensitivity of @xmath1180 @xmath0jy per 16 beam .
this makes it unique in that it covers the entire core of the shapley supercluster in a single observational campaign , and implies that agn and galaxies forming stars at rates as low as 1.0 m@xmath1 year@xmath2 are detected .
the radio data are complemented by new optical imaging in @xmath3 and @xmath18 .
the areal coverage and uniform sensitivity of these data represent the key advantages of this study : populations of active galaxies down to low activity levels may be studied across the full range of environmental conditions within the supercluster .
this enables stronger conclusions about the potential correlation of increased activity within individual galaxies to regions of dynamical activity within the clusters which make up the supercluster . for the sake of consistency with prior papers investigating the radio galaxy populations of nearby clusters , a cosmology with @xmath19 km s@xmath2
mpc@xmath2 and @xmath20 has been adopted .
this yields a luminosity distance of 186.8 mpc to the supercluster ( using @xmath21 ) , meaning @xmath22 kpc . for reference
, these values become @xmath23 mpc and @xmath24 kpc for the _ wmap _ cosmology with @xmath25 km s@xmath2 mpc@xmath2 and @xmath26 , @xmath27 .
the initial radio observations were performed in 2001 june with the vla in its cnb configuration .
this configuration , observing at a frequency of 1.4 ghz , is well suited to the study of star - forming galaxies in the shapley supercluster with an angular resolution of @xmath1115 ( 12.4 kpc ) , meaning low surface brightness emission spread over the disk of a galaxy will not be missed .
the extended north arm of the configuration also results in a more circular beam .
the program was scheduled over five days , each consisting of a 6-hour block centered on transit of a3558 ( see table [ tbl - radobs ] for a listing of all pointings and observation dates ) .
unfortunately , data were lost due to thunderstorms and strong winds stowing the array on 2001 june 20 and mechanical problems on 2001 june 21 . to replace these data ,
a single additional 6-hour track was scheduled during the subsequent cnb configuration of the vla in 2002 september .
the conditions for these observations were less favorable , as they were performed during daytime where the sun is a source of additional noise .
the observational strategy consisted of observing the core of the supercluster through a mosaic of individual vla pointings . these were arranged in a hexagonal grid to provide nearly uniform sensitivity across the entire area , with an 18 spacing between adjacent pointings .
this mosaic strategy is the same as that employed by the nrao vla sky survey ( nvss , * ? ? ?
* ) and faint images of the radio sky at twenty centimeters ( first , * ? ? ?
* ) although with a tighter grid spacing . originally , 70 pointings were planned but this was pared down to 63 due to the lost observing time .
the seven pointings removed from the grid were taken from the edges to minimize the effect on the final mosaic coverage .
the final 63 pointings are summarized in table [ tbl - radobs ] , and a footprint of the covered area may be viewed in figure [ fig - foot ] .
the radio data were obtained in line mode at 1.4 ghz , consisting of 28 channels each of 3.125 mhz bandwidth ( two sets of seven channels at each polarization ) .
each pointing center was visited only once to minimize the amount of time spent moving among pointings . to account for the varying system temperature of the vla at 1.4 ghz as a function of elevation
, the dwell times were adjusted in order to achieve a similar sensitivity at each pointing .
the minimum dwell time ( for sources observed near transit ) was 20 minutes , while sources at the beginning and ends of the runs were observed for 27 minutes .
although the author sure thought he was clever in doing this , he failed to respect the importance of good ( @xmath28 ) coverage . for the most part
, the pointings were prioritized such that the edges of the supercluster corresponded to those with poorer ( @xmath28 ) coverage . to improve the ( @xmath28 ) coverage of a few central pointings ,
they were revisited for brief periods ( @xmath29 minutes ) during the 29 september 2002 observations .
the effect of the ( @xmath28 ) coverage will be addressed in more detail below .
flux calibration was achieved using 3c286 , with calibration of phases and bandpasses performed from roughly hourly observations of the nearby calibrator source j1316 - 336 .
the nrao s astronomical image processing system ( aips ) was used to calibrate and image the data , following the usual procedures for 1.4 ghz data .
briefly , the data for each pointing were calibrated ( including antenna - specific weights ) and reduced individually .
ideally , one would handle the ( @xmath28 ) data for each pointing in a manner which both produces a consistent beam across all pointings and does so with good characteristics ( i.e. , a nice gaussian with minimal sidelobes ) .
this is where the effect of moving to each position in the pointing grid only once is most relevant . for several of the pointings at the edges of the pointing grid ,
the resulting beam is highly elongated .
this renders it virtually impossible to obtain a consistent beam across all pointings , necessitating the compromise of choosing parameters which produce beams with the same area and using a circular restoring beam of equal area in the final maps .
table [ tbl - radobs ] indicates the initial beam sizes .
each pointing was imaged in @xmath29 fields , with four fields covering the primary beam and additional fields dedicated to outlying bright sources .
iterations of imaging and self - calibration were performed until the final images for each pointing were created . at this point ,
the cleaned sources were restored with a 16 circular beam for combining into the final mosaic .
the rms noise of these individual pointings ranged from about 75 to 125 @xmath0jy beam@xmath2 .
the final mosaic was created by combining the reduced images for all 63 pointings . at this stage , the flanking fields dedicated to bright sources were ignored and only the four fields covering the primary beam for each pointing were used .
the contribution of each image to the mosaic was properly weighted by its signal - to - noise , as discussed in @xcite , with the images truncated at the point where the response of the vla had dropped to @xmath30 of its peak ( i.e. , each pointing images about a 40 diameter ) .
the noise characteristics of the final mosaic are quite good , with a typical rms of about 80 @xmath0jy beam@xmath2 over the area described as within 2 mpc of the centers of the individual abell clusters .
the noisiest region within this area has a local noise around 115 @xmath0jy beam@xmath2 and is found to the east of the center of abell 3558 .
it is caused by several bright sources , particularly the quasar at j133019.1 - 312259 ( e.g. , * ? ? ?
the cleanest regions have rms noise as low as 62 @xmath0jy beam@xmath2 .
the final mosaic is available upon request from the author .
the catalog of radio sources was then created using the task sad , which identifies peaks in a map and fits them with gaussians .
parameters of importance are saved from these fits , including the source position , peak flux density , integral flux density , major and minor axis , plus position angle .
sad also creates a residual map , which was inspected to identify any sources for which gaussian fits failed ( usually strong extended sources ) .
these were manually added to the radio source catalog .
the relative flux scale was found to be consistent with the nvss via comparison of the fluxes of unresolved sources .
optical images in both @xmath3 and @xmath18 ( hereafter referred to simply as `` @xmath4 '' ) filters were obtained in 2002 march and april using the cerro tololo inter - american observatory 1.5 m telescope with the cassegrain focus ccd imager . using the @xmath31 focus provided a wide field of view of 148 with 044 pixels .
even so , the large area of the core of the supercluster required numerous individual telescope pointings : 152 fields in each @xmath3 and @xmath4 , arranged in a simple grid with a 14 spacing .
figure [ fig - foot ] depicts the coverage of the optical fields in relation to the radio mosaic .
data for this project were collected on the nights of march 24 and 25 ( kindly provided by michael ledlow during a run for a related program ) , and march 29 , march 31 , and april 1 .
sky brightness was an issue , as these dates straddled the full moon .
the exposures were prioritized such that the @xmath3 images were collected during the darker conditions , and exposure times were adjusted slightly for the conditions on a nightly basis .
the @xmath4 exposures ranged from 120 to 150 seconds , while the @xmath3 exposures were either 180 or 240 seconds . the seeing varied from 11 ( about the best possible with the @xmath31 focus ) to 20 , with most exposures falling in the @xmath32 range .
data reduction followed the standard reduction steps .
the images were bias subtracted and then flat fielded using twilight frames collected during several of the nights .
the astrometry was registered using about 20 unsaturated usno a2.0 stars per field @xcite , yielding rms errors of under 025 .
the photometry was set by observations of `` selected area '' fields from @xcite , performed nightly and at a range of airmass .
the derived relationships were : @xmath33 @xmath34 where @xmath35 is the background - subtracted counts from the source , @xmath36 is the integration time measured in seconds , and @xmath37 is the airmass . the consistency of the photometry was checked using stars at the edges of the science exposures , where the overlap of the pointings produced numerous comparison objects ( see below ) .
lists of optical sources were generated by field using sextractor @xcite .
the most important parameters proved to be those related to the background as scattered light was an issue , particularly for the @xmath4 images collected during times of high sky brightness .
this produced images with a gradient in the apparent background which often changed sharply along one edge of the ccd .
choosing a background mesh size too large failed to properly respond to this effect , while smaller background mesh sizes performed poorly around the larger real objects . in this latter case , some of the counts corresponding to the bigger elliptical galaxies and brighter saturated stars
get included in the background resulting in extracted magnitudes for these objects which are fainter than their true magnitudes . since
one of the main uses of the optical magnitudes is to select galaxies brighter than prescribed limits , the chosen mesh size erred more toward correct handling of the scattered light .
this means that the reported magnitudes for the brightest elliptical galaxies ( @xmath38 ) are up to @xmath39 mags fainter than their real values .
the derived magnitudes for all fields correspond to the fixed aperture size of 159 in radius , which is the gunn - oke aperture @xcite at the assumed redshift of the supercluster ( @xmath21 ) .
initially , the color term in the above photometric equations was ignored and later introduced when the output catalogs were merged .
further , when the source catalogs for the individual fields were merged the magnitudes were also corrected for galactic extinction using the values of @xcite , based on the @xmath40 and @xmath41 values for the center of each 148 field . limiting magnitudes , based on inspection of number count histograms , ranged from 18.5 to 19.5 in @xmath4 and from 19.5 to 21 in @xmath3 . the consistency of the photometry from field to field was then checked using the regions of overlap .
the magnitude errors reported by sextractor are based simply on counting statistics , so this step essentially determines the errors resulting from application of the above photometric equations to the collected data .
it would also reveal any magnitude zero point shifts as would occur under non - photometric conditions .
all designated stars whose photometry was not flagged by sextractor were used .
the consistency of the @xmath3 photometry was excellent , with a standard deviation under 0.05 mag for objects with magnitudes brighter than @xmath42 and still under 0.15 mag for galaxies as faint as @xmath43 .
as previously suggested , the @xmath4 photometry was less consistent with a standard deviation of less than 0.1 mag out to @xmath44 and rising to 0.2 mag at @xmath45 . for simplicity
, these findings have been generalized and an additional error of 0.05 mag in @xmath3 and 0.1 mag in @xmath4 have been included in the reported photometry .
lastly , the merged catalog of all optical sources was edited to remove duplicates ( see additional discussion below ) .
the radio and optical source catalogs were then merged to create the final list of radio galaxies .
the adopted conventions for creation of the radio galaxy catalog were : * the radio peak flux must be greater than 330 @xmath0jy , chosen as a 5@xmath17 detection in the regions of the mosaic map with the lowest noise .
* either the radio peak flux or integral flux must be greater than five times the local rms noise , as evaluated in a 6 box centered on the peak in the radio emission .
* the galaxy must be brighter than @xmath46 , which corresponds to @xmath47 in the adopted cosmology .
this limit is one magnitude fainter than that used in @xcite and related papers , allowing the present study to sample cluster radio galaxies down to fainter levels .
it is also safely brighter than the completeness limit of the worst @xmath4 band fields .
although radio sources with detected optical counterparts fainter than 17.36 are present in the data , most such radio galaxies are safely assumed background objects .
* the separation between radio and optical position must be less than 7 ( except in cases of extended powerful radio galaxies ) .
the probability of chance coincidence of radio and optical objects satisfying the above criteria was evaluated by shifting the optical catalog by arbitrary amounts and re - performing the correlation . from this analysis
, it is expected that about 4 sources in the final list are false .
however , the false detection probability for sources within 2 mpc of the cluster centers and with @xmath48 and @xmath49 w hz@xmath2 is much lower .
these are the limits used for the comparison sample in the radio galaxy fraction analysis ( see section [ sec - rgfrac ] ) , and the false detection probability corresponding to the 7 separation limit approximately matches that used in construction of that comparison sample . perhaps the most critical step in source identification was a thorough visual inspection .
the radio contours were overlaid on the optical images along with markers identifying sources in each the radio and optical catalogs . in the case of the radio data ,
this step allowed for extended radio galaxies to be identified in addition to any other problem cases not fit by sad .
the extended radio sources are often missed because the peak in their emission does not necessarily coincide with the optical location of the galaxy . for the optical data ,
the visual inspection provided a check on the sextractor results including the star / galaxy segregation for problem objects such as close pairs of faint stars and galaxies .
these false bright objects were removed from the catalog of optical sources .
the use of markers for each of the radio and optical catalogs also enabled the removal of any duplications which were missed in the previous steps .
finally , a small number of optical galaxies in close proximity to bright stars were absent from the sextractor catalogs .
magnitudes for these galaxies were determined manually by summing counts within irregular shaped apertures designed to avoid contamination by the nearby stars .
the final list is presented in table [ tbl - radgals ] .
this table includes the optical position , magnitudes and color , integral radio flux , peak radio flux , a marker indicating whether the radio emission was unresolved , local noise at the location of the radio source , and separation between the optical and radio positions .
those sources for which the aips task jmfit found a minimum size of zero for the major axis were assumed to be unresolved .
for such sources , the fitted peak flux density is a better representation of the true flux of the source than the fitted integral radio flux ( e.g. , see * ? ? ? * ) .
in addition , the nasa / ipac extragalactic database ( ned ) was searched to determine whether velocity information was available for each of the 210 entries in table [ tbl - radgals ] .
a total of 123 galaxies had public velocity information , including at least 104 members of the shapley supercluster ( see next section ) .
this information is also presented in table [ tbl - radgals ] . in all subsequent text ,
the 210 radio - detected galaxies will be referred to simply as `` radio galaxies , '' with the 104 spectroscopically - confirmed members of the shapley supercluster called `` cluster radio galaxies '' or sometimes `` cluster members . ''
to determine which radio galaxies belong to the shapley supercluster , the individual cluster recession velocities and dispersions ( @xmath15 ) of @xcite were used . allowing galaxies within @xmath50 of any cluster / substructure , the minimum and maximum velocities describing the supercluster are 10934 km s@xmath2 and 17684 km s@xmath2 . in practice , these limits correspond to those for the brighter galaxies in a3558 ( @xmath51 km s@xmath2 and @xmath52 km s@xmath2 ; * ? ? ? * ) . as noted previously ,
there are 104 cluster member radio galaxies identified in the present study , nearly doubling the total identified in prior studies .
these include all of the cluster radio galaxies reported in the prior radio studies of the supercluster @xcite , with the exception of two spirals with low integral flux.jy beam@xmath2 rms .
] of the 17 radio galaxies with velocities formally placing them outside the supercluster , 6 are background radio sources and 11 are foreground .
the majority of these foreground objects are likely related to the supercluster , having velocities greater than 9000 km s@xmath2 and typically located in the eastern region of the studied area .
the other three foreground galaxies belong to the hydra - centaurus wall ( @xmath53 km s@xmath2 , see * ? ? ?
a further two radio galaxies were associated with galaxy pairs in ned .
although the galaxies hosting the radio emission did not have ned velocities , their companions did have cluster velocities and hence they may tentatively be considered cluster members which would bring the total to 106 .
there are 87 remaining radio galaxies for which no velocity data are available . of these , 19 are probable background galaxies on the basis of their @xmath54 colors ( see below ) .
table [ tbl - radgals ] includes nine galaxies with radio emission extended beyond their optical sizes , seven of which are cluster members ( see figures [ fig-132145 ] - [ fig-133542 ] ) . of the cluster members , three have @xmath55 w hz@xmath2 and
are thus in the realm of classical fanaroff - riley objects @xcite : j132357 - 313845 , j132802 - 314521 , and j133331 - 314058 . in each case
, the radio morphology appears to be `` head - tail '' in nature .
j132357 - 313845 , depicted in figure [ fig-132357 ] , is discussed at length in @xcite .
similarly , j133331 - 314058 ( figure [ fig-133331 ] ) is analyzed in detail in @xcite .
the third possible head - tail source , j132802 - 314521 , consists of a strong core centered on the galaxy plus what appears to be a diffuse tail ( figure [ fig-132802 ] ) .
it is possible that these features are unrelated and the potential tail is a background source .
however , the luminosity of the core portion of the radio emission is @xmath56 w hz@xmath2 .
such high luminosities are usually associated with extended emission , lending credence to the association of the diffuse emission as a low surface brightness extension of the source . with this emission
, the luminosity becomes @xmath57 w hz@xmath2 .
similarly , the optical magnitude of the galaxy ( @xmath58 ) is typical of fanaroff - riley sources .
a caveat to this interpretation is that the higher resolution observations of @xcite ( 10@xmath59 5 ) did not resolve the core emission , as might be expected if it were a jet associated with the diffuse emission .
the detection of up to three head - tail sources across the clusters at the core of the shapley supercluster is typical of rich clusters in general ( e.g. , see * ? ? ?
an additional four extended radio galaxies belonging to the cluster have general morphologies which would place them among fanarof - riley objects , although in each case the luminosity is below @xmath60 w hz@xmath2 .
these include : j132145 - 310300 ( figure [ fig-132145 ] ) , with emission extended to the north and south suggestive of weak `` jets '' but @xmath61 of only @xmath62 w hz@xmath2 ; j132206 - 314616 ( figure [ fig-132206 ] and see * ? ? ?
* ) , which has the appearance of a compact double but a luminosity of only @xmath63 w hz@xmath2 , much lower than the @xmath64 w hz@xmath2 typical of such sources ; j133048 - 314325 ( figure [ fig-133048 ] ) , with @xmath65 w hz@xmath2 and discussed in more detail in section [ ssec - evolution ] ; and j133542 - 315354 ( figure [ fig-133542 ] ) at @xmath66 w hz@xmath2 , which might be contaminated by emission from faint background galaxies to the west .
the remaining two extended radio galaxies are strong background agn : j132210 - 312805 ( figure [ fig-132210 ] ) exhibits the morphology of a strong double - lobed radio source , and j132950 - 312258 ( figure [ fig-132950 ] ) with a ned velocity of 58775 km s@xmath2 implying @xmath67 w hz@xmath2 .
an advantage to having optical data in two filters is that the cluster red sequence ( also referred to as the e / s0 ridge line ) may be identified .
this is useful in characterizing the radio galaxies , particularly those for which spectra are not available .
galaxies which lie along the red sequence are most likely supercluster member ellipticals , while galaxies above the red sequence ( i.e. , larger @xmath54 for a given @xmath4 magnitude ) may safely be assumed to be background objects .
bluer objects may be star - forming galaxies associated with the supercluster , but may also be foreground or background objects . in most cases ,
ned velocities enable this determination .
the red sequence was fit using the method of @xcite , with a few small variations .
briefly , the red sequence is parametrized as a line such that @xmath68 where @xmath69 is the @xmath54 color for a given galaxy , @xmath70 is its @xmath4 magnitude , and @xmath71 and @xmath72 are the intercept and slope , respectively . to fit the red sequence , the deviation of each galaxy from a specified @xmath71 and @xmath72 is determined .
the distribution of these deviations is evaluated using the biweight location and scale @xcite , and a range of @xmath71 and @xmath72 are searched to find the pair that minimizes these quantities .
all galaxies brighter than @xmath73 were used in the fits . instead of using a radial cutoff of 0.2@xmath74 as in @xcite
, in the present paper the red sequence was fit for galaxies within 2 mpc of the adopted cluster centers .
in addition , the galaxies used in fitting the red sequence were iteratively 3@xmath17 clipped to remove outliers .
this clipping generally only removed very red background galaxies .
the resulting fits are depicted in figure [ fig - redseq ] and summarized in table [ tbl - rsfit ] .
the fits for the individual clusters have steeper slopes than those of similar redshift clusters reported in @xcite .
the relationship derived in that study predicts a slope of @xmath75 , shallower than all of the fitted slopes in the present study . however , this difference is unlikely to be significant .
a simple test is to evaluate the colors predicted by the different fits in table [ tbl - rsfit ] for representative @xmath4 magnitudes of cluster galaxies . for galaxies with @xmath76 from 12.7 to 17.3 ,
the fits for the individual clusters produce colors which vary by only 0.1 less than the photometric errors for the galaxies . for context , the derived relation for the full galaxy sample regardless of radial separation from the cluster cores has a slope of @xmath77 , comparable to the generalized results of @xcite . as the primary purpose of the red sequence fits for this paper
is the identification of background galaxies and categorization of cluster radio galaxies as star - forming or agn , further analysis of the red sequence fits is deferred .
colors for the radio galaxies presented in table [ tbl - radgals ] are depicted in figure [ fig - colors ] .
the change in the nature of the galaxy activity is apparent .
the optically - brightest galaxies predominantly lie on the cluster red sequence and are presumably powered by agn .
moving to fainter optical magnitudes , bluer colors are more prevalent as star formation becomes the origin for the radio emission .
one of the key science drivers for studying the shapley supercluster is to assess the fraction of active galaxies in relation to other clusters of galaxies . for this purpose ,
the analysis presented in ( * ? ? ?
* hereafter mo03 ) for a2255 is performed for the individual clusters in the shapley supercluster using the 18 nearby clusters presented in @xcite as a comparison sample .
briefly , the radio galaxy fraction for a given cluster is evaluated as the number of radio galaxies with luminosity greater than a prescribed limit of @xmath78 w hz@xmath2 divided by the total number of galaxies .
this radio luminosity is based on the completeness limit of the nvss at the most distant cluster of the comparison sample , and assumes a spectral index of 0.7 for the k correction ( where @xmath79 ) . in the shapley supercluster
, it corresponds to a flux of 1.64 mjy which is well above the lower limits of the full radio mosaic .
velocity information is applied in order to remove foreground and background objects from the radio galaxies which comprise the numerator of the fraction .
the allowed velocity ranges for each cluster , taken from @xcite and corresponding to @xmath50 about each cluster s systemic velocity , were : a3556 , 12428 16286 km s@xmath2 ; a3558 , 10934 17391 km s@xmath2 ; a3562 , 11753 17231 km s@xmath2 ; sc1327 - 312 , 12771 16971 km s@xmath2 ; and sc1329 - 313 , 12520 15921 km s@xmath2 .
complete velocity information is unavailable for the denominator , so the galaxy counts are corrected for contamination as in mo03 by assuming @xmath80 , where @xmath81 galaxies steradian@xmath2 .
this value of @xmath82 was derived directly from regions outside the comparison sample clusters .
the analysis is performed for galaxies with @xmath48 and within 2 mpc projected separation of the respective cluster centers .
the adopted center positions were ( j2000 coordinates ) : a3556 , 13:24:06.2 -31:39:38 ; a3558 , 13:27:54.8 -31:29:32 ; a3562 , 13:33:31.8 -31:40:23 ; sc1327 - 312 , 13:29:47.0 -31:36:29 ; and sc1329 - 313 , 13:31:36.0 -31:48:46 . as in mo03 ,
the radio galaxy fractions are evaluated for optically bright objects ( @xmath8 ) , intermediate objects ( @xmath83 , a range which includes @xmath84 ) , and faint objects ( @xmath85 ) .
the actual testing is performed using a @xmath86 statistic which evaluates the probability that a given cluster deviates from the fraction predicted by the pooled comparison sample .
table [ tbl - abs ] is provided to assist the reader in identifying which galaxies from table [ tbl - radgals ] enter the radio galaxy fraction analysis .
in addition to the galaxies optical positions , it includes their absolute @xmath4 magnitudes , radio luminosities ( computed using integral fluxes for resolved sources and peak fluxes for unresolved sources , as described in section [ sec - id ] ) , and projected distance to each of the five clusters .
there are 40 galaxies which meet the requirements outlined above , plus an additional seven for which velocities are not available .
of these seven , two are the aforementioned members of galaxy pairs whose partners have measured velocities placing them within the supercluster . as such , they are probable members of the supercluster .
the remaining five galaxies include four potential cluster members and a probable background galaxy , evaluated on the basis of their optical colors .
analysis of radio galaxy fractions has been performed both with and without the six possible supercluster galaxies which lack direct velocity measurements ( the presumed background galaxy was removed from consideration ) . as they are distributed across the supercluster , at most two such galaxies
enter the calculation for any individual cluster .
consequently , they have little effect on the calculated fractions ( see tables [ tbl - brightfracs ] - [ tbl - faintfracs ] ) .
the latter portion of table [ tbl - abs ] includes the same information for cluster radio galaxies which did not enter the radio galaxy fraction analysis .
results of the radio galaxy fraction analysis are presented in tables [ tbl - brightfracs ] , [ tbl - intfracs ] , and [ tbl - faintfracs ] for the bright , intermediate , and faint optical magnitude bins .
there is frequent evidence for significant variation in the radio galaxy fraction in the shapley clusters . in general
, there appears to be a deficit of radio galaxies in the intermediate optical magnitude bin ( table [ tbl - intfracs ] ) .
this holds for all five clusters at greater than the 95% confidence level , although inclusion of the few galaxies without velocities reduces the significance below 95% for a3562 and sc1329 - 313 .
the comparison sample indicates that about 15% of galaxies in this optical magnitude range should be radio sources , while the values computed for individual clusters in the shapley supercluster were all under 10% , even when the radio galaxies lacking velocities were included . to assess the importance of cosmic variance in the computed significance levels ,
the fractions were recalculated under the assumption that all counted galaxies were cluster members ( i.e. , no background correction and hence a lower limit to the radio galaxy fraction ) and using twice the assumed background correction .
the resulting ranges for the calculated significance of any excess or deficit of radio galaxies are included in tables [ tbl - brightfracs ] - [ tbl - faintfracs ] .
it can be seen that the background correction does not appear to cause the deficit of intermediate optical magnitude radio galaxies .
the reduction in activity among intermediate optical magnitude galaxies is offset somewhat by an increase in the probability for the brightest galaxies to host radio sources , although number statistics limit the significance of this statement .
the clusters generally have about 10 to 20 galaxies in this magnitude range , of which the comparison sample indicates that 24% should be radio sources . in a3556 ,
more than half of all bright galaxies are radio detections , translating to an increase in activity significant at about 99% confidence . with the exception of sc1329 - 313 , the other clusters have marginal excesses significant at about the 85% level .
these results are much more dependent on the background correction . as with a2255
, the most significant variation is found in the optically - fainter galaxies .
the optical luminosity function insures that there are many such galaxies per cluster , and in general the fraction of these galaxies which are also radio emitters is quite small
less than 2% .
both a3556 and a3558 have radio galaxy fractions consistent with this figure , but a3562 has a very strong excess of radio galaxies among its fainter galaxy population with nearly 11% of these galaxies being radio sources .
the significance of this excess is well over 99.9% , regardless of the background correction or whether a single galaxy without measured velocity is included .
as would be expected based on the large overlap in areas covered by the 2 mpc radial sampling limit ( e.g. , refer to figure [ fig - foot ] ) , sc1327 - 312 and sc1329 - 313 also have larger radio galaxy fractions than would be expected based on the comparison sample , at 98.5% and 99.9% significance respectively .
the significance level for each of these three clusters would be higher if their allowed velocity ranges for radio galaxies were greater ( i.e. , as less rich systems within the supercluster they have lower @xmath15 and consequently some radio galaxies with velocities consistent with supercluster membership are excluded from consideration as being outside the allowed range of @xmath87 for the individual clusters ) .
an advantage to performing one study for the entire core of the supercluster is that any variations in radio galaxy fraction across the clusters may be assessed in light of environmental factors . to this end
, a map showing the distribution of the radio galaxy fractions was created . at each point in the map
the radio galaxy fraction was computed within a 0.5 mpc radius , using a radio detection threshold of @xmath78 w hz@xmath2 and corrected galaxy counts with magnitudes @xmath48 ( i.e. , luminosity and magnitude limits identical to the preceding analysis ) .
the choice of a 0.5 mpc sampling region was motivated by the desire to include large enough areas to have meaningful numbers of sources while still avoiding edge effects within the main body of the supercluster .
including regions beyond 2 mpc of the respective cluster centers adds three additional radio galaxies assumed to be cluster members but for which no velocity information was available , and two further cases where radio detections without velocity measurements were assumed to be background sources on the basis of their red colors .
thus , contamination by foreground / background radio galaxies is still minimal .
the resulting map is presented in figure [ fig - fracn ] , where the grey - scale indicates the radio galaxy fraction . to help indicate where high fractions result from low number statistics ( e.g. , only two local galaxies of
which one is a radio source ) , contours of galaxy surface density are provided .
the increase in activity in a3562 is readily apparent , particularly in the direction of a3558 .
the limits of @xmath78 w hz@xmath2 and @xmath48 for the fractions analysis were set by the available comparison sample data .
the present study includes much deeper data , so it is illustrative to consider fainter sources . figure [ fig - fract ] parallels figure [ fig - fracn ] in that it is for sources with @xmath48 , but in this case radio sources with @xmath88 w hz@xmath2 are considered detections . the lower limit on radio luminosity corresponds to 0.5 mjy ( i.e. , 5@xmath17 detections in the noisier regions of the mosaic ) while the upper limit prevents direct overlap with the radio galaxies used in figure [ fig - fracn ] .
figure [ fig - fracp ] is for the optically faintest galaxies in the present study , those with @xmath89 .
the radio detection limit is again @xmath90 w hz@xmath2 ; none of these fainter galaxies have @xmath91 w hz@xmath2 .
as would be expected , most of the optically - fainter galaxies lack spectroscopic velocity measurements and are evaluated as cluster or non - cluster sources on the basis of @xmath54 color .
the identification of 210 radio galaxies in the core of the shapley supercluster provides a good database for evaluation of galaxy activity .
only 34 of these radio galaxies may be removed from consideration as galaxies merely seen in projection on the supercluster , half of which are known foreground / background objects on the basis of existing optical spectroscopy with the other half appearing too red to be consistent with supercluster membership . of 104 confirmed cluster members ,
at least 40 may be used for a statistical analysis of radio galaxy fractions which parallels that of mo03 .
the analysis of radio galaxy fractions indicates varied activity across the supercluster . in general , there are two broad findings which need to be understood : 1 ) the decrease in the radio galaxy fraction for galaxies with optical magnitudes near @xmath92 , generally seen across the entire core of the supercluster ( i.e. , the whole area studied ) ; and 2 ) the strong increase in the radio galaxy fraction for optically - faint galaxies in the vicinity of a3562 .
these two points will now be discussed in light of environmental effects and prior theoretical and observational findings .
examination of the locations of the radio galaxies ( figures [ fig - fracn ] - [ fig - fracp ] ) points to the region around a3562 as one of high activity .
this is especially the case for the abundant optically - fainter galaxies , where the increase in the radio galaxy fraction for a3562 was significant at very high confidence .
the fact that this increase corresponds to a specific region within the core of the supercluster indicates that there must be some difference in environment intrinsic to this region . in mo03
, it was argued that a2255 had a larger active galaxy population as a result of an ongoing cluster - cluster merger .
the a2255 merger axis is believed to be perpendicular to the observer and viewed very near the time when the cores of the merging partners are coincident . according to the simulations of @xcite , at this particular time galaxies will cross the shock front formed between the merging clusters and experience a spike in ram pressure , thereby instigating starbursts .
recent x - ray analysis supports a similar picture in regards to the merger history and timing for a3562 .
@xcite used a 6-pointing _ xmm _ mosaic to study the hydrodynamics of a3562 and the sc1329 - 313 group .
they found an interesting assortment of properties , including temperature and entropy substructure and evidence for disruption of the core of sc1329 - 313 . on @xmath11100 kpc scales , the emission of a3562 and sc1329 - 313 point towards one another , indicating interaction .
the properties are well described by a scenario in which sc1329 - 313 has recently passed to the north of the core of a3562 , traveling toward the west , and been deflected to the south and towards the observer .
the motion was supersonic , having an implied mach number @xmath931.3 .
this passage would then have created a shock wave , as well as having produced oscillations of the core of a3562 .
examination of figure [ fig - fracn ] is qualitatively consistent with this scenario .
the region associated with an excess of radio galaxies corresponds to the expected location of the shock associated with the passage of sc1329 - 313 . in the @xcite model ,
early in the merger the shock front acts as a buffer for any gas - rich galaxies in sc1329 - 313 , protecting them from ram pressure removal of their gas .
however , at a time around core passage the shock front decelerates and the galaxies pass through it , thereby experiencing a sharp increase in ram pressure leading to starbursts .
an important consideration to this model for the shapley supercluster , however , is the pre - merger nature of the active galaxies .
should they have been members of either a3562 or sc1329 - 313 , the high efficiency of ram pressure stripping ( e.g. , * ? ? ?
* ) implies that it would be reasonable to expect that they were already gas poor and not likely to undergo a subsequent starburst .
consequently , the cluster merger itself may be indicative of larger - scale interactions : as the pair of clusters merge along a filament , outlying groups and field galaxies along the filament are also thrown into the mix .
it is these galaxies which would cause the increase in the observed radio galaxy fraction . in a similar light ,
the merger may induce activity in the outlying groups through tidal interactions @xcite . a picture along these same lines
is evoked for the unusually high fraction of radio galaxies in a2125 @xcite , and seems to fit for the curious radio galaxy j133048 - 314325 in the shapley supercluster .
figure [ fig-133048 ] depicts j133048 - 314325 , which superficially has the radio morphology of a wide - angle tail ( wat ) but a radio luminosity a factor of ten lower than such sources .
other galaxies are easily visible nearby this source , the radio galaxy j133051 - 314417 and a face - on spiral galaxy ( j133047 - 314339 ) within the radio contours of j133048 - 314325 ( note that this spiral galaxy was designated the radio counterpart in lower resolution observations by * ? ? ?
these three galaxies appear to be a compact group , as indicated by their small spatial and velocity separation ( a maximum separation of just 1 , or about 50 kpc ; ned velocities of 13394 , 13226 , and 13815 km s@xmath2 in order of increasing galaxy ra ) .
a potentially useful analogy for this system is that of stephan s quintet in which a high velocity intruder galaxy ( ngc 7318a ) has interacted with the intragroup medium ( igm ) around ngc 7318a and ngc 7319 .
radio continuum emission from this system ( e.g. , * ? ? ?
* ) arises from a combination of sources , including the seyfert nucleus of ngc 7319 , star formation , and the large ( @xmath94 kpc ) shock caused by the interaction .
much of the star formation , interestingly enough , is within the igm and not simply associated with individual galaxies of stephan s quintet .
while this star formation does contribute to the 1.4 ghz radio continuum , the majority of the radio emission is associated with a radio ridge delineating the shock front ( about 1.5 mjy for star formation , 35 mjy for the shock , and 29 mjy for the seyfert galaxy ngc 7319 ) .
inspection of the radio contours of j133048 - 314325 suggests a similar interaction of the sources , potentially explaining their activity .
j133051 - 314417 would be the intruder galaxy , with support for this interpretation being its @xmath95 km s@xmath2 velocity offset from the other galaxies and the general shape of the radio emission which is suggestive of j133051 - 314417 having passed through the igm in a southeasterly direction .
the extended morphology of j133048 - 314325 would then be the combination of possible agn emission from the galaxy itself , plus shock and star formation emission . that j133048 - 314325 might be an agn
is suggested by its optical properties , which include what might be an unresolved optical nucleus coincident with the peak in the radio emission , its blue color @xmath96 , and an unusually bright absolute magnitude of @xmath97 .
it is possible that the unresolved optical component and bright total magnitude arise from a foreground star , as no emission lines are noted in the spectroscopy reported by @xcite , the literature source for the ned velocity .
however , it would be a remarkable coincidence if a foreground star were to lie in projection directly on the nucleus of a galaxy which has moderately strong radio emission .
the portion of the j133048 - 314325 emission that would be associated with the shock would then be the ridge to the southeast .
using tvstat to evaluate the flux associated with this feature produces a radio luminosity of @xmath98 w hz@xmath2 , comparable to that of the shock ridge in stephan s quintet ( @xmath99 w hz@xmath2 ) .
its size , @xmath1160 kpc , is also similar to that of the shock ridge in stephan s quintet .
a contribution to the total emission along the northwest edge could then come from star formation in the spiral galaxy j133047 - 314339 .
the entire system lies about 11 to the northwest of the cataloged position of sc1329 - 313 , very near the peak in the galaxy surface density indicated in figure [ fig - fracn ] .
thus , it is reasonable to think that the hypothesized interaction akin to that seen in stephan s quintet is the result of tidal forces from the larger merger of sc1329 - 313 with a3562 .
this hypothesis can easily be tested by future observations .
additional radio observations at shorter wavelengths would reveal any spectral index variations , where any shock emission would have steeper spectral index ( e.g. , the shock ridge in stephan s quintet has @xmath100 * ? ? ?
higher resolution observations might also help separate any agn emission associated with j133048 - 314325 .
finally , @xcite use infrared observations and long - slit optical spectroscopy to reveal regions associated with shocked gas and those associated with photoionization from star formation .
returning to our discussion of radio galaxy fractions , there is less direct evidence for the `` preferred '' location of activity around sc1329 - 313 when examining fractions at fainter optical and radio limits .
however , these results may still be interpretted within the framework described above .
the inclusion of galaxies with fainter radio luminosities while keeping the optical limit at @xmath48 ( figure [ fig - fract ] ) produces a more uniform distribution of radio galaxy activity within the supercluster . for
the faintest optical galaxies studied , those with @xmath101 between -19 and -20 , the radio sources are again found mainly within a3562 ( figure [ fig - fracp ] ) . in each of these cases , the star formation rates implied by the radio emission are @xmath102 m@xmath1 yr@xmath2 @xcite .
this would be a substantial amount of star formation in the fainter galaxies ( i.e. , those with @xmath103 ) , and a typical star formation rate for normal brighter spirals such as the milky way .
thus , it again appears that the optically fainter galaxies are likely to be starbursts , and that these types of galaxies are found preferentially in the vicinity of a3562
. why would there be different effects on the faint and intermediate magnitude populations ?
the answer to this question might lie in the different galaxy morphologies that are included within these magnitude ranges .
galaxies in the intermediate magnitude range will be larger and more likely to have significant bulges .
it is a common prediction of evolutionary models that bulges stabilize galaxies against transformation ( e.g. , galaxy - galaxy mergers and harassment , * ? ? ?
* ; * ? ? ?
this is because the predicted evolution is usually driven by the rapid inflow of gas to the nuclei of galaxies , where it creates a nuclear starburst .
thus , brighter galaxies with bulges are more resistant to evolutionary changes than fainter , disk - dominated galaxies .
although this can easily explain why more dramatic change is seen in the fainter galaxy population , it is still difficult to understand why there would be a deficit of intermediate magnitude radio galaxies in the shapley supercluster relative to other clusters .
it is possible that the radio galaxies in this magnitude range found in other clusters represent the steady trickle of isolated field galaxies that enter the clusters , and that the dramatic merger environment of the shapley supercluster precludes such objects . are these results consistent with the prior radio studies of the core of the shapley supercluster ?
* hereafter v00 ) examined a3558 and argued for a reduction in galaxy activity in this region , and ( * ? ? ?
* g04 ) studied a3562 and found a level of activity normal for clusters in general .
v00 noted a reduced fraction of radio galaxies in a3558 relative to the cluster survey of ( * ? ? ?
* lo96 ) . in the present study
, there is a reduced fraction of radio galaxies in the intermediate optical magnitude range for a3558 while the fractions among the brighter and fainter optical magnitudes are normal . in figure [ fig - rlf ] , the cumulative radio luminosity function ( rlf ) for the shapley supercluster and that for a3558 alone are plotted against the rlf derived for the comparison sample from @xcite . in each case , the rlf is for galaxies with @xmath48 and within 2 mpc projected separation of their cluster centers ( hence the a3558 rlf is a subset of the presented shapley supercluster rlf ) .
the lo96 rlf was constructed using only elliptical galaxies with @xmath104 , so it is not immediately comparable to those in the figure .
however , applying a simple scaling which amounts to saying that 13.5% of all cluster galaxies with @xmath48 are ellipticals with @xmath104 produces an excellent agreement with the bright end of the rlf derived from the @xcite comparison sample .
since essentially all radio galaxies with @xmath55 w hz@xmath2 are bright ellipticals , this agreement is expected . as can be seen in figure [ fig - rlf ] ,
there does appear to be a deficit of the more luminous radio galaxies in a3558 and in the shapley supercluster core region in general .
this might have been expected based on the result of a lower fraction of intermediate optical magnitude radio galaxies , although that finding was irrespective of radio luminosity and was countered slightly by the increased fraction of optically bright radio galaxies .
the difficulty in assessing any deficit of strong radio galaxies relative to the lo96 rlf is that galaxies with such high luminosities are intrinsically rare , so demonstrating a statistically significant absence for a specific cluster is hampered by number statistics . in their analysis , v00 attempted to perform a more direct comparison with the lo96 rlf by correcting total galaxy counts to estimate the number of ellipticals and s0s in the a3558 environment .
they arrived at a net of 185 elliptical and s0 galaxies , of which 17 were identified radio galaxies .
integration of the lo96 rlf would predict 26 , and thus the deficit appears significant at about 97% confidence .
however , the lo96 rlf was constructed for ellipticals only and lo96 noted that only @xmath113 of their sample of 188 radio sources were s0 galaxies with the remainder being ellipticals . should just 24 of the 185 estimated counts used in the v00 analysis be s0 galaxies , the significance of the deficit drops under 90% .
it is instructive to examine this issue more closely . following the lo96 prescription ,
all galaxies within 0.6 mpc of the center of a3558 were selected ( i.e. , 0.3@xmath105 ) .
these were further culled to remove galaxies with @xmath106 and @xmath107 , for more direct comparison with the v00 analysis .
this produced a list of 43 galaxies , of which three had radio luminosities high enough for consideration .
a simple integration of the lo96 rlf would predict about 6 radio galaxies , and thus yields a deficit that would be claimed with 91% significance .
the morphologies of the 43 selected galaxies were assessed visually using the @xmath4 images along with ellipticities determined by sextractor . for most of these galaxies , ned also provided a morphology ( including confirmation of cluster membership via a published velocity ) and the assignments were consistent .
removal of non - ellipticals left only 24 galaxies , including the three radio galaxies .
this is almost exactly the fraction predicted by the lo96 rlf .
this analysis underscores the uncertainties involved in comparing luminosity functions constructed using small numbers of galaxies .
note that although the preceding discussion was based on photometry , the alternate procedure of v00 using spectroscopy is similarly effected : selecting galaxies on the basis of non - emission line spectra will include s0 galaxies as well as `` passive spirals '' @xcite and potentially even normal spirals whose fiber spectra sample just their bulge components . in summary , although there may be a reduction in the fraction of a3558 s elliptical galaxy population which host radio sources , it is questionable whether this reduction is significant .
g04 noted only marginal evidence for an increase in the radio - emitting faint galaxy population in a3562 , whereas in the current work it is reported as highly significant .
similarly , their analysis of the data presented in mo03 suggested that the radio galaxy fraction in a2255 was only marginally greater than the comparison clusters .
this difference in interpretation is caused primarily by the adopted statistical methods .
using the same magnitude ranges described above , g04 calculated the fraction of radio galaxies for each cluster .
they then calculated the fraction for the full comparison sample as the mean of these values , with the standard deviation representing the error .
the procedure of mo03 ( and used here ) treats the comparison sample as a whole .
in essence , it is an application of the binomial distribution where the comparison sample is used to determine the expected probability . using numbers for faint galaxies from table
[ tbl - faintfracs ] , 20 of 1104.6 comparison sample galaxies were radio detections so the statistical question is : what is the probability that finding 11 of 113.2 galaxies ( i.e. , the results for a3562 ) is consistent with this parent distribution ?
the @xmath86 nature of the actual test takes into account that the `` true fraction , '' i.e. , the 20 of 1104.6 from the comparison sample , is not known with certainty as would be assumed if one used the standard binomial probability function .
these differences in applied statistical tests are the reason that g04 ascribed only marginal significance to the excess activity in a2255 .
note that in the case of a3562 , the g04 numbers for radio detections and total galaxies do not produce a significant result regardless of the statistical method used .
presumably this is the result of differences in the areas surveyed .
although the radio data presented herein identify the region around sc1329 - 313 to have an excess of active galaxies , they only indirectly address the level of this activity .
the important question with respect to cluster evolutionary models , particularly those applied to the butcher - oemler effect , is whether the active galaxies are starbursts or more normal star - forming galaxies which have just entered the supercluster . in some studies ,
clusters of galaxies at moderate redshifts ( @xmath108 ) show evidence for starbursts ( e.g. , * ? ? ?
* ) while other studies suggest star formation is gradually extinguished as stripping removes the gaseous halos from galaxies , thereby removing the reservoir needed for future star formation ( strangulation , e.g. , * ? ? ?
unfortunately , the data presented herein do not allow for straightforward testing of such issues as the radio emission may include an agn component . in the case of the optically fainter galaxies ( @xmath109 )
, a good case can be made for starbursts on the basis of their relatively high radio luminosities and the fact that agn are rare in low mass galaxies .
for the brighter radio galaxies , ideally one would use optical spectroscopy in comparison with models ( as done in * ? ? ? * ) or data at other wavelengths such as the uv in order to quantify the past - to - present star formation and fraction of total stellar mass created in recent starbursts ( e.g. , * ? ? ?
* ; * ? ? ?
as alluded to in the case of the fainter galaxies , one possible available avenue is to rely on the strength of the radio emission to indicate the level of present star formation and the @xmath4 magnitude as a proxy for total stellar mass .
the ratio of these quantities would then be a measure of the relative strength of the current star formation .
the radio - to - optical flux ratio , @xmath110 , was calculated using the definition presented in @xcite : @xmath111 , where @xmath112 is the flux density at 6940@xmath113 determined using the @xmath4 magnitude via @xmath114 . to reduce contamination due to agn ,
galaxies brighter than @xmath115 were removed along with those redder than @xmath116 ( i.e. , remove probable red sequence objects down to the fainter magnitudes studied ) .
in addition a minimum 1.4 ghz flux of 0.5 mjy was required to remove potential bias caused by variation in the sensitivity of the radio data across the supercluster .
after removal of galaxies with published velocities placing them outside the supercluster , this resulted in a set of 83 galaxies for testing ( 48 of which did have published velocities indicating supercluster membership ) .
specific regions were compared via a wilcox test to determine whether any exhibited evidence for higher @xmath110 values , and hence more likely greater relative levels of star formation .
no highly significant variation in @xmath110 values across the core of the supercluster was found , although there is slight evidence that those in the vicinity of sc1329 - 313 do have higher @xmath110 .
this result was always less than @xmath117 , and slightly dependent on sub - samples used ( e.g. , how the regions were defined in ra and dec , whether or not galaxies without published velocities were included , what optical magnitude ranges were compared , etc . ) .
hence , this simple statistical analysis is suggestive of increased starburst activity associated with sc1329 - 313 , although this clearly needs to be confirmed through a more direct study .
this paper has presented a comprehensive radio study of the core of the shapley supercluster .
the vla has been used to map a nearly 7 square degree area through a mosaic of pointings which provide moderately uniform noise characteristics across the entire area . in conjunction with optical imaging , radio detections for 210 galaxies with @xmath5
were presented .
these include 104 galaxies with velocities placing them within the supercluster , which approximately doubles the previously known total of radio galaxies in the core of the shapley supercluster .
in addition , 2 radio galaxies are members of cataloged pairs whose companions have supercluster velocities and 8 radio galaxies are likely associated with the supercluster but are formally placed in the foreground ( 9000 km s@xmath2 @xmath118 10934 km s@xmath2 ) .
of those radio galaxies without velocity measurements , 68 have optical colors which do not rule out supercluster membership . across the entire core of the supercluster , intermediate optical magnitude ( @xmath83 )
galaxies appear less likely to host radio sources than their counterparts in a large comparison sample .
about 5% of such galaxies were found to be radio sources with @xmath49 w hz@xmath2 , whereas 15% of the comparison sample were .
this deficit is offset somewhat by the brighter galaxies ( @xmath8 ) of the shapley supercluster being more likely to be radio sources . while these results generally pertained to the entire region surveyed ,
a more dramatic effect was observed in the fainter galaxies ( @xmath85 ) localized around a3562 and sc1329 - 313 .
these galaxies were much more likely to be radio sources than anticipated on the basis of comparison clusters , with the high significance of the result being unaffected by changes in the assumed background .
on the basis of their blue colors and radio luminosities , these galaxies are presumably starbursts .
it is fascinating that this very region is identified by recent x - ray analysis as the location of an ongoing merger of sc1329 - 313 with a3562 and the rest of the supercluster .
the remainder of the supercluster does not exhibit any statistical excess of radio galaxies among its optically fainter population , underscoring the potential importance of cluster mergers in galaxy evolution .
this interpretation is consistent with radio studies of other active clusters , notably a2125 and a2255 .
examination of the galaxies with extended radio emission also revealed two interesting candidates for future observation .
j132802 - 314521 is a strong radio source with a luminous core coincident with the galaxy ( @xmath119 w hz@xmath2 ) .
this emission appears to connect to larger - scale diffuse emission , suggesting this source might be another head - tail radio galaxy within the supercluster .
extended radio emission around the galaxy j133048 - 314325 is a potential analog to the well - known system stephan s quintet . in this interpretation , j133051 - 314417
would have interacted with j133048 - 314325 and the nearby spiral j133047 - 314339 .
the resulting radio emission would be the combination of emission from the individual galaxies as well as shocks in the igm .
interestingly , this source is near the peak of the galaxy distribution centered on sc1329 - 313 , suggesting that environmental processes in groups involved with larger - scale mergers may be important for galaxy evolution .
the author thanks an anonymous referee , whose careful reading of the manuscript led to improvements in the clarity of the statistical analysis and additional insight into the extended radio galaxies .
much of this work was completed while i held a national research council associateship at nasa s goddard space flight center .
i also acknowledge the support of nasa through the american astronomical society s small research grant program , which supplied the computing resources to analyze the radio data .
most significantly , this project benefitted greatly from the assistance of michael ledlow .
mike s enthusiastic support of the project included invaluable comments on observing proposals and assistance with the data collection .
i am unable to express enough how sad his passing was , and how great a loss it is to his friends and colleagues .
venturi , t. , bardelli , s. , morganti , r. , & hunstead , r.w .
1998 , , 298 , 1113 venturi , t. , bardelli , s. , morganti , r. , & hunstead , r.w .
2000 , , 314 , 594 ( v00 ) venturi , t. , bardelli , s. , dallacasa , d. , brunetti , g. , giacintucci , s. , hunstead , r.w . , & morganti , r. 2003 , , 402 , 913 xu , c.k . , lu , n. , condon , j.j .
, dopita , m. , & tuffs , r.j .
2003 , , 595 , 665 r r r c c 1a & 13 28 48.0 & -30 41 00 & 23 @xmath59 11 & 06/18/2001 + 2b & 13 27 35.4 & -30 50 00 & 20 @xmath59 12 & 06/18/2001 + 2c & 13 30 00.6 & -30 50 00 & 26 @xmath59 10 & 06/20/2001 + 3a & 13 23 57.1 & -30 58 45 & 37 @xmath59 7 & 06/21/2001 + 3b & 13 26 22.5 & -30 58 58 & 29 @xmath59 9 & 06/22/2001 + 3c & 13 28 48.0 & -30 59 00 & 19 @xmath59 14 & 06/18/2001
+ 3d & 13 31 13.5 & -30 58 58 & 20 @xmath59 12 & 06/20/2001 + 4a & 13 22 43.8 & -31 07 34 & 28 @xmath59 9 & 06/18/2001 + 4b & 13 25 09.5 & -31 07 52 & 29 @xmath59 9 & 06/21/2001 + 4c & 13 27 35.2 & -31 08 00 & 17 @xmath59 14 & 06/18/2001 + 4d & 13 30 00.8 & -31 08 00 & 18 @xmath59 14 & 06/18/2001 + 4e & 13 32 26.5 & -31 07 52 & 20 @xmath59 13 & 06/20/2001 + & & & & 06/22/2001 + 4f & 13 34 52.2 & -31 07 34 & 19 @xmath59 13 & 06/22/2001 + 5a & 13 21 30.3 & -31 16 21 & 27 @xmath59 9 & 06/17/2001 + 5b & 13 23 56.1 & -31 16 45 & 24 @xmath59 10 & 06/17/2001 + 5c & 13 26 22.1 & -31 16 58 & 21 @xmath59 12 & 09/29/2002 + 5d & 13 28 48.0 & -31 17 00 & 17 @xmath59 15 & 06/18/2001 + 5e & 13 31 13.9 & -31 16 58 & 16 @xmath59 15 & 06/20/2001 + & & & & 06/22/2001 + 5f & 13 33 39.9 & -31 16 45 & 18 @xmath59 15 & 06/22/2001 + 5 g & 13 36 05.7 & -31 16 21 & 16 @xmath59 16 & 06/22/2001 + 6a & 13 22 42.6 & -31 25 34 & 23 @xmath59 11 & 06/17/2001 + 6b & 13 25 08.8 & -31 25 52 & 18 @xmath59 14 & 06/17/2001 + 6c & 13 27 34.9 & -31 26 00 & 17 @xmath59 16 & 06/17/2001 + & & & & 09/29/2002 + 6d & 13 30 01.1 & -31 26 00 & 19 @xmath59 13 & 06/18/2001 + 6e & 13 32 27.2 & -31 25 52 & 16 @xmath59 15 & 06/22/2001 + & & & & 09/29/2002 + 6f & 13 34 53.4 & -31 25 34 & 17 @xmath59 15 & 06/22/2001 + 7a & 13 21 28.8 & -31 34 21 & 20 @xmath59 13 & 06/17/2001 + 7b & 13 23 55.2 & -31 34 45 & 17 @xmath59 14 & 06/17/2001 + 7c & 13 26 21.6 & -31 34 58 & 16 @xmath59 16 & 09/29/2002 + 7d & 13 28 48.0 & -31 35 00 & 16 @xmath59 16 & 06/18/2001 + 7e & 13 31 14.4 & -31 34 58 & 18 @xmath59 14 & 06/22/2001 + & & & & 09/29/2002 + 7f & 13 33 40.8 & -31 34 45 & 16 @xmath59 15 & 06/20/2001 + & & & & 09/29/2002 + 7 g & 13 36 07.2 & -31 34 21 & 17 @xmath59 15 & 06/22/2001 + 8a & 13 20 14.8 & -31 43 05 & 18 @xmath59 14 & 06/17/2001 + 8b & 13 22 41.4 & -31 43 34 & 17 @xmath59 15 & 06/17/2001 + 8c & 13 25 08.1 & -31 43 52 & 17 @xmath59 15 & 09/29/2002 + 8d & 13 27 34.7 & -31 44 00 & 18 @xmath59 14 & 09/29/2002 + 8e & 13 30 01.3 & -31 44 00 & 17 @xmath59 15 & 09/29/2002 + 8f & 13 32 28.0 & -31 43 52 & 17 @xmath59 15 & 06/20/2001 + & & & & 09/29/2002 + 8 g & 13 34 54.6 & -31 43 34 & 18 @xmath59 14 & 06/22/2001 + 9a & 13 21 27.4 & -31 52 21 & 17 @xmath59 15 & 06/17/2001 + 9b & 13 23 54.2 & -31 52 45 & 18 @xmath59 14 & 09/29/2002 + 9c & 13 26 21.1 & -31 52 58 & 19 @xmath59 13 & 09/29/2002 + 9d & 13 28 48.0 & -31 53 00 & 21 @xmath59 12 & 06/18/2001 + 9e & 13 31 14.9 & -31 52 58 & 25 @xmath59 10 & 06/22/2001 + 9f & 13 33 41.8 & -31 52 45 & 20 @xmath59 13 & 09/29/2002 + 9 g & 13 36 08.6 & -31 52 21 & 20 @xmath59 13 & 06/22/2001 + 10a & 13 22 40.2 & -32 01 34 & 16 @xmath59 16 & 06/17/2001 + 10b & 13 25 07.3 & -32 01 52 & 25 @xmath59 10 & 09/29/2002 + 10c & 13 27 34.4 & -32 02 00 & 17
@xmath59 15 & 06/18/2001 + 10d & 13 30 01.6 & -32 02 00 & 24 @xmath59 11 & 06/18/2001 + 10e & 13 32 28.7 & -32 01 52 & 25 @xmath59 10 & 06/22/2001 + 10f & 13 34 55.8 & -32 01 34 & 22 @xmath59 12 & 06/22/2001 + 11a & 13 21 25.9 & -32 10 21 & 20 @xmath59 13 & 06/17/2001 + 11b & 13 23 53.3 & -32 10 45 & 29 @xmath59 9 & 06/17/2001 + 11c & 13 26 20.6 & -32 10 58 & 30 @xmath59 9 & 09/29/2002 + 11d & 13 28 48.0 & -32 11 00 & 28 @xmath59 9 & 06/18/2001 + 11e & 13 31 15.4 & -32 10 58 & 35 @xmath59 7 & 09/29/2002 + 11f & 13 33 42.7 & -32 10 45 & 33 @xmath59 8 & 06/20/2001 + 12a & 13 22 39.0 & -32 19 34 & 22 @xmath59 12 & 06/17/2001 + 12d & 13 30 01.8 & -32 20 00 & 34 @xmath59 8 & 06/18/2001 + 12e & 13 32 29.4 & -32 19 52 & 21 @xmath59 12 & 06/20/2001 + 12f & 13 34 57.0 & -32 19 34 & 26 @xmath59 10 & 06/21/2001 + r r
r r r r r c r r r c 13 19 03.14 & -31 34 42.9 & [email protected] & [email protected] & [email protected] & 1.50 & 0.61 & r & 0.20 & 4.8 & & a + 13 19 07.85 & -31 40 10.8 & [email protected] & [email protected] & [email protected] & 0.94 & 0.67 & u & 0.14 & 5.2 & 13968 & + 13 19 15.67 & -31 51 18.8 & [email protected] & [email protected] & [email protected] & 5.14 & 3.40 & r & 0.15 & 4.5 & & a + 13 19 35.06 & -31 31 39.4 & [email protected] & [email protected] & [email protected] & 1.76 & 1.73 & r & 0.13 & 2.3 & & a + 13 19 35.35 & -31 40 21.3 & [email protected] & [email protected] & [email protected] & 0.63 & 0.70 & u & 0.09 & 2.9 & & + 13 20 23.40 & -31 26 31.8 & [email protected] & [email protected] & [email protected] & 1.01 & 0.45 & r & 0.11 & 4.3 & & a + 13 20 24.47 & -32 02 48.0 & [email protected] & [email protected] & [email protected] & 1.00 & 0.70 & r & 0.12 & 1.1 & & a + 13 20 27.89 & -31 36 07.1 & [email protected] & [email protected] & [email protected] & 0.48 & 0.36 & r & 0.09 & 2.7 & & + 13 20 54.63 & -31 44 33.4 & [email protected] & [email protected] & [email protected] & 7.94 & 7.12 & r & 0.07 & 1.2 & 16249 & + 13 20 55.59 & -31 43 59.3 & [email protected] & [email protected] & [email protected] & 6.65 & 5.12 & r & 0.07 & 2.1 & 15829 & + 13 20 55.64 & -31 27 24.6 & [email protected] & [email protected] & [email protected] & 0.84 & 0.75 & r & 0.09 & 0.8 & & + 13 21 12.30 & -31 08 14.9 & [email protected] & [email protected] & [email protected] & 1.33 & 1.58 & u & 0.12 & 1.1 & & + 13 21 45.02 & -31 03 00.6 & [email protected] & [email protected] & [email protected] & 7.80 & 4.93 & r & 0.12 & 1.1 & 15071 & b , c + 13 21 52.65 & -31 45 54.4 & [email protected] & [email protected] & [email protected] & 1.44 & 1.03 & r & 0.07 & 3.5 & 14705 & + 13 21 58.97 & -31 38 37.8 & [email protected] & [email protected] & [email protected] & 0.45 & 0.45 & u & 0.07 & 2.5 & & + 13 22 01.07 & -31 26 44.6 & [email protected] & [email protected] & [email protected] & 0.28 & 0.40 & u & 0.08 & 2.2 & & + 13 22 03.25 & -31 02 24.7 & [email protected] & [email protected] & [email protected] & 0.77 & 0.58 & r & 0.11 & 4.7 & & + 13 22 06.75 & -31 46 16.4 & [email protected] & [email protected] & [email protected] & 7.34 & 2.29 & r & 0.07 & & 14229 & b , c + 13 22 10.13 & -31 40 52.1 & [email protected] & [email protected] & [email protected] & 1.85 & 0.96 & r & 0.07 & 3.0 & & + 13 22 10.92 & -31 28 05.1 & [email protected] & [email protected] & [email protected] & 98.87 & 22.00 & r & 0.07 & & & a , b , c + 13 22 19.65 & -31 22 05.7 & [email protected] & [email protected] & [email protected] & 0.55 & 0.72 & u & 0.07 & 3.0 & & + 13 22 22.85 & -31 37 43.6 & [email protected] & [email protected] & [email protected] & 1.28 & 1.09 & r & 0.07 & 0.2 & & + 13 22 23.73 & -31 04 32.7 & [email protected] & [email protected] & [email protected] & 2.38 & 0.91 & r & 0.09 & 1.7 & 4512 & + 13 22 39.55 & -32 15 20.1 & [email protected] & [email protected] & [email protected] & 1.11 & 1.06 & u & 0.08 & 2.9 & & a + 13 22 41.29 & -32 18 50.6 & [email protected] & [email protected] & [email protected] & 0.50 & 0.40 & r & 0.09 & 3.0 & & a + 13 22 42.81 & -32 23 42.2 & [email protected] & [email protected] & [email protected] & 1.54 & 0.84 & r & 0.11 & 2.7 & & + 13 22 49.71 & -32 26 11.6 & [email protected] & [email protected] & [email protected] & 2.92 & 3.18 & u & 0.12 & 3.7 & 13731 & + 13 22 52.61 & -32 30 05.6 & [email protected] & [email protected] & [email protected] & 1.11 & 0.85 & r & 0.15 & 1.1 & & a + 13 23 00.09 & -31 23 13.9 & [email protected] & [email protected] & [email protected] & 1.17 & 0.84 & r & 0.07 & 1.4 & & + 13 23 03.86 & -31 23 59.1 & [email protected] & [email protected] & [email protected] & 2.61 & 1.27 & r & 0.06 & 2.4 & & + 13 23 15.11 & -31 27 36.0 & [email protected] & [email protected] & [email protected] & 5.18 & 4.02 & r & 0.07 & 1.4 & 9068 & + 13 23 29.07 & -31 50 39.6 & [email protected] & [email protected] & [email protected] & 4.25 & 3.85 & r & 0.08 & 1.7 & 14306 & + 13 23 33.56 & -32 16 03.8 & [email protected] & [email protected] & [email protected] & 0.47 & 0.55 & r & 0.08 & 4.7 & & + 13 23 54.22 & -32 10 56.3 & [email protected] & [email protected] & [email protected] & 5.76 & 3.05 & r & 0.08 & 1.3 & 15295 & + 13 23 57.58 & -31 38 45.2 & [email protected] & [email protected] & [email protected] & 55.54 & 2.43 & r & 0.08 & 3.2 & 15054 & b , c + 13 24 02.09 & -31 19 13.8 & [email protected] & [email protected] & [email protected] & 0.53 & 0.50 & r & 0.07 & 1.6 & & + 13 24 05.04 & -31 04 19.6 & [email protected] & [email protected] & [email protected] & 1.11 & 1.26 & u & 0.08 & 4.2 & & a + 13 24 06.68 & -31 40 12.0 & [email protected] & [email protected] & [email protected] & 10.39 & 9.06 & r & 0.08 & 0.9 & 14397 & + 13 24 16.30 & -32 12 54.6 & [email protected] & [email protected] & [email protected] & 0.44 & 0.56 & r & 0.09 & 4.6 & & + 13 24 16.88 & -31 10 09.9 & [email protected] & [email protected] & [email protected] & 2.04 & 1.66 & r & 0.08 & 2.5 & 10057 & + 13 24 22.77 & -31 42 23.7 & [email protected] & [email protected] & [email protected] & 13.90 & 9.83 & r & 0.09 & 0.7 & 14370 & + 13 24 32.46 & -31 06 58.2 & [email protected] & [email protected] & [email protected] & 2.50 & 2.07 & r & 0.08 & 1.0 & & + 13 24 33.93 & -31 40 14.6 & [email protected] & [email protected] & [email protected] & 0.43 & 0.50 & u & 0.09 & 3.0 & & + 13 24 37.96 & -31 52 39.0 & [email protected] & [email protected] & [email protected] & 0.73 & 0.55 & r & 0.10 & 1.7 & & a + 13 24 39.14 & -32 06 33.4 & [email protected] & [email protected] & [email protected] & 0.63 & 0.59 & u & 0.09 & 2.4 & 13650 & + 13 24 42.77 & -32 05 48.2 & [email protected] & [email protected] & [email protected] & 1.63 & 0.85 & r & 0.09 & 1.2 & & + 13 24 47.57 & -31 20 29.9 & [email protected] & [email protected] & [email protected] & 1.66 & 0.90 & r & 0.07 & 2.9 & 9626 & + 13 24 50.44 & -31 33 37.8 & [email protected] & [email protected] & [email protected] & 1.16 & 0.80 & r & 0.08 & 2.4 & 14773 & + 13 24 57.54 & -30 51 01.1 & [email protected] & [email protected] & [email protected] & 0.58 & 0.70 & r & 0.13 & 3.5 & & + 13 25 04.19 & -32 06 57.7 & [email protected] & [email protected] & [email protected] & 0.94 & 0.64 & r & 0.09 & 3.0 & & + 13 25 07.47 & -32 05 19.0 & [email protected] & [email protected] & [email protected] & 0.37 & 0.56 & u & 0.09 & 2.9 & & + 13 25 08.77 & -31 00 35.5 & [email protected] & [email protected] & [email protected] & 1.02 & 0.46 & r & 0.09 & 4.2 & & b + 13 25 15.31 & -30 55 07.3 & [email protected] & [email protected] & [email protected] & 0.62 & 0.78 & r & 0.11 & 1.8 & & + 13 25 29.81 & -31 15 59.8 & [email protected] & [email protected] & [email protected] & 0.46 & 0.40 & u & 0.08 & 1.7 & & + 13 25 42.26 & -32 03 44.0 & [email protected] & [email protected] & [email protected] & 0.44 & 0.53 & r & 0.09 & 2.8 & & + 13 25 50.32 & -32 08 54.2 & [email protected] & [email protected] & [email protected] & 0.62 & 0.64 & r & 0.09 & 4.1 & 13853 & + 13 26 17.31 & -31 38 46.0 & [email protected] & [email protected] & [email protected] & 5.93 & 5.24 & r & 0.08 & 1.4 & 39812 & + 13 26 18.81 & -30 57 37.0 & [email protected] & [email protected] & [email protected] & 5.31 & 4.60 & r & 0.09 & 2.5 & 14212 & + 13 26 25.80 & -31 27 20.9 & [email protected] & [email protected] & [email protected] & 0.68 & 0.77 & u & 0.08 & 3.3 & 13749 & + 13 26 28.67 & -30 44 44.4 & [email protected] & [email protected] & [email protected] & 4.62 & 2.57 & r & 0.14 & 3.8 & & b + 13 26 35.69 & -30 59 36.4 & [email protected] & [email protected] & [email protected] & 0.54 & 0.46 & r & 0.08 & 3.8 & 14273 & + 13 26 44.47 & -32 01 50.9 & [email protected] & [email protected] & [email protected] & 0.33 & 0.42 & u & 0.08 & 5.0 & & + 13 26 48.60 & -31 34 18.6 & [email protected] & [email protected] & [email protected] & 2.15 & 1.34 & r & 0.08 & 2.7 & 14657 & + 13 26 49.79 & -31 23 44.5 & [email protected] & [email protected] & [email protected] & 9.31 & 6.37 & r & 0.09 & 1.9 & 15190 & + 13 26 49.94 & -30 37 56.9 & [email protected] & [email protected] & [email protected] & 1.91 & 0.64 & r & 0.24 & 3.5 & 14555 & + 13 26 52.95 & -32 00 28.6 & [email protected] & [email protected] & [email protected] & 0.26 & 0.40 & u & 0.08 & 3.7 & 14389 & + 13 26 53.88 & -31 18 55.9 & [email protected] & [email protected] & [email protected] & 1.40 & 1.08 & r & 0.10 & 1.1 & & d + 13 26 54.01 & -31 18 39.5 & [email protected] & [email protected] & [email protected] & 0.46 & 0.50 & r & 0.10 & 2.6 & 14512 & d + 13 26 57.51 & -32 11 52.3 & [email protected] & [email protected] & [email protected] & 0.78 & 0.55 & r & 0.10 & 4.6 & 13823 & + 13 26 58.44 & -30 45 10.7 & [email protected] & [email protected] & [email protected] & 0.55 & 0.80 & u & 0.12 & 1.2 & & + 13 27 03.35 & -31 13 31.5 & [email protected] & [email protected] & [email protected] & 1.13 & 0.51 & r & 0.08 & 2.7 & 16387 & + 13 27 04.06 & -31 32 16.3 & [email protected] & [email protected] & [email protected] & 0.43 & 0.39 & u & 0.08 & 3.9 & & + 13 27 29.66 & -31 23 24.3 & [email protected] & [email protected] & [email protected] & 3.47 & 2.93 & r & 0.09 & 0.6 & 14281 & + 13 27 29.66 & -30 33 31.8 & [email protected] & [email protected] & [email protected] & 4.22 & 4.46 & r & 0.24 & 1.2 & 14150 & + 13 27 34.96 & -31 32 59.2 & [email protected] & [email protected] & [email protected] & 1.06 & 0.79 & u & 0.09 & 1.9 & 12815 & + 13 27 45.50 & -31 47 51.0 & [email protected] & [email protected] & [email protected] & 0.44 & 0.48 & u & 0.08 & 1.6 & 13186 & + 13 27 51.48 & -31 14 45.1 & [email protected] & [email protected] & [email protected] & 1.56 & 1.29 & r & 0.08 & 2.7 & 14151 & + 13 27 54.95 & -31 32 18.6 & [email protected] & [email protected] & [email protected] & 9.81 & 9.26 & r & 0.10 & 0.6 & 15481 & + 13 27 56.84 & -31 29 44.6 & [email protected] & [email protected] & [email protected] & 5.51 & 4.60 & r & 0.10 & 1.7 & 14061 & + 13 28 02.63 & -31 45 21.3 & [email protected] & [email protected] & [email protected] & 32.37 & 21.98 & r & 0.08 & 1.3 & 12857 & b , c , e + 13 28 03.55 & -30 40 50.9 & [email protected] & [email protected] & [email protected] & 1.46 & 0.85 & r & 0.12 & 4.6 & 14996 & + 13 28 08.02 & -32 08 46.1 & [email protected] & [email protected] & [email protected] & 2.38 & 1.58 & r & 0.08 & 3.5 & 21712 & + 13 28 10.37 & -31 20 57.4 & [email protected] & [email protected] & [email protected] & 0.92 & 0.50 & r & 0.09 & 3.4 & 15690 & + 13 28 10.85 & -32 12 51.6 & [email protected] & [email protected] & [email protected] & 0.62 & 0.36 & r & 0.10 & 6.8 & 43169 & + 13 28 25.02 & -31 51 28.3 & [email protected] & [email protected] & [email protected] & 2.68 & 1.13 & r & 0.09 & 2.1 & 4389 & b + 13 28 25.75 & -30 38 08.2 & [email protected] & [email protected] & [email protected] & 1.20 & 0.42 & r & 0.11 & 3.6 & & f + 13 28 29.84 & -32 14 35.6 & [email protected] & [email protected] & [email protected] & 2.69 & 2.36 & r & 0.09 & 0.5 & & + 13 28 30.11 & -31 47 15.7 & [email protected] & [email protected] & [email protected] & 0.52 & 0.45 & u & 0.09 & 4.5 & 9971 & + 13 28 30.27 & -32 04 42.5 & [email protected] & [email protected] & [email protected] & 0.24 & 0.40 & u & 0.08 & 2.7 & & a + 13 28 33.79 & -31 26 28.3 & [email protected] & [email protected] & [email protected] & 0.86 & 0.53 & r & 0.10 & 2.8 & 14522 & + 13 28 34.83 & -31 44 44.4 & [email protected] & [email protected] & [email protected] & 0.62 & 0.43 & r & 0.10 & 0.9 & 9920 & + 13 28 35.78 & -31 30 54.3 & [email protected] & [email protected] & [email protected] & 0.65 & 0.60 & r & 0.11 & 2.4 & 14003 & + 13 28 38.47 & -32 10 36.7 & [email protected] & [email protected] & [email protected] & 0.65 & 0.42 & u & 0.08 & 4.0 & & + 13 28 38.72 & -31 20 49.3 & [email protected] & [email protected] & [email protected] & 0.61 & 0.41 & r & 0.10 & 0.9 & 15625 & + 13 28 39.26 & -31 39 41.1 & [email protected] & [email protected] & [email protected] & 0.83 & 0.96 & u & 0.11 & 2.7 & 13314 & + 13 28 50.26 & -30 47 23.1 & [email protected] & [email protected] & [email protected] & 0.67 & 0.52 & r & 0.08 & 2.3 & & + 13 28 50.91 & -32 07 53.1 & [email protected] & [email protected] & [email protected] & 2.00 & 1.56 & r & 0.08 & 2.4 & 15073 & + 13 28 52.29 & -32 09 45.8 & [email protected] & [email protected] & [email protected] & 7.29 & 6.05 & r & 0.08 & 1.8 & 12441 & + 13 28 54.30 & -31 55 11.1 & [email protected] & [email protected] & [email protected] & 0.77 & 0.70 & u & 0.08 & 1.9 & 11937 & + 13 28 57.71 & -31 51 34.9 & [email protected] & [email protected] & [email protected] & 1.60 & 1.78 & u & 0.09 & 1.2 & 16532 & g + 13 28 59.29 & -31 44 15.7 & [email protected] & [email protected] & [email protected] & 1.58 & 1.20 & r & 0.10 & 2.7 & 13550 & + 13 29 04.33 & -31 47 52.4 & [email protected] & [email protected] & [email protected] & 0.69 & 0.67 & r & 0.09 & 0.8 & 15411 & + 13 29 12.29 & -31 18 39.9 & [email protected] & [email protected] & [email protected] & 0.55 & 0.35 & r & 0.10 & 1.0 & 13610 & + 13 29 14.98 & -31 16 44.9 & [email protected] & [email protected] & [email protected] & 1.07 & 0.86 & r & 0.10 & 3.5 & 13573 & b + 13 29 17.00 & -32 02 33.9 & [email protected] & [email protected] & [email protected] & 1.13 & 0.85 & r & 0.08 & 2.7 & 13834 & + 13 29 20.43 & -30 46 37.4 & [email protected] & [email protected] & [email protected] & 1.62 & 1.51 & r & 0.09 & 1.2 & 14398 & + 13 29 28.57 & -31 39 25.7 & [email protected] & [email protected] & [email protected] & 2.30 & 1.81 & r & 0.12 & 1.5 & 14413 & + 13 29 31.44 & -31 37 29.0 & [email protected] & [email protected] & [email protected] & 0.88 & 0.70 & r & 0.11 & 3.1 & 15135 & + 13 29 34.78 & -30 32 43.6 & [email protected] & [email protected] & [email protected] & 1.58 & 1.27 & r & 0.14 & 3.7 & 4575 & + 13 29 37.24 & -30 57 56.5 & [email protected] & [email protected] & [email protected] & 0.83 & 0.49 & r & 0.10 & 2.2 & 10428 & + 13 29 43.22 & -30 28 22.9 & [email protected] & [email protected] & [email protected] & 1.31 & 0.87 & r & 0.25 & 5.5 & & + 13 29 50.90 & -31 22 58.2 & [email protected] & [email protected] & [email protected] & 16.72 & 6.12 & r & 0.11 & 0.5 & 58775 & b , c + 13 29 54.12 & -31 08 43.9 & [email protected] & [email protected] & [email protected] & 0.56 & 0.33 & u & 0.10 & 0.5 & & + 13 29 54.20 & -32 00 59.3 & [email protected] & [email protected] & [email protected] & 1.03 & 0.78 & r & 0.08 & 0.9 & 14012 & + 13 29 55.10 & -31 19 56.4 & [email protected] & [email protected] & [email protected] & 3.87 & 1.54 & r & 0.11 & 2.9 & 14815 & + 13 29 57.07 & -31 11 08.1 & [email protected] & [email protected] & [email protected] & 1.97 & 1.84 & r & 0.10 & 0.9 & 16628 & + 13 29 59.97 & -32 25 59.2 & [email protected] & [email protected] & [email protected] & 0.63 & 1.57 & r & 0.11 & 4.2 & & a + 13 30 05.14 & -31 24 36.8 & [email protected] & [email protected] & [email protected] & 8.40 & 2.67 & r & 0.12 & 3.7 & 13780 & + 13 30 05.20 & -31 13 53.4 & [email protected] & [email protected] & [email protected] & 0.95 & 0.84 & r & 0.11 & 1.2 & 15943 & + 13 30 22.06 & -30 43 52.1 & [email protected] & [email protected] & [email protected] & 1.71 & 0.94 & r & 0.13 & 2.3 & & + 13 30 24.83 & -30 41 37.1 & [email protected] & [email protected] & [email protected] & 1.33 & 1.24 & r & 0.14 & 2.8 & 14491 & + 13 30 32.80 & -31 34 33.8 & [email protected] & [email protected] & [email protected] & 1.48 & 0.63 & r & 0.12 & 0.5 & 14478 & b + 13 30 35.76 & -30 41 07.8 & [email protected] & [email protected] & [email protected] & 1.25 & 1.18 & r & 0.16 & 6.3 & & + 13 30 39.21 & -30 48 08.7 & [email protected] & [email protected] & [email protected] & 0.59 & 0.53 & r & 0.11 & 5.6 & & + 13 30 40.65 & -31 42 32.1 & [email protected] & [email protected] & [email protected] & 0.53 & 0.34 & r & 0.09 & 1.9 & 15489 & + 13 30 40.74 & -31 51 56.0 & [email protected] & [email protected] & [email protected] & 0.43 & 0.35 & r & 0.08 & 3.5 & 14407 & + 13 30 41.47 & -31 34 17.3 & [email protected] & [email protected] & [email protected] & 1.56 & 1.33 & r & 0.10 & 6.1 & 15323 & d + 13 30 42.03 & -31 34 28.7 & [email protected] & [email protected] & [email protected] & 1.69 & 0.97 & r & 0.10 & 3.0 & 15215 & d , h + 13 30 46.20 & -32 18 42.7 & [email protected] & [email protected] & [email protected] & 0.93 & 0.68 & r & 0.10 & 1.8 & & + 13 30 47.73 & -31 51 06.5 & [email protected] & [email protected] & [email protected] & 0.55 & 0.40 & r & 0.08 & 1.7 & 14636 & + 13 30 48.93 & -31 43 25.7 & [email protected] & [email protected] & [email protected] & 7.53 & 1.42 & r & 0.09 & 1.3 & 13226 & b , c , i + 13 30 51.04 & -31 44 17.9 & [email protected] & [email protected] & [email protected] & 0.67 & 0.57 & u & 0.09 & 3.4 & 13815 & + 13 30 51.82 & -31 04 40.3 & [email protected] & [email protected] & [email protected] & 3.78 & 2.79 & r & 0.10 & 1.6 & & + 13 30 52.20 & -31 46 07.0 & [email protected] & [email protected] & [email protected] & 1.68 & 1.68 & r & 0.08 & 1.1 & 13832 & + 13 30 52.70 & -31 27 13.2 & [email protected] & [email protected] & [email protected] & 0.78 & 0.95 & u & 0.11 & 1.8 & 14612 & + 13 30 53.83 & -31 30 21.7 & [email protected] & [email protected] & [email protected] & 0.50 & 0.40 & r & 0.10 & 3.5 & 14409 & + 13 30 56.16 & -31 33 42.9 & [email protected] & [email protected] & [email protected] & 0.56 & 0.39 & r & 0.09 & 3.9 & 14348 & + 13 31 00.75 & -31 44 53.9 & [email protected] & [email protected] & [email protected] & 2.36 & 2.03 & r & 0.08 & 0.4 & 11976 & + 13 31 01.99 & -31 40 36.3 & [email protected] & [email protected] & [email protected] & 0.76 & 0.72 & u & 0.09 & 1.9 & 13180 & + 13 31 16.90 & -31 35 38.4 & [email protected] & [email protected] & [email protected] & 1.96 & 0.86 & r & 0.09 & 0.6 & 14578 & b + 13 31 17.43 & -31 24 34.7 & [email protected] & [email protected] & [email protected] & 1.40 & 1.01 & r & 0.09 & 4.3 & 13763 & + 13 31 19.47 & -31 13 44.4 & [email protected] & [email protected] & [email protected] & 3.79 & 1.99 & r & 0.09 & 0.7 & 15710 & b + 13 31 20.22 & -31 44 34.6 & [email protected] & [email protected] & [email protected] & 2.37 & 1.16 & r & 0.08 & 3.4 & 13333 & + 13 31 22.42 & -31 22 10.4 & [email protected] & [email protected] & [email protected] & 0.52 & 0.43 & r & 0.09 & 2.1 & 14027 & + 13 31 27.58 & -31 49 14.7 & [email protected] & [email protected] & [email protected] & 1.64 & 1.59 & u & 0.08 & 1.9 & 12928 & + 13 31 37.79 & -31 33 12.0 & [email protected] & [email protected] & [email protected] & 2.92 & 2.28 & r & 0.08 & 0.8 & 14194 & j + 13 31 43.87 & -30 55 44.3 & [email protected] & [email protected] & [email protected] & 1.25 & 0.73 & r & 0.12 & 3.5 & & f + 13 31 53.37 & -31 55 36.1 & [email protected] & [email protected] & [email protected] & 0.58 & 0.72 & u & 0.07 & 0.8 & & + 13 31 54.87 & -31 54 04.1 & [email protected] & [email protected] & [email protected] & 1.94 & 1.13 & r & 0.07 & 1.9 & 13043 & + 13 31 59.81 & -31 49 22.2 & [email protected] & [email protected] & [email protected] & 0.89 & 0.89 & u & 0.07 & 2.7 & 12963 & + 13 32 00.41 & -32 24 01.8 & [email protected] & [email protected] & [email protected] & 1.20 & 0.99 & r & 0.12 & 5.8 & & k + 13 32 01.52 & -31 13 58.0 & [email protected] & [email protected] & [email protected] & 1.68 & 1.77 & u & 0.08 & 2.4 & & + 13 32 03.17 & -31 46 48.6 & [email protected] & [email protected] & [email protected] & 4.30 & 3.74 & r & 0.08 & 1.0 & 13107 & + 13 32 12.62 & -31 05 42.8 & [email protected] & [email protected] & [email protected] & 0.52 & 0.47 & r & 0.07 & 0.8 & & + 13 32 15.15 & -31 13 29.8 & [email protected] & [email protected] & [email protected] & 0.76 & 0.49 & u & 0.07 & 4.0 & 44703 & + 13 32 20.74 & -32 07 37.6 & [email protected] & [email protected] & [email protected] & 1.27 & 1.05 & r & 0.09 & 1.5 & & a + 13 32 30.07 & -31 03 15.8 & [email protected] & [email protected] & [email protected] & 3.07 & 2.47 & r & 0.08 & 0.6 & 13772 & + 13 32 35.44 & -31 31 45.5 & [email protected] & [email protected] & [email protected] & 1.80 & 1.20 & r & 0.07 & 3.3 & 15193 & + 13 32 36.13 & -32 20 01.3 & [email protected] & [email protected] & [email protected] & 2.22 & 1.61 & r & 0.10 & 1.6 & & a + 13 32 38.27 & -31 01 41.3 & [email protected] & [email protected] & [email protected] & 1.19 & 1.34 & u & 0.09 & 1.1 & 13189 & + 13 32 42.95 & -31 55 13.7 & [email protected] & [email protected] & [email protected] & 0.42 & 0.33 & u & 0.08 & 4.1 & 14483 & + 13 32 43.89 & -31 47 01.3 & [email protected] & [email protected] & [email protected] & 0.76 & 0.68 & u & 0.07 & 1.0 & 12984 & + 13 32 46.48 & -31 37 03.3 & [email protected] & [email protected] & [email protected] & 1.83 & 1.37 & r & 0.08 & 2.2 & 12315 & + 13 32 46.75 & -31 57 59.2 & [email protected] & [email protected] & [email protected] & 2.22 & 1.76 & r & 0.08 & 1.5 & 11106 & + 13 33 00.85 & -30 53 07.4 & [email protected] & [email protected] & [email protected] & 2.05 & 1.35 & r & 0.16 & 2.8 & 10561 & + 13 33 01.00 & -31 13 29.3 & [email protected] & [email protected] & [email protected] & 1.22 & 1.18 & r & 0.07 & 1.2 & 10981 & + 13 33 01.33 & -31 44 19.2 & [email protected] & [email protected] & [email protected] & 1.69 & 1.63 & r & 0.07 & 2.2 & & + 13 33 04.16 & -31 39 04.0 & [email protected] & [email protected] & [email protected] & 2.06 & 1.55 & r & 0.07 & 0.1 & 14120 & + 13 33 14.20 & -32 12 08.2 & [email protected] & [email protected] & [email protected] & 4.55 & 1.96 & r & 0.08 & 2.6 & & a + 13 33 17.99 & -31 39 11.1 &
[email protected] & [email protected] & [email protected] & 0.93 & 0.77 & r & 0.08 & 2.7 & & h , l + 13 33 23.53 & -31 24 23.3 & [email protected] & [email protected] & [email protected] & 0.25 & 0.39 & u & 0.07 & 1.8 & 13103 & + 13 33 27.77 & -31 02 16.2 & [email protected] & [email protected] & [email protected] & 0.66 & 0.77 & u & 0.11 & 2.0 & & a , m + 13 33 28.21 & -32 01 07.1 & [email protected] & [email protected] & [email protected] & 0.85 & 0.91 & r & 0.07 & 0.5 & & + 13 33 31.63 & -31 40 58.6 & [email protected] & [email protected] & [email protected] & 103.34 & 35.58 & r & 0.09 & 9.6 & 14438 & b , c + 13 33 34.76 & -31 40 20.5 & [email protected] & [email protected] & [email protected] & 0.51 & 0.40 & u & 0.09 & 0.7 & 14653 & + 13 33 40.7 & -31 19 57 . &
[email protected] & [email protected] & [email protected] & 6.25 & 4.82 & r & 0.08 & 1.0 & & n + 13 33 40.94 & -31 54 23.4 & [email protected] & [email protected] & [email protected] & 0.71 & 0.62 & r & 0.07 & 3.0 & & + 13 33 50.43 & -31 28 38.0 & [email protected] & [email protected] & [email protected] & 0.67 & 0.61 & u & 0.08 & 3.0 & 10994 & + 13 33 57.54 & -31 29 05.5 & [email protected] & [email protected] & [email protected] & 1.17 & 0.78 & r & 0.08 & 2.2 & & o + 13 33 58.73 & -31 58 16.1 & [email protected] & [email protected] & [email protected] & 1.29 & 1.20 & u & 0.07 & 0.4 & 14689 & + 13 34 00.12 & -32 22 56.1 & [email protected] & [email protected] & [email protected] & 1.36 & 0.91 & r & 0.11 & 3.7 & & + 13 34 06.6 & -31 43 39 . & [email protected] & [email protected] & [email protected] & 1.05 & 0.86 & r & 0.08 & 2.6 & & n + 13 34 08.12 & -31 16 55.6 & [email protected] & [email protected] & [email protected] & 0.90 & 0.45 & r & 0.07 & 1.8 & & + 13 34 18.21 & -31 49 08.1 & [email protected] & [email protected] & [email protected] & 0.71 & 0.72 & r & 0.07 & 3.0 & & + 13 34 19.43 & -31 59 14.4 & [email protected] & [email protected] & [email protected] & 0.28 & 0.40 & u & 0.07 & 2.3 & 10908 & + 13 34 29.81 & -31 36 55.2 & [email protected] & [email protected] & [email protected] & 0.40 & 0.34 & u & 0.06 & 2.6 & & + 13 34 31.48 & -32 05 09.0 & [email protected] & [email protected] & [email protected] & 1.00 & 0.59 & r & 0.07 & 2.3 & 12647 & + 13 34 36.19 & -31 41 06.7 & [email protected] & [email protected] & [email protected] & 1.47 & 1.23 & r & 0.07 & 5.4 & 11407 & + 13 34 39.94 & -31 32 54.5 & [email protected] & [email protected] & [email protected] & 5.37 & 3.99 & r & 0.06 & 6.4 & 16490 & + 13 34 40.45 & -31 29 43.2 & [email protected] & [email protected] & [email protected] & 0.66 & 0.61 & r & 0.06 & 3.8 & & + 13 34 40.80 & -31 26 43.6 & [email protected] & [email protected] & [email protected] & 1.57 & 1.55 & r & 0.06 & 1.0 & 14156 & + 13 34 43.00 & -32 22 14.0 & [email protected] & [email protected] & [email protected] & 1.12 & 0.67 & r & 0.09 & 1.5 & 12322 & + 13 34 43.81 & -31 45 52.2 & [email protected] & [email protected] & [email protected] & 0.29 & 0.41 & u & 0.08 & 1.9 & & + 13 34 57.25 & -32 26 31.6 & [email protected] & [email protected] & [email protected] & 0.78 & 0.85 & r & 0.11 & 0.6 & & + 13 35 03.06 & -31 39 18.6 & [email protected] & [email protected] & [email protected] & 15.49 & 14.40 & r & 0.07 & 1.1 & 15205 & + 13 35 12.61 & -31 33 45.0 & [email protected] & [email protected] & [email protected] & 2.33 & 1.37 & r & 0.07 & 1.4 & 13673 & + 13 35 15.57 & -31 30 48.2 & [email protected] & [email protected] & [email protected] & 0.59 & 0.39 & u & 0.06 & 5.7 & & + 13 35 18.70 & -32 32 30.4 & [email protected] & [email protected] & [email protected] & 4.71 & 2.85 & r & 0.15 & 4.9 & & + 13 35 18.90 & -31 34 57.5 & [email protected] & [email protected] & [email protected] & 0.71 & 0.81 & u & 0.07 & 1.3 & 58050 & + 13 35 29.53 & -30 52 56.1 & [email protected] & [email protected] & [email protected] & 2.07 & 2.46 & u & 0.21 & 1.5 & 12117 & + 13 35 32.44 & -31 07 56.2 & [email protected] & [email protected] & [email protected] & 2.45 & 2.97 & u & 0.08 & 2.2 & & a + 13 35 42.55 & -31 53 54.6 & [email protected] & [email protected] & [email protected] & 14.41 & 4.18 & r & 0.08 & 3.1 & 14385 & b , c + 13 35 47.72 & -32 27 23.9 & [email protected] & [email protected] & [email protected] & 2.07 & 1.16 & r & 0.15 & 3.8 & & a + 13 35 53.20 & -31 29 41.4 & [email protected] & [email protected] & [email protected] & 0.40 & 0.34 & r & 0.07 & 4.6 & & + 13 35 56.85 & -32 01 50.0 & [email protected] & [email protected] & [email protected] & 1.99 & 0.83 & r & 0.09 & 0.6 & & a + 13 35 59.75 & -31 23 04.6 & [email protected] & [email protected] & [email protected] & 0.26 & 0.39 & u & 0.07 & 6.1 & & + 13 36 18.04 & -31 46 45.3 & [email protected] & [email protected] & [email protected] & 2.58 & 2.12 & r & 0.08 & 0.7 & 12263 & + 13 36 18.70 & -31 48 31.2 & [email protected] & [email protected] & [email protected] & 2.76 & 2.32 & r & 0.07 & 0.7 & 11692 & + 13 36 20.97 & -31 34 39.7 & [email protected] & [email protected] & [email protected] & 1.00 & 0.95 & u & 0.08 & 3.7 & & + 13 36 26.46 & -31 06 36.4 & [email protected] & [email protected] & [email protected] & 0.80 & 0.84 & r & 0.12 & 1.0 & & + l r r r a3556 & 2.86 & -0.080 & 317 + a3558 & 2.69 & -0.067 & 542 + a3562 & 2.79 & -0.072 & 378 + sc1327 - 312 & 3.05 & -0.092 & 585 + sc1329 - 313 & 3.15 & -0.098 & 488 + all data & 2.49 & -0.054 & 1696 + r
r r c r r r r r c 13 22 06.75 & -31 46 16.4 & -22.1 & 3.02 & 1.30 & 3.75 & 7.21 & 4.86 & 5.99 & + 13 23 29.07 & -31 50 39.6 & -22.5 & 1.75 & 0.67 & 2.98 & 6.35 & 4.03 & 5.12 & + 13 23 54.22 & -32 10 56.3 & -22.0 & 2.37 & 1.55 & 3.25 & 6.23 & 4.07 & 4.96 & + 13 23 57.58 & -31 38 45.2 & -22.4 & 22.9 & 0.10 & 2.54 & 6.05 & 3.68 & 4.85 & + 13 24 06.68 & -31 40 12.0 & -22.9 & 4.28 & 0.03 & 2.46 & 5.95 & 3.59 & 4.75 & + 13 24 22.77 & -31 42 23.7 & -22.5 & 5.72 & 0.22 & 2.32 & 5.78 & 3.42 & 4.57 & + 13 24 32.46 & -31 06 58.2 & -20.7 & 1.03 & 1.64 & 2.42 & 5.95 & 3.64 & 4.94 & a + 13 26 18.81 & -30 57 37.0 & -22.3 & 2.19 & 2.51 & 1.88 & 5.06 & 2.93 & 4.21 & + 13 26 48.60 & -31 34 18.6 & -20.3 & 0.89 & 1.73 & 0.74 & 4.26 & 1.88 & 3.11 & + 13 26 49.79 & -31 23 44.5 & -22.1 & 3.83 & 1.90 & 0.74 & 4.32 & 1.97 & 3.27 & + 13 27 29.66 & -31 23 24.3 & -22.7 & 1.43 & 2.29 & 0.40 & 3.91 & 1.59 & 2.89 & + 13 27 54.95 & -31 32 18.6 & -22.5 & 4.04 & 2.43 & 0.14 & 3.57 & 1.20 & 2.47 & + 13 27 56.84 & -31 29 44.6 & -22.8 & 2.27 & 2.48 & 0.02 & 3.57 & 1.21 & 2.50 & + 13 28 02.63 & -31 45 21.3 & -22.1 & 13.3 & 2.50 & 0.79 & 3.47 & 1.18 & 2.25 & + 13 28 50.91 & -32 07 53.1 & -21.4 & 0.82 & 3.29 & 1.99 & 3.24 & 1.66 & 1.97 & + 13 28 52.29 & -32 09 45.8 & -21.4 & 3.00 & 3.35 & 2.08 & 3.27 & 1.74 & 2.00 & b + 13 28 57.71 & -31 51 34.9 & -20.9 & 0.73 & 3.12 & 1.28 & 2.93 & 0.91 & 1.67 & c + 13 29 28.57 & -31 39 25.7 & -20.4 & 0.95 & 3.39 & 1.10 & 2.56 & 0.24 & 1.42 & + 13 29 55.10 & -31 19 56.4 & _
-22.0 _ & 1.59 & 3.81 & 1.36 & 2.50 & 0.82 & 1.78 & + 13 29 57.07 & -31 11 08.1 & _
-21.0 _ & 0.81 & 3.97 & 1.58 & 2.69 & 1.26 & 2.14 & + 13 30 05.14 & -31 24 36.8 & -21.6 & 3.46 & 3.86 & 1.40 & 2.32 & 0.62 & 1.53 & + 13 30 42.03 & -31 34 28.7 & -22.2 & 0.70 & 4.18 & 1.78 & 1.81 & 0.59 & 0.91 & + 13 30 48.93 & -31 43 25.7 & -23.7 & 3.10 & 4.24 & 1.96 & 1.72 & 0.74 & 0.56 & d + 13 30 51.82 & -31 04 40.3 & -20.5 & 1.56 & 4.63 & 2.24 & 2.45 & 1.72 & 2.23 & a + 13 30 52.20 & -31 46 07.0 & -20.3 & 0.69 & 4.28 & 2.04 & 1.70 & 0.84 & 0.48 & + 13 31 00.75 & -31 44 53.9 & -20.8 & 0.97 & 4.37 & 2.10 & 1.60 & 0.88 & 0.42 & e + 13 31 16.90 & -31 35 38.4 & -20.7 & 0.81 & 4.54 & 2.15 & 1.44 & 0.95 & 0.68 & + 13 31 19.47 & -31 13 44.4 & -21.7 & 1.56 & 4.76 & 2.30 & 1.92 & 1.49 & 1.74 & + 13 31 20.22 & -31 44 34.6 & -20.7 & 0.98 & 4.57 & 2.29 & 1.40 & 1.06 & 0.27 & + 13 31 37.79 & -31 33 12.0 & -21.4 & 1.20 & 4.77 & 2.36 & 1.25 & 1.18 & 0.77 & f + 13 31 54.87 & -31 54 04.1 & -20.6 & 0.80 & 4.97 & 2.80 & 1.22 & 1.60 & 0.33 & + 13 32 01.52 & -31 13 58.0 & -21.6 & 0.73 & 5.16 & 2.71 & 1.62 & 1.80 & 1.74 & a + 13 32 03.17 & -31 46 48.6 & -22.4 & 1.77 & 5.03 & 2.75 & 0.98 & 1.52 & 0.30 & + 13 32 30.07 & -31 03 15.8 & -20.2 & 1.26 & 5.63 & 3.19 & 1.95 & 2.38 & 2.32 & + 13 32 35.44 & -31 31 45.5 & -20.2 & 0.74 & 5.38 & 2.96 & 0.73 & 1.79 & 1.05 & + 13 32 46.48 & -31 37 03.3 & _
-21.0 _ & 0.75 & 5.48 & 3.09 & 0.51 & 1.89 & 0.94 & + 13 32 46.75 & -31 57 59.2 & -22.2 & 0.91 & 5.54 & 3.37 & 0.99 & 2.17 & 0.87 & g + 13 33 04.16 & -31 39 04.0 & -20.1 & 0.85 & 5.66 & 3.29 & 0.30 & 2.08 & 1.05 & + 13 33 31.63 & -31 40 58.6 & _
-21.0 _ & 42.5 & 5.95 & 3.59 & 0.03 & 2.37 & 1.28 & + 13 33 40.7 & -31 19 57 . & -20.6 & 2.57 & 6.13 & 3.68 & 1.02 & 2.59 & 1.94 & a + 13 34 39.94 & -31 32 54.5 & -20.5 & 2.21 & 6.68 & 4.28 & 0.81 & 3.09 & 2.09 & + 13 35 03.06 & -31 39 18.6 & -22.4 & 6.38 & 6.92 & 4.53 & 0.96 & 3.33 & 2.23 & + 13 35 12.61 & -31 33 45.0 & -21.1 & 0.96 & 7.03 & 4.62 & 1.11 & 3.43 & 2.40 & + 13 35 42.55 & -31 53 54.6 & -21.9 & 5.93 & 7.35 & 5.06 & 1.53 & 3.83 & 2.60 & + 13 35 56.86 & -32 01 50.0 & -20.4 & 0.82 & 7.56 & 5.33 & 1.86 & 4.09 & 2.82 & h + 13 36 18.04 & -31 46 45.3 & -22.1 & 1.06 & 7.70 & 5.36 & 1.78 & 4.14 & 2.97 & + 13 36 18.70 & -31 48 31.2 & -21.4 & 1.14 & 7.71 & 5.38 & 1.80 & 4.16 & 2.97 & i + 13 19 07.85 & -31 40 10.8 & _
-21.0 _ & 0.28 & 3.14 & 5.57 & 9.09 & 6.73 & 7.89 & + 13 20 54.63 & -31 44 33.4 & -21.1 & 3.27 & 2.03 & 4.48 & 7.97 & 5.61 & 6.75 & + 13 20 55.59 & -31 43 59.3 & -21.4 & 2.74 & 2.02 & 4.47 & 7.96 & 5.60 & 6.74 & + 13 21 45.02 & -31 03 00.6 & -21.8 & 3.21 & 2.35 & 4.13 & 7.71 & 5.37 & 6.66 & + 13 21 52.65 & -31 45 54.4 & -21.6 & 0.59 & 1.44 & 3.89 & 7.36 & 5.01 & 6.14 & + 13 22 49.71 & -32 26 11.6 & -22.1 & 1.31 & 2.44 & 4.24 & 7.08 & 5.00 & 5.80 & + 13 24 39.14 & -32 06 33.4 & -21.5 & 0.24 & 1.38 & 2.75 & 5.73 & 3.55 & 4.46 & + 13 24 50.44 & -31 33 37.8 & -20.5 & 0.48 & 0.55 & 1.95 & 5.50 & 3.13 & 4.34 & + 13 25 50.32 & -32 08 54.2 & -22.4 & 0.26 & 1.82 & 2.35 & 5.06 & 2.96 & 3.77 & + 13 26 25.80 & -31 27 20.9 & -21.4 & 0.32 & 1.59 & 0.95 & 4.54 & 2.17 & 3.44 & + 13 26 35.69 & -30 59 36.4 & -20.1 & 0.22 & 2.54 & 1.70 & 4.85 & 2.73 & 4.01 & + 13 26 49.94 & -30 37 56.9 & -20.3 & 0.79 & 3.51 & 2.64 & 5.28 & 3.46 & 4.64 & + 13 26 52.95 & -30 00 28.6 & -21.7 & 0.16 & 5.22 & 4.46 & 6.53 & 5.10 & 6.16 & + 13 26 54.01 & -31 18 39.5 & -19.8 & 0.19 & 2.05 & 0.84 & 4.34 & 2.03 & 3.33 & + 13 26 57.51 & -32 11 52.3 & -20.8 & 0.32 & 2.40 & 2.18 & 4.41 & 2.49 & 3.13 & + 13 27 03.35 & -31 13 31.5 & -21.4 & 0.47 & 2.28 & 0.96 & 4.32 & 2.07 & 3.37 & + 13 27 29.66 & -30 33 31.8 & -21.1 & 1.74 & 3.92 & 2.78 & 5.08 & 3.44 & 4.55 & + 13 27 34.96 & -31 32 59.2 & -21.8 & 0.33 & 2.22 & 0.27 & 3.78 & 1.40 & 2.66 & + 13 27 45.50 & -31 47 51.0 & -22.5 & 0.20 & 2.34 & 0.91 & 3.66 & 1.40 & 2.42 & + 13 27 51.48 & -31 14 45.1 & -19.6 & 0.64 & 2.68 & 0.73 & 3.82 & 1.63 & 2.91 & + 13 28 03.55 & -30 40 50.9 & -20.5 & 0.60 & 3.85 & 2.41 & 4.57 & 2.96 & 4.05 & + 13 28 10.37 & -31 20 57.4 & -20.9 & 0.38 & 2.74 & 0.46 & 3.53 & 1.28 & 2.57 & + 13 28 33.79 & -31 26 28.3 & -20.7 & 0.35 & 2.90 & 0.44 & 3.22 & 0.92 & 2.22 & + 13 28 35.78 & -31 30 54.3 & -20.1 & 0.27 & 2.88 & 0.44 & 3.16 & 0.80 & 2.10 & + 13 28 38.72 & -31 20 49.3 & -22.1 & 0.25 & 3.03 & 0.63 & 3.24 & 1.06 & 2.33 & + 13 28 39.26 & -31 39 41.1 & -19.9 & 0.40 & 2.87 & 0.69 & 3.08 & 0.73 & 1.91 & + 13 28 54.30 & -31 55 11.1 & -20.8 & 0.29 & 3.12 & 1.41 & 3.00 & 1.08 & 1.73 & + 13 28 59.29 & -31 44 15.7 & -20.2 & 0.65 & 3.09 & 1.00 & 2.87 & 0.63 & 1.66 & + 13 29 04.33 & -31 47 52.4 & -20.5 & 0.28 & 3.16 & 1.17 & 2.84 & 0.72 & 1.60 & + 13 29 12.29 & -31 18 39.9 & _
-21.0 _ & 0.23 & 3.40 & 0.98 & 2.95 & 0.95 & 2.13 & + 13 29 14.98 & -31 16 44.9 & -21.5 & 0.44 & 3.45 & 1.06 & 2.96 & 1.03 & 2.18 & + 13 29 17.00 & -32 02 33.9 & -21.9 & 0.47 & 3.45 & 1.85 & 2.89 & 1.33 & 1.61 & + 13 29 20.43 & -30 46 37.4 & -20.8 & 0.67 & 4.25 & 2.31 & 3.77 & 2.48 & 3.40 & + 13 29 31.44 & -31 37 29.0 & -20.3 & 0.36 & 3.43 & 1.09 & 2.54 & 0.17 & 1.43 & + 13 29 54.20 & -32 00 59.3 & -21.0 & 0.42 & 3.80 & 2.00 & 2.50 & 1.21 & 1.23 & + 13 30 05.20 & -31 13 53.4 & -20.2 & 0.39 & 4.00 & 1.58 & 2.55 & 1.13 & 1.97 & + 13 30 24.83 & -30 41 37.1 & -21.4 & 0.55 & 4.95 & 2.86 & 3.52 & 2.74 & 3.41 & + 13 30 32.80 & -31 34 33.8 & -22.4 & 0.61 & 4.08 & 1.68 & 1.91 & 0.49 & 0.97 & + 13 30 40.65 & -31 42 32.1 & -19.3 & 0.22 & 4.15 & 1.86 & 1.80 & 0.64 & 0.66 & + 13 30 40.74 & -31 51 56.0 & -20.7 & 0.18 & 4.19 & 2.07 & 1.89 & 0.95 & 0.60 & + 13 30 41.47 & -31 34 17.3 & -21.2 & 0.64 & 4.17 & 1.77 & 1.82 & 0.58 & 0.92 & + 13 30 47.73 & -31 51 06.5 & -21.4 & 0.23 & 4.26 & 2.11 & 1.80 & 0.96 & 0.52 & + 13 30 51.04 & -31 44 17.9 & -21.5 & 0.23 & 4.26 & 1.99 & 1.70 & 0.78 & 0.52 & + 13 30 52.70 & -31 27 13.2 & -20.6 & 0.39 & 4.33 & 1.88 & 1.80 & 0.83 & 1.16 & + 13 30 53.83 & -31 30 21.7 & -21.0 & 0.21 & 4.32 & 1.89 & 1.74 & 0.77 & 1.01 & + 13 30 56.16 & -31 33 42.9 & -19.9 & 0.23 & 4.33 & 1.92 & 1.67 & 0.74 & 0.85 & + 13 31 01.99 & -31 40 36.3 & -19.8 & 0.30 & 4.38 & 2.05 & 1.58 & 0.82 & 0.54 & + 13 31 17.43 & -31 24 34.7 & -20.7 & 0.58 & 4.61 & 2.15 & 1.62 & 1.12 & 1.21 & + 13 31 22.42 & -31 22 10.4 & -19.6 & 0.21 & 4.69 & 2.22 & 1.64 & 1.23 & 1.32 & + 13 31 27.58 & -31 49 14.7 & -22.1 & 0.65 & 4.66 & 2.44 & 1.38 & 1.23 & 0.09 & + 13 31 59.81 & -31 49 22.2 & -19.8 & 0.37 & 5.00 & 2.76 & 1.06 & 1.53 & 0.25 & + 13 32 38.27 & -31 01 41.3 & -20.0 & 0.55 & 5.74 & 3.31 & 2.00 & 2.50 & 2.42 & + 13 32 42.95 & -31 55 13.7 & -20.5 & 0.14 & 5.48 & 3.28 & 0.90 & 2.07 & 0.77 & + 13 32 43.89 & -31 47 01.3 & -20.7 & 0.28 & 5.46 & 3.16 & 0.60 & 1.93 & 0.72 & + 13 33 01.00 & -31 13 29.3 & -20.4 & 0.50 & 5.80 & 3.33 & 1.37 & 2.35 & 1.96 & + 13 33 23.53 & -31 24 23.3 & -20.9 & 0.16 & 5.93 & 3.48 & 0.80 & 2.36 & 1.66 & + 13 33 34.76 & -31 40 20.5 & -22.3 & 0.16 & 5.99 & 3.62 & 0.03 & 2.41 & 1.32 & + 13 33 50.43 & -31 28 38.0 & -20.9 & 0.25 & 6.19 & 3.75 & 0.61 & 2.60 & 1.73 & + 13 33 58.73 & -31 58 16.1 & -20.4 & 0.49 & 6.29 & 4.07 & 0.93 & 2.85 & 1.57 & + 13 34 31.48 & -32 05 09.0 & -20.8 & 0.41 & 6.67 & 4.52 & 1.38 & 3.30 & 2.01 & + 13 34 36.19 & -31 41 06.7 & -22.6 & 0.61 & 6.63 & 4.26 & 0.68 & 3.05 & 1.93 & + 13 34 40.80 & -31 26 43.6 & -21.3 & 0.65 & 6.73 & 4.29 & 0.99 & 3.14 & 2.23 & + 13 34 43.00 & -32 22 14.0 & -21.4 & 0.46 & 6.98 & 5.00 & 2.20 & 3.83 & 2.56 & + 13 35 29.53 & -30 52 56.1 & -21.5 & 1.01 & 7.61 & 5.16 & 2.66 & 4.23 & 3.71 & + l r r r r@ - r a3556 & 7 & 12.7 & _ 98.7 _ & 92.3 & 99.9 + a3558 & 8 & 21.7 & 82.0 & 63.6 & 93.7 + a3562 & 5 & 11.7 & 85.7 & 59.9 & 98.1 + sc1327 - 312 & 8 & 19.7 & 89.9 & 75.5 & 97.4 + sc1329 - 313 & 3 & 12.7 & 0.1 & 34.0 & 43.8 + comparison & 40 & 169.4 & & & + l r r r r@ - r a3556 & 0 & 34.3 & _ 98.5 _ & 96.9 & 99.2 + a3558 & 3 & 87.3 & _ 99.7 _ & 99.4 & 99.8 + a3562 & 3 & 58.3 & _ 96.0 _ & 92.8 & 97.8 + extended & 5 & & 81.3 & 69.5 & 88.8 + sc1327 - 312 & 4 & 108.3 & _ 99.9 _ & 99.7 & 99.9 + extended & 6 & & _ 99.2 _ & 98.6 & 99.5 + sc1329 - 313 & 4 & 83.3 & _ 98.8 _ & 98.0 & 99.3 + extended & 6 & & 94.3 & 90.7 & 96.6 + comparison & 87 & 582.4 & & & + l r r r r@ - r a3556 & 1 & 81.2 & 29.7 & 1.0 & 49.9 + extended & 2 & & 32.6 & 2.5 & 67.5 + a3558 & 3 & 180.2 & 10.9 & 10.0 & 28.4 + extended & 4 & & 29.3 & 9.6 & 50.1 + a3562 & 11 & 113.2 & _
@xmath12199.9 _ & @xmath12199.9 & @xmath12199.9 + extended & 12 & & _ @xmath12199.9 _ & @xmath12199.9 & @xmath12199.9 + sc1327 - 312 & 9 & 196.2 & _ 98.5 _ & 96.2 & 99.6 + extended & 11 & & _ 99.9 _ & @xmath12199.9 & @xmath12199.9 + sc1329 - 313 & 9 & 142.2 & _ 99.9 _ &
@xmath12199.9 & @xmath12199.9 + extended & 11 & & _ @xmath12199.9 _ & @xmath12199.9 & @xmath12199.9 + comparison & 20 & 1104.6 & & & + | the core of the shapley supercluster ( a3556 , a3558 , sc1327 - 312 , sc1329 - 313 , and a3562 ) is an ideal region in which to study the effects of cluster mergers on the activity of individual galaxies .
this paper presents the most comprehensive radio continuum investigation of the region , relying on a 63-pointing mosaic obtained with the very large array yielding an areal coverage of nearly 7 square degrees .
the mosaic provides a typical sensitivity of about 80 @xmath0jy at a resolution of 16 , enabling detection of galaxies with star formation rates as low as 1 m@xmath1 year@xmath2 .
the radio data are complemented by optical imaging in @xmath3 and @xmath4 , producing a catalog of 210 radio - detected galaxies with @xmath5 ( @xmath6 ) .
at least 104 of these radio - detected galaxies are members of the supercluster on the basis of public velocity measurements . across the entire core of the supercluster , there appears to be a significant deficit of radio galaxies at intermediate optical magnitudes ( @xmath7 ) .
this deficit is offset somewhat by an increase in the frequency with which brighter galaxies ( @xmath8 ) host radio sources .
more dramatic is the highly significant increase in the probability for fainter galaxies ( @xmath9 ) in the vicinity of a3562 and sc1329 - 313 to be associated with radio emission .
the radio and optical data for these sources strongly suggest that these active galaxies are powered by star formation . in conjunction with recent x - ray analysis ,
this is interpreted as young starbursts related to the recent merger of sc1329 - 313 with a3562 and the rest of the supercluster . |
bulk silicon is an indirect band - gap material emitting in the infrared region .
radiative lifetimes of excited carrier are very long , causing a predominant de - excitation via fast non - radiative recombinations .
moreover , silicon has a significant free carrier absorption and auger recombination rates which makes the use of this material in optoelectronic applications very problematic . during the last decade ,
several breakthroughs have increased the hopes of using nanostructured silicon as an optical active material.@xcite the basic idea has been to take advantage of the reduced dimensionality of the nanocrystalline phase ( 1 - 5 nm in size ) where quantum confinement , band folding and surface effects play a crucial role.@xcite indeed , it has been found that si nanocrystals band - gap increases with decreasing size with a luminescence external efficiency in excess of 23%.@xcite moreover , optical gain has been already demonstrated in a large variety of experimental conditions.@xcite nevertheless , si nanocrystals ( si - nc ) still have a memory of the indirect band gap of the bulk phase and this is evidenced by the clearly observed structures related to momentum - conserving phonons.@xcite this drawback can be circumvented by introducing an isoelectronic impurities .
@xcite indeed , in a series of recent very interesting papers , fujii and collaborators @xcite have proved the possibility of a detailed control of the si - nc photoluminescence by the simultaneous doping with _ n_- and _ p_- type impurities .
in particular , they have showed that a ( b and p ) codoped si - nc always has an higher photoluminescence intensity than that of both a single ( b or p ) doped and of an undoped nanocrystal . besides , under resonant excitation condition , the codoped samples did not exhibit structures related to momentum - conserving phonons suggesting that , in this case , the quasi - direct optical transitions are predominant.@xcite from a theoretical point of view , investigations of impurities in silicon nanostructures are very few when compared to the large number of papers reporting calculations for pure , undoped systems ; moreover , most of them are based on semi - empirical approaches .
an handful number of first - principles studies has been devoted to quantum confinement effects in single - doped si - nc.@xcite these works have basically shown that i ) the si - nc ionization energy is virtually size independent , ii ) the impurity formation energy is greater for smaller nanocrystals and iii ) impurity segregation strongly affects the conductance properties of nanostructures.@xcite we have recently started a systematic study of the electronic and optical properties of codoped si - nc .
our first results @xcite show that codoped si - nc undergo a minor structural distortion around the impurities and that the formation energies are always smaller than those of the corresponding single - doped cases .
moreover , we have found that codoping reduces the band - gap with respect to the undoped nanocrystals , showing the possibility of an impurity based engineering of the si - nc optical properties.@xcite we report here a comprehensive theoretical study of the structural , electronic and optical properties of b and p simultaneously doped si nanocrystals using _ ab initio _ density functional theory .
the paper is organized as follows .
section [ comp ] is focused on the description of the theoretical and computational methods , section [ res ] is dedicated to the discussion of our results .
the results will be presented and discussed in the following order : i ) structural properties ( subsection [ geometri ] ) , ii ) formation energies ( subsection [ forme ] ) and iii ) electronic ( subsection [ elpr ] ) and iv ) optical properties ( subsection [ opti ] ) for all the considered si nanocrystals . concerning the optical properties we discuss both absorption and emission spectra obtained within a single - particle approach ( [ abs ] ) and , in one case , with many - body methods considering self - energy corrections and the bethe - salpeter scheme ( [ lum ] ) .
finally , in section [ conc ] we summarize our results .
our results are obtained within a plane - wave pseudopotential dft scheme , using the quantum - espresso package @xcite .
the impurity states are calculated in an approximately spherical si - nc , built by considering all the bulk si atoms contained in a sphere ( centered on a si ion ) with a diameter ranging from 1.1 nm ( si@xmath0h@xmath1 ) to 1.79 nm ( si@xmath2h@xmath3 ) .
the surface dangling bonds are saturated with hydrogens . following the work of fujii et al .
@xcite , we have located the b and p impurities in substitutional positions just below the nanocrystal surface .
it is worth mentioning that this arrangement represents the most stable configuration , as confirmed by theoretical and experimental works.@xcite full relaxation with respect to the atomic positions has been allowed for both doped and undoped systems .
all the dft calculations are performed within the generalized gradient approximation ( gga ) using vanderbilt ultrasoft @xcite pseudopotentials for the determination of both the structural and electronic properties ( see sec .
[ geometri ] , sec .
[ forme ] and sec .
[ elpr ] ) whereas norm - conserving pseudopotential within the local density approximation ( lda ) at the relaxed geometry have been used to evaluate the optical properties ( see sec .
[ opti ] ) .
this choice is due to the fact that although vanderbilt ultrasoft pseudopotentials allow the treatment of several hundreds of atoms per unit cell in the atomic relaxation process , the removal of the norm - conservation condition is a well known problem for the calculation of the optical transition matrix elements @xcite .
each nanocrystal has been placed in a large supercell in order to prevent interactions between the periodic replicas ( about 6 of vacuum separates neighbor nanocrystals ) .
structural , electronic and optical properties , as well as the impurity formation energies , are investigated as a function of the size and for several impurity positions within the si - nc .
starting from the si@xmath4h@xmath5 nanocrystal,@xcite the formation energy of the neutral b or / and p impurities can be defined as the energy needed to insert one b or / and one p atom within the nanocrystal after removing one / two si atoms ( transferred to the chemical reservoir , assumed to be bulk si)@xcite + @xmath6 where e is the total energy of the system , @xmath7 the total energy per atom of bulk si and @xmath8 the total energy per atom of the impurity ( we consider the total energy per atom in the tetragonal b@xmath9 crystal for b , as in ref . and the orthorhombic black phosphorus for p , as in ref . ) .
the integers @xmath10 and @xmath11 can be set to either @xmath12 or @xmath13 .
in particular , @xmath14 is the choice when a b impurity is present in the nanocrystal ( @xmath12 otherwise ) and @xmath15 for a p impurity ( @xmath12 otherwise ) . with this prescription , eq.[deltahf ]
can be used for both the single doping and the codoping case .
the calculations of the optical properties have been done both in the ground and in the first excited states , where the excited state corresponds to the electronic configuration in which the highest occupied single - particle state ( homo ) contains a hole , while the lowest unoccupied single - particle state ( lumo ) contains the corresponding electron.@xcite it is worth pointing out that an undoped and relaxed si - nc have t@xmath16 symmetry ; in the presence of doping , this high symmetry is generally lost due to the presence of the impurity atoms . moreover , because of the significant differences in the charge density of the ground and excited states , the actual atomic relaxations in the two cases are different .
the nanocrystal optical response is evaluated for both the ground and the excited state relaxed geometries computing the imaginary part of the dielectric function ( @xmath17 ) through the fermi golden rule .
the emission spectra have been calculated using the excited state atomic positions and the ground state electronic configuration ( more details can be found in section [ lum ] ) .
it should be noted that although @xmath17 should only be used for calculating the nanocrystal absorption coefficient , it can also be used for getting a first approximation to the emission spectra simply because the emission can be viewed as the time reversal of the absorption.@xcite in other words , once the relaxed atomic positions corresponding to to an hole in the homo and an electron in the lumo have been found , these atomic positions are used for the calculation of @xmath17 whose main features are also those of the emission spectra .
it is worth mentioning that the photoluminescence spectra can be derived using the well known van roosbroeck and shockley @xcite relation which , again , involves @xmath17 .
however , such a calculation requires the knowledge of the electron and hole populations , at the working temperature , in the lumo and homo states respectively .
the populations , in turn , depends on the actual dynamics in the excitation and emission processes , including the non radiative electron hole recombinations . in this work
we have not considered any particular dynamics so that our emission spectra contains only the informations related to both the transition energies and the oscillator strengths . .[tablesi87 ] bond lengths ( in ) around the impurity sites for the undoped , single doped and codoped si@xmath18h@xmath19 nanocrystal ( diameter 1.50 nm ) .
b and p impurities have been substitutionally located at subsurface positions ( see fig . [ si85bp ] ) .
si@xmath20 and si@xmath21 refer to two surface and two inner si atoms around this site respectively .
[ cols="<,^,^,^,^,^ " , ] the differences between full many - body calculations and single - particle results are of 0.55 ev and 0.41 ev for absorption and emission energy gaps respectively , and of 0.14 ev between the two stokes shifts .
it is interesting to note that the homo - lumo transition in the emission spectrum at 2.20 ev is almost dark while an important excitonic peak is evident at about 2.75 ev ( see fig .
[ si33abs - ems ] ) , again red - shifted with respect to the first absorption peak . as expected , what comes out is the importance of fully taking into account the many - body aspect of the problem in order to overcome the limits of the single - particle approach .
the structural , electronic and optical properties of si nanocrystals codoped with b and p impurities have been studied also going beyond the single - particle approach .
we have considered si - nc of different size and with the impurities located at different distances .
we show that codoping is always energetically favored with respect to simple b- or p - doping and that the two impurities tend to occupy nearest neighbor sites near the surface rather than other positions inside the nanocrystal itself .
our results demonstrate that the codoped nanocrystals present valence and conduction band - edge states which are localized on the two impurities respectively and band - gaps always lower in energy with respect to that of undoped si nanocrystals . besides , the electronic properties show a dependence on both nanocrystal size and impurity - impurity distance .
the impurity located band - edge states originate absorption thresholds in the visible region which are shifted lower in energy with respect to the undoped case .
moreover , the emission spectra show a stokes shift with respect to the absorption which is due to the structural relaxation after the creation of the electron - hole pair .
our results make evident the presence of electronic quasi - direct optical transitions between donor and acceptor states that can help to understand the experimental outcomes and makes it possible to engineer the absorption and emission properties of si nanocrystals .
we acknowledge the support of the miur prin ( 2005 ) italy , of the cnr - cnism `` progetto innesco '' of the crui vigoni project ( 2005 - 2006 ) italy - germany and of eu nanoquanta network of excellence ( nmp4-ct-2004 - 500198 ) .
all the calculations were performed at cineca - bologna ( `` iniziativa calcolo parallelo del cnr - infm '' ) , cicaia - modena and `` campus computational grid''-universit di napoli `` federico ii '' .
s. ossicini , l. pavesi , and f. priolo , _ light emitting silicon for microphotonics _ , springer - berlin ( 2003 ) .
v. n. borisenko and s. ossicini , _ what is what in the nanoworld _ , wiley - vch , weinheim ( 2004 ) .
o. bisi , s. ossicini , and l. pavesi , surf .
reports * 38 * , 5 ( 2000 ) .
b. gelloz , a. kojima , and n. koshida , appl .
. lett . * 87 * , 031107 ( 2005 ) .
l. pavesi , l. dal negro , c. mazzoleni , g. franz , and f. priolo , nature * 408 * , 440 ( 2000 ) .
l. dal negro , m. cazzanelli , l. pavesi , s. ossicini , d. pacifici , g. franz , and f. priolo , appl .
. lett . * 82 * , 4636 ( 2003 ) .
j. ruan , p. m. fauchet , l. dal negro , m. cazzanelli , and l. pavesi , appl .
. lett . * 83 * , 5479 ( 2003 ) .
m. cazzanelli , d. kovalev , l. dal negro , z. gaburro , and l. pavesi , phys .
* 93 * , 207042 ( 2004 ) .
k. luterov , k. dohnalov , v. ervek , i. pelant , j .-
likforman , o. crgut , p. gilliot , and b. hnerlage , appl .
. lett . * 84 * , 3280 ( 2004 ) .
_ properties of porous silicon _ , ed . by l. t. canham ( iee inspec , london 1997 ) . c. delerue , g. allan , and m. lannoo , phys .
b * 64 * , 193402 ( 2001 ) .
m. fujii , y. yamaguchi , y. takase , k. ninomiya , and s. hayashi , appl .
. lett . * 87 * , 211919 ( 2005 ) .
m. fujii , k. toshikiyo , y. takase , y. yamaguchi , and s. hayashi , j. appl .
* 94 * , 1990 ( 2003 ) .
m. fujii , y. yamaguchi , y. takase , k. ninomiya , and s. hayashi , appl .
. lett . * 85 * , 1158 ( 2004 ) .
d. v. melnikov and j. r. chelikowsky , phys .
* 92 * , 046802 ( 2004 ) .
g. cantele , e. degoli , e. luppi , r. magri , d. ninno , g. iadonisi , and s. ossicini , phys .
b*72 * , 113303 ( 2005 ) .
z. zhou , m. l. steigerwald , r. a. friesner , l. brus , and m. s. hybertsen , phys .
rev . b*71 * , 245308 ( 2005 ) .
m. v. fernndez - serra , ch .
adessi , and x. balse , phys .
lett . * 96 * , 166805 ( 2006 ) .
s. ossicini , e. degoli , f. iori , e. luppi , r. magri , g. cantele , f. trani , and d. ninno , appl .
phys . lett . *
87 * , 173120 ( 2005 ) .
s. ossicini , f. iori , e. degoli , e. luppi , r. magri , r. poli , g. cantele , f. trani , and d. ninno , j. select .
topics in quantum electr .
* 12 * , 1585 ( 2006 ) .
s. baroni , a. dal corso , s. de gironcoli , p. giannozzi , c. cavazzoni , g. ballabio , s. scandolo , g. chiarotti , p. focher , a. pasquarello , k. laasonen , a. trave , r. car , n. marzari , and a. kokalj , http://www.pwscf.org/. l. colombi ciacchi and m. c. payne , phys .
lett . * 95 * , 196101 ( 2005 ) .
e. garrone , f. geobaldo , p. rivolo , g. amato , l. boarino , m. chiesa , e. giamello , r. gobetto , p. ugliengo , and a. vitale , adv . mat . * 17 * , 528 ( 2005 ) .
d. vanderbilt , phys .
b * 41 * r7892 ( 1990 ) .
we have checked that the use of vanderbilt or norm - conserving pseudopotentials does not affect the calculated structural properties of the si - nc .
e. degoli , g. cantele , e. luppi , r. magri , d. ninno , o. bisi , and s. ossicini , phys .
b * 69 * , 155411 ( 2004 ) . s. b. zhang and j. e. northrup , phys .
lett . * 67 * , 2339 ( 1991 ) .
x. luo , s. b. zhang , and s .- h .
wei , phys .
90 * , 026103 ( 2003 ) .
l. g. wang and a. zunger , phys .
b * 66 * , 161202(r ) ( 2002 ) .
a. franceschetti and s.t .
pantelides , phys .
b * 68 * , 033313 ( 2003 ) .
a. puzder , a.j .
williamson , j.c .
grossman , and g. galli , j. am .
125 * , 2786 ( 2003 ) .
g. onida , l. reining , and a. rubio , rev . of mod
. phys . * 74 * , 601 ( 2002 ) , and references therein .
e. luppi , f. iori , r. magri , o. pulci , e. degoli , s. ossicini , and v. olevano , phys .
b * 75 * , 033303 ( 2007 ) .
f. bassani and g. pastori parravicini , _ electronic states and optical transitions in solids _
, pergamon press , new york 1975 .
w. van roosbroeck , w. shockley , phys .
94 , 1558 ( 1954 ) .
l. hedin , phys . rev . * 139 * , a796 ( 1965 ) .
f. trani , d. ninno , and g. iadonisi , phys .
b * 75 * , 033312 ( 2007 ) .
p. y. yu and m. cardona , _ fundamentals of semiconductors _ , springer - berlin ( 2001 ) .
d. ninno , f. trani , g. cantele , k. hameeuw , g. iadonisi , e. degoli , and s. ossicini , europhys .
lett . * 74 * , 519 ( 2006 ) .
f. trani , d. ninno , g. cantele , g. iadonisi , k. hameeuw , e. degoli , and s. ossicini , phys .
b * 73 * , 245430 ( 2006 ) .
r. m. wentzcovitch , k. j. chang , and m. l. cohen , phys .
b * 34 * , 1071 ( 1986 ) .
f. iori , e. degoli , e. luppi , r. magri , i. marri , g. cantele , d. ninno , f. trani , and s. ossicini , j. lumin .
* 121 * , 335 ( 2006 ) .
r. w. godby and i. d. white phys .
lett . * 80 * , 3161 ( 1998 ) .
a. franceschetti , l.w .
wang , and a. zunger , phys .
lett . * 83 * , 1269 ( 1999 ) .
weissker , j. furthmller , and f. bechstedt , phys .
. lett . * 90 * , 085501 ( 2003 ) .
we note that the difference that arises in calculating the energy gap using the homo - lumo gap or the @xmath22-scf method strongly shrinks as the dimension of the nanocrystal increases @xcite .
we have used the non selfconsistent g@xmath23w@xmath23 approach within the rpa plasmon pole approximation .
we use a planewave - frequency space code .
exc code , v. olevano , http://www.bethe-salpeter.org .
c. delerue , m. lannoo , and g. allan , phys .
lett . * 84 * , 2457 ( 2000 ) . c. d. spataru , s. ismail - beigi , l. x. benedict , and s. g. louie .
lett . * 92 * , 077402 ( 2004 ) .
e. chang , g. bussi , a. ruini , and e. molinari , phys .
* 92 * , 196401 ( 2004 ) .
m. bruno , m. palummo , a. marini , r. del sole , v. olevano , a. n. kholod , and s. ossicini , phys .
b * 72 * , 153310 ( 2005 ) .
m. bruno , m. palummo , a. marini , r. del sole , and s. ossicini , phys .
rev . lett .
* 98 * , 036807 ( 2007 ) . | we show that the optical and electronic properties of nanocrystalline silicon can be efficiently tuned using impurity doping . in particular ,
we give evidence , by means of ab - initio calculations , that by properly controlling the doping with either one or two atomic species , a significant modification of both the absorption and the emission of light can be achieved .
we have considered impurities , either boron or phosphorous ( doping ) or both ( codoping ) , located at different substitutional sites of silicon nanocrystals with size ranging from 1.1 nm to 1.8 nm in diameter .
we have found that the codoped nanocrystals have the lowest impurity formation energies when the two impurities occupy nearest neighbor sites near the surface .
in addition , such systems present band - edge states localized on the impurities giving rise to a red - shift of the absorption thresholds with respect to that of undoped nanocrystals .
our detailed theoretical analysis shows that the creation of an electron - hole pair due to light absorption determines a geometry distortion that in turn results in a stokes shift between adsorption and emission spectra . in order to give a deeper insight in this effect , in one case we have calculated
the absorption and emission spectra going beyond the single - particle approach showing the important role played by many - body effects .
the entire set of results we have collected in this work give a strong indication that with the doping it is possible to tune the optical properties of silicon nanocrystals . |
the stellar initial mass function ( imf ) , an essential ingredient for numerous astrophysical problems , is commonly assumed to be invariant with time and with galactic properties .
this _ ansatz _ follows from a lack of evidence for its variation , despite searches covering a wide range of environments ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , more than from theoretical understanding of the processes that regulate the mass distribution of stars .
however , several recent studies have questioned the idea of a universal imf in both nearby and distant galaxies .
considering a few examples in the local universe , a truncated and/or steeper imf has been invoked to explain an apparent correlation between the masses of clusters and the masses of their largest members ( @xcite , though see @xcite for an opposing view ) , or between galaxy colors and h@xmath0 equivalent widths @xcite , as well as between the ratio of h@xmath0 to fuv luminosity ( @xmath3 ) and the surface brightness @xcite or the star formation rates ( sfrs ; * ? ? ?
* ; * ? ? ?
* ) in galaxies
. however , these claims remain controversial , and alternative interpretations have been offered , without resorting to imf variations @xcite . in this letter
we focus on observations of systematic variations in @xmath3 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) that have been used to argue strongly for a non - universal imf ( e.g. * ? ?
* ; * ? ? ?
recombination lines such as h@xmath0 ultimately come from the ionizing photons produced primarily by very massive stars , while fuv luminosity is driven by less massive stars ( @xmath4 m@xmath5for h@xmath0 , @xmath6 m@xmath5 for fuv , using a @xcite imf ) . as a result , the ratio @xmath3 can be used as a probe for the massive end of the imf , although the interpretation of observations is complicated by additional factors ( e.g. dust ) .
recently , both ( * ? ? ?
* hereafter m09 ) and ( * ? ?
* hereafter l09 ) reported that @xmath3 varies systematically with galaxy properties , which may imply a variation of the imf .
* hereafter b09 ) have cautioned that dust correction or a bursty star formation history ( sfh ) can produce the observed trends even if the underlying imf is universal , but this idea has not been explored in detail with models . while they can not rule out variations in the imf , l09 also argue that the underlying cause for the observed trend is not clear and stochastic effects at low sfr need to be explored further . following these suggestions ( see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , in this work we ask whether the observed @xmath3 is consistent with a random but incomplete sampling of a canonical or modified imf . to test these hypotheses , we combine observations from b09 , l09 and m09 with models from slug @xcite , a novel fully - stochastic code for synthetic photometry of stellar clusters and galaxies .
we compile a sample of 457 galaxies from b09 , l09 and m09 with integrated h@xmath0 and fuv luminosities , together with corrections for dust extinction and [ ] contamination . due to differential absorption ,
particular care is required when performing dust corrections . here
we follow the procedure adopted by each author in their studies , a choice motivated by the fact that the different dust extinctions are consistent across the three samples .
we refer to the original works for additional discussion . similarly , a careful analysis of the selection biases should be a prerequisite for comparison with models .
unfortunately , these data - sets have been assembled adopting different selection criteria ( a nearly complete volume limited sample for l09 , a subset of galaxies selected from the mass function for m09 and a sample of galaxies with multi - wavelength observations for b09 ) that are difficult to characterize .
we will therefore emphasize only those results that are believed to be less sensitive to selection biases .
the imf describes the relative probability with which stars at any mass are formed . at high sfrs ,
the large number of stars guarantees a nearly complete sampling of the imf , and the maximum stellar mass @xmath7 that can be found in a galaxy approaches the theoretical upper end @xmath8 of the imf . at low sfrs instead , due to the smaller number of stars that are formed , the probability to find massive stars decreases .
the imf is not fully sampled , and @xmath9 . in this regime ,
stochastic effects become important to describe the observed @xmath3 distribution , regardless of the imf functional form .
in addition to finite sampling in mass , @xmath3 is also affected by finite sampling in time since massive stars experience evolutionary phases of short duration ( e.g. wn or wc phases ) during which their ionizing luminosities can vary significantly . since the majority of stars is believed to form in embedded star clusters ( e.g. * ? ? ?
* ) , two additional effects should be considered .
first , the maximum stellar mass in a cluster can not exceed the cluster mass @xmath10 .
second , at low sfrs , a time - averaged continuous sfr results in a series of small bursts " separated in time and associated with the formation of a new cluster .
provided that stars form with a modest spread in ages , gaps in the ages of clusters increase the probability to observe a galaxy when the most massive stars have already left the main sequence . because it treats these types of stochastic sampling correctly
, slug is the ideal tool to test whether incomplete sampling of a canonical imf can reproduce the deficiency of h@xmath0 observed in low - luminosity galaxies without resorting to a modified imf . for
any given sfh , a fraction @xmath11 of the total stellar population is assumed to form in clusters . in this case , slug randomly draws a set of clusters from a cluster mass function ( cmf ) @xmath12 and populates them with stars randomly selected from an imf @xmath13 over some specified mass interval .
currently , all the stars within a cluster are assumed to be coeval .
the remaining fraction @xmath14 of the total stellar population are formed in the field , simulated by randomly selecting stars from an imf .
the ensemble of clusters and stars is then evolved with time and at each time step the luminosity of individual stars is combined to quantify the number of ionizing photons and @xmath15 for the simulated galaxy . in this work ,
we consider two classes of models .
the `` kroupa model '' is based on a universal @xcite imf in the mass interval @xmath16 m@xmath5 and a cmf with @xmath17 between @xmath18 m@xmath5 . to test the effects of clustering we run two sets of simulations , one with @xmath19 ( all stars in the field , similar to the monte carlo in l09 ) and one with @xmath20 ( all stars in clusters ) .
these extreme cases bracket all possible solutions , although the observations of @xcite suggest that reality is likely closer to @xmath20 than @xmath19 .
the `` igimf model '' is instead based on the integrated galactic initial mass function ( igimf ; * ? ? ?
* ) , a modified imf in which @xmath7 is a function of @xmath10 that , in turn , depends on the galaxy sfr .
the igimf implementation follows the minimal-1 and minimal-2 formulations @xcite , with a @xcite imf and a cmf with @xmath17 , minimum mass of @xmath21 m@xmath5 and maximum mass derived from the sfr averaged over @xmath22 year . in both the kroupa @xmath20 and the igimf model , the imf integrated over a galaxy is modulated by the galaxy star formation and the cmf . while in the @xmath20
model the relations between these quantities emerge from a pure stochastic realization of the imf and cmf , the igimf relies on prescriptions for these correlations that imply a truncation in the cmf and the imf . for both models ,
we adopt stellar libraries with the padova agb tracks at solar metallicity @xcite , the @xcite implementation of @xcite and @xcite atmospheres , and @xcite winds .
we run @xmath23 simulated galaxies randomly drawn from a schechter function , similar to the observed uv luminosity function @xcite , with sfr @xmath24 .
we use 10 time steps of 80 myr per galaxy to improve the statistics .
each of them is independent since the 80 myr interval ensures that all massive stars that form in one time step are gone by the next one . to further improve the statistics at higher luminosity (
observations are biased towards higher @xmath15 ) , we add 2000 galaxies uniformly distributed in logarithmic bins of sfr . due to escape or dust absorption before ionization , only a fraction @xmath25 of the lyman continuum photons produces ionizations that ultimately yield h@xmath0 photons . to account for this effect
, we correct the h@xmath0 luminosity in the models by a factor @xmath26 ( b09 ) .
although the exact value for @xmath25 is uncertain ( l09 ) , any choice in the range @xmath27 ( e.g. b09 , * ? ? ?
* ) would leave our basic results unchanged .
we further emphasize that reasonable choices of metallicity , stellar libraries , time steps , or even imf ( kroupa versus salpeter ) yield very similar results .
slug simulations and theoretical predictions are displayed in figure [ lumlum ] .
stochastic effects are evident in the distributions , with simulated galaxies scattered in proximity to the expected luminosities for a fully sampled imf .
since previous studies ( e.g. * ? ? ?
* ) have highlighted the importance of the minimum mass in the cmf ( @xmath28 ) for this type of calculations , we display in the top panel of figure [ lumlum ] the mean values for additional simulations that are similar to the @xmath20 model , but with @xmath29 .
the main result of this analysis is summarized in figure [ lumlum ] . at @xmath30 ,
observed galaxies lie close to the value of @xmath3 expected for both a fully - sampled kroupa imf and the igimf , but at fainter luminosity and lower sfr the data deviates from the fully - sampled kroupa imf curve towards the fully - sampled igimf curve .
this trend has been taken as evidence in support of the igimf , but both imfs are subject to stochastic effects . when we properly include these using slug , we see that realizations drawn from the igimf are completely inconsistent with a significant fraction of the observed galaxies , particularly at @xmath31 .
conversely , realizations drawn from a kroupa imf span a larger range of luminosity and overlap with most of the observed sample , with clustering responsible for a further increase in the luminosity spread ( compare the @xmath20 and @xmath19 models ) .
the width in the simulated distributions follows from the treatment of @xmath7 and clustering . for a universal imf
, @xmath7 can assume any value up to @xmath8 , regardless of the sfr .
at low sfrs , realizations that lack massive stars are frequent and skew the distribution to low @xmath32 and low @xmath15 . at the same time , realizations with massive stars are still possible and some models are distributed near or even above the theoretical expectation for a fully sampled imf .
the narrower scatter found for the @xmath19 model emphasizes that stochastic sampling of the imf alone can not reproduce the entire range of observed luminosity ( cf .
l09 ) . in the @xmath20 model ,
simulated galaxies that lie at the lowest @xmath32 for any given @xmath15 have an excess of older and massive clusters .
this is because at low sfr , a massive cluster represents a significant event in the galaxy sfh that , on average , increases the time interval between the formations of clusters and produces an intrinsic level of burstiness that results in a wider luminosity distribution ( see also * ? ? ?
this effect is obviously amplified in models with @xmath29 , as evident from the lower h@xmath33fuv luminosity in these simulations .
note that irregular sfhs are typical of dwarf galaxies @xcite and models with bursty star formation may reproduce the observed luminosity distribution equally well .
conversely , in the igimf theory , at lower sfrs , @xmath9 .
further , only clusters with low mass can be drawn .
the narrower mass range that is accessible translates in a narrower spread in luminosity , and none of the models can significantly exceed the expected luminosities for a fully sampled igimf ( see l09 ) .
ccccccc data & 10 & 37.1 & & 0.64 & & + kroupa @xmath20 & 241817 & 37.0 & 37.7 & 0.86 & 0.5995 & 0.5 + kroupa @xmath19 & 362593 & 37.3 & 37.7 & 0.55 & 0.5515 & 0.6 + igimf & 64253 & 35.6 & 36.4 & 0.72 & 0.0034 & 2.9 + data & 86 & 38.5 & & 0.49 & & + kroupa @xmath20 & 25571 & 38.3 & 38.7 & 0.55 & 0.1073 & 1.6 + kroupa @xmath19 & 24705 & 38.5 & 38.7 & 0.30 & 0.1205 & 1.6 + igimf & 9363 & 37.7 & 38.2 & 0.45 & 0.0000 & 6.9 + data & 153 & 39.5 & & 0.45 & & + kroupa @xmath20 & 4248 & 39.5 & 39.7 & 0.40 & 0.3524 & 0.9 + kroupa @xmath19 & 4238 & 39.6 & 39.7 & 0.29 & 0.0002 & 3.7 + igimf & 3576 & 39.2 & 39.6 & 0.37 & 0.0003 & 3.6 + data & 135 & 40.5 & & 0.35 & & + kroupa @xmath20 & 3494 & 40.6 & 40.7 & 0.34 & 0.4042 & 0.8 + kroupa @xmath19 & 3562 & 40.6 & 40.7 & 0.29 & 0.0090 & 2.6 + igimf & 3296 & 40.4 & 40.7 & 0.33 & 0.0365 & 2.1 + [ tab1 ] due to incompleteness , statistical comparisons between data and simulations are not straightforward .
this is complicated by the fact that , at any sfr , both h@xmath0 and fuv luminosities are subject to scatter and a truly independent variable is lacking .
nevertheless , in figure [ histo ] and table [ tab1 ] , we attempt to quantify the agreement between models and observations by comparing the statistics of the h@xmath0 distributions in intervals of @xmath15 .
considering the center of the h@xmath0 distributions , the igimf and @xmath20 realizations diverge , moving towards lower fuv luminosity . at the lowest @xmath15 ,
the two distributions are separated by @xmath34 dex , with the data clearly favoring the kroupa imf .
figure [ histo ] demonstrates that a truncated imf and cmf as in the igimf produce systematically lower @xmath32 than observed .
this is particularly evident for the interval @xmath35 that is characterized by large enough statistics ( @xmath36 galaxies ) and likely not affected by severe incompleteness . to quantify this claim ,
we perform a kolmogorov - smirnov test comparing the observed distribution with 10000 random sub - samples of the models , extracted to have a size comparable to the data .
we report the results in table [ tab1 ] .
we see that the observed data and the kroupa @xmath20 models are generally consistent with being drawn from the same parent distribution , while the hypothesis that the data and the igimf models originate from the same parent distribution can be ruled out .
the @xmath19 model also appears to be inconsistent with the data , though not by as much as the igimf model . finally , in figure [ chi ] we compare for each individual data point the observed @xmath37 with the simulated @xmath38 , averaged over an interval of fuv that is the larger of 0.1 dex and the error on the observed @xmath15 .
this difference is then normalized to @xmath39 where @xmath40 is the error on @xmath37 and @xmath41 is the standard deviation of the models .
this quantity is defined such that a distribution centered on zero with dispersion of unity indicates perfect agreement between models and data ( see right panels ) . in figure
[ chi ] , the center of the @xmath20 model ( top panel ) is consistent with the observed galaxies within 2@xmath42 for @xmath43 and within 3@xmath42 for brighter
luminosity , where @xmath40 becomes comparable to @xmath41 .
conversely , due to the narrower scatter , the @xmath19 model ( central panel ) is only partially consistent with observations , as previously suggested by the ks test
. also , simulated galaxies have higher @xmath32 compared to the observed galaxies . based on results of two independent tests , we favor the @xmath20 model over the @xmath19 simulations .
the igimf model ( bottom panel ) reproduces instead the observed @xmath32 only at high @xmath15 and there is a clear systematic offset between the data and the models at @xmath43 .
although the usual notion of probability associated to @xmath42 does not apply since the h@xmath0 distributions are not gaussian , we conclude that observations are better described by models based on the kroupa imf than on the igimf , particularly at fainter luminosity and when clusters are included .
data are very sparse below @xmath44 , where the separation between a universal imf and a steeper / truncated imf is most evident , precluding us from concluding that no variation in the imf occurs at very low luminosity .
moreover , we have not explored correlations with the galaxy surface brightness , suggested by m09 to be the physical quantity related to the imf variation . also , selection biases are not well characterized and a putative population of galaxies with @xmath45 and @xmath46 would call into question a universal imf .
further we assume that all stars born in an individual cluster are coeval , and future work will be required to investigate the effects of relaxing this assumption .
conversely , the observed existence of galaxies with @xmath47 and @xmath48 poses a direct challenge to the truncation in the imf and cmf as currently implemented in igimf theory .
as illustrated in figure [ lumlum ] , if the igimf model is correct , it should be impossible for galaxies to occupy this region of luminosity space .
the conclusion is strengthened by the fact that we have implemented only a minimal igimf .
the standard igimf would predict even less @xmath32 , exacerbating the discrepancy with the observations .
uncertainties in the measurements , particularly in the dust corrections , together with the fact that our models are highly idealized ( simple sfh , single metallicity , lack of any feedback ) , may explain the existence of galaxies non - overlapping with igimf models , but the discrepancy we found is most likely larger than the errors at faint @xmath32 .
using slug , a fully stochastic code for synthetic photometry in star clusters and galaxies , we have compared the h@xmath0 to fuv luminosities in a sample of @xmath1 nearby galaxies with models from a universal kroupa imf and a modified imf , the igimf .
our principal findings are : i ) simulated galaxies based on a kroupa imf and stochastic sampling of stellar masses are consistent with the observed @xmath32 distribution ; ii ) only models where stars are formed in clusters account for the full scatter in the observed luminosity ; iii ) realizations based on a truncated imf as currently implemented in the igimf underestimate the mean h@xmath0 luminosity .
based on this result , and since other factors not included in our simulations ( e.g. dust , escape fraction of ionizing radiation , bursty sfhs ) can mimic some of the features of the observed luminosity distribution ( b09 , l09 ) , we conclude that present observations of the integrated luminosity in nearby galaxies are consistent with a universal imf and do not demand a truncation at its upper end .
while we show that the current igimf implementation provides a poor description of available observations , our work is consistent with the fundamental idea behind the igimf , i.e. that the sfr , the cmf , and the imf jointly produce the observed luminosity distribution of galaxies .
however , our analysis emphasizes that the correlations between these quantities emerge naturally from stochastic sampling and that a further modification to the imf in clusters as proposed in the igimf model is not needed to account for the integrated luminosities in galaxies .
further , our calculation highlights how clusters introduce the level of burstiness required to fully account for the observed luminosities , owing to the combined effects of the cluster age distribution and the short life - time of massive stars .
time dependence might be a key element currently missing in models of the integrated imf in galaxies .
we are most grateful to a.boselli and g.meurer for sharing their data and for valuable comments .
we acknowledge useful discussions with m.cervio , j.lee , p.kroupa , j.scalo , x.prochaska , j.werk , g.gavazzi and k.schlaufman .
we thank f.bigiel for motivating us to write slug .
rlds is supported under a nsf graduate research fellowship .
mrk acknowledges support from : alfred p. sloan fellowship ; nsf grants ast-0807739 and career-0955300 ; nasa astrophysics theory and fundamental physics grant nnx09ak31 g ; _ spitzer space telescope _ theoretical research program grant .
bastian , n. , covey , k. r. , & meyer , m. r. 2010 , , 48 , 339 boselli , a. , boissier , s. , cortese , l. , buat , v. , hughes , t. m. , & gavazzi , g. 2009 , , 706 , 1527 bressan , a. , fagotto , f. , bertelli , g. , & chiosi , c. 1993 , , 100 , 647 calzetti , d. , chandar , r. , lee , j. c. , elmegreen , b. g. , kennicutt , r. c. , & whitmore , b. 2010 , , 719 , l158 cervio , m. , & valls - gabaud , d. 2003 , , 338 , 481 chabrier , g. 2003 , , 115 , 763 corbelli , e. , verley , s. , elmegreen , b. g. , & giovanardi , c. 2009 , , 495 , 479 da silva , r. l. , fumagalli , m. , & krumholz , m. 2011 , arxiv:1106.3072 eldridge , j. j. 2011 , arxiv:1106.4311 fumagalli , m. , da silva , r. , krumholz , m. , & bigiel , f. 2011 , astronomical society of the pacific conference series , 440 , 155 gunawardhana , m. l. , et al .
2011 , , in press , arxiv:1104.2379 haas , m. r. , & anders , p. 2010
, , 512 , a79 hillier , d. j. , & miller , d. l. 1998 , , 496 , 407 hirashita , h. , buat , v. , & inoue , a. k. 2003 , , 410 , 83 hoversten , e. a. , & glazebrook , k. 2008 , , 675 , 163 kennicutt , r. c. , jr .
1998 , , 36 , 189 kroupa , p. 2001 , , 322 , 231 kroupa , p. , & weidner , c. 2003 , , 598 , 1076 lada , c. j. , & lada , e. a. 2003 , , 41 , 57 lamb , j. b. , oey , m. s. , werk , j. k. , & ingleby , l. d. 2010 , , 725 , 1886 lee , j. c. , et al .
2009 , , 706 , 599 lee , j. c. , gil de paz , a. , tremonti , c. , kennicutt , r. , & the local volume legacy team 2010 , arxiv:1011.2181 lejeune , t. , cuisinier , f. , & buser , r. 1997 , , 125 , 229 maeder , a. , & meynet , g. 1988 , , 76 , 411 meurer , g. r. , et al .
2009 , , 695 , 765 pflamm - altenburg , j. , weidner , c. , & kroupa , p. 2007
, , 671 , 1550 pflamm - altenburg , j. , weidner , c. , & kroupa , p. 2009
, , 395 , 394 salpeter , e. e. 1955 , , 121 , 161 smith , l. j. , norris , r. p. f. , & crowther , p. a. 2002 , , 337 , 1309 weidner , c. , & kroupa , p. 2006
, , 365 , 1333 weidner , c. , kroupa , p. , & pflamm - altenburg , j. 2011 , , 412 , 979 weisz , d. r. , skillman , e. d. , cannon , j. m. , et al .
2008 , , 689 , 160 williams , j. p. , & mckee , c. f. 1997 , , 476 , 166 wyder , t. k. , et al .
2005 , , 619 , l15 | recent observations indicate a lower h@xmath0 to fuv ratio in dwarf galaxies than in brighter systems , a trend that could be explained by a truncated and/or steeper imf in small galaxies .
however , at low star formation rates ( sfrs ) , the h@xmath0 to fuv ratio can vary due to stochastic sampling even for a universal imf , a hypothesis that has , prior to this work , received limited investigation . using slug , a fully stochastic code for synthetic photometry in star clusters and galaxies , we compare the h@xmath0 and fuv luminosity in a sample of @xmath1 nearby galaxies with models drawn from a universal kroupa imf and a modified imf , the integrated galactic initial mass function ( igimf ) .
once random sampling and time evolution are included , a kroupa imf convolved with the cluster mass function reproduces the observed @xmath2 distribution at all fuv luminosities , while a truncated imf as implemented in current igimf models underpredicts the @xmath2 luminosity by more than an order of magnitude at the lowest sfrs .
we conclude that the observed luminosity is the result of the joint probability distribution function of the sfr , cluster mass function , and a universal imf , consistent with parts of the igimf theory , but that a truncation in the imf in clusters is inconsistent with the observations .
future work will examine stochastic star formation and its time dependence in detail to study whether random sampling can explain other observations that suggest a varying imf . |
the dirac equation for relativistic electrons , in spite of its fundamental importance in physics , is far from being thoroughly understood @xcite .
one of its special features is that , even in the absence of fields , there exists a spectrum of positive and negative electron energies .
this peculiarity is a source of various problems .
one of them is a phenomenon of zitterbewegung ( trembling motion ) which arises from an interference of positive and negative energy states @xcite .
there exist attempts to circumvent this duality .
one of them is the foldy and wouthuysen transformation ( fwt ) which , for the case of no external fields , allows one to break the dirac equation into separate equations for positive and negative energies @xcite . in their original paper ,
fw remarked that a functional kernel , which transforms functions from the original representation to the fw representation , is characterized by a non - locality in coordinate space of the order of the compton wavelength @xmath0
. rose in his book @xcite put this statement on a quantitative basis by showing that the second moment of the kernel is equal to @xmath1 ( see below ) .
this result was not followed by other investigations and it is by now not well known .
also , the moments are often used in statistical physics but they are a rather unorthodox way to characterize quantum mechanical properties .
it was shown later that the fw transformation is not unique .
in other words , there exist other transformations capable of separating positive and negative energy states in the dirac equation , both in the absence of fields and in the presence of an external magnetic field @xcite . in the following we consider , in addition to the fwt , a two - step transformation proposed by moss and okninski ( mo , ref .
@xcite ) .
let us consider , as a matter of example , an average value of the time - dependent velocity operator @xmath2 in the heisenberg picture , @xmath3 where @xmath4 , @xmath5 is the dirac hamiltonian and @xmath6 is an arbitrary dirac spinor .
according to the dirac equation one has @xmath7 , where @xmath8 ( @xmath9 ) are the standard dirac @xmath10 matrices .
as is well known , @xmath8 do not commute with @xmath5 , so the velocity depends on time also in the absence of fields , which results in the zitterbewegung mentioned above .
the operator @xmath11 in eq .
( [ iv(t ) ] ) stands for a unitary transformation , for example the fwt .
equation ( [ iv(t ) ] ) expresses the well known fact that a unitary transformation does not change an average value . in our case
it means that a transformation will not eliminate the zitterbewegung , which makes physical sense .
we wrote down eq .
( [ iv(t ) ] ) to illustrate symbolically that , if one transforms the operators : @xmath12 , one should also transform the functions : @xmath13 .
important operators transformed according to fwt are given in the original fw paper and quoted in many textbooks . here , we concentrate on the transformed functions which were not analyzed in the literature .
our work has two objectives .
the first is to characterize in various ways the non - locality of the foldy - wouthuysen transformation for the dirac equation . in doing this
we continue the work of rose but use means more typical for quantum mechanics .
the second objective is to investigate , as a matter of example , another separating `` transformation and to compare its properties to those of fwt .
such a comparison should give an idea of what one can expect of various ' ' separating transformations .
our subject is of relevance for two reasons .
first , with the rise of graphene `` there is nowadays a great deal of interest in relativistic - type wave equations @xcite .
second , it is now possible to simulate the dirac equation with the trapped ions and cold atoms interacting with the laser radiation , where one can tailor much more ' ' user friendly `` values of the basic parameters @xmath14 and @xmath15 @xcite . in fact , a proof - of - principle experiment simulating 1 + 1 dirac equation and the resulting zitterbewegung was carried out by gerritsma _
et al . _
if one deals with effective parameters by employing narrow - gap semiconductors , the non - locality is determined by an ' ' effective `` compton wavelength @xmath16 , where @xmath17 is the effective electron mass and @xmath18 is the maximum electron velocity . in narrow - gap semiconductors
one can deal with @xmath19 and @xmath20 cm / s , so that the non - locality can be a few orders of magnitude larger than that in a vacuum , see ref .
the band structure of graphene corresponds to the ' ' ultra - relativistic case ( no gap ) and a similar maximum velocity @xmath18 , so the non - locality may be even larger .
in this section we establish forms of foldy - wouthuysen and moss - okninski transformations and define corresponding transformation kernels .
next we characterize nonlocal properties of the kernels by calculating their moments .
the dirac equation for a free relativistic electron is @xmath21 , where @xmath22 is the momentum and @xmath23 and @xmath24 are @xmath10 dirac matrices in the standard notation .
the eigenenergies are @xmath25 with @xmath26 . for the field - free case
the momentum components are good quantum numbers @xmath27 . in the fw transformation an initial function @xmath6
is transformed to the new representation @xmath28 with the use of the operator @xmath29 defined as @xmath30 there is @xmath31 . to find the mo transformation , we follow moss and okninski @xcite and introduce first the unitary operator @xmath32 which transforms the dirac equation to the form @xmath33 , in which @xmath34 the hamiltonian @xmath35 has zeroes on the diagonal . for a four - component function @xmath36 there is @xmath37 . the two - step mo transformation is defined using the @xmath35 hamiltonian in the following way @xmath38 in the field - free case the operators @xmath29 and @xmath39 are number matrices . following foldy and wouthuysen
we introduce a function transformation from the old to the new representation @xmath40 . inserting the unity operators we obtain @xmath41 which defines the transformation kernel @xmath42 in eqs .
( [ uk_phi ] ) and ( [ uk_k ] ) @xmath11 stands for each of the two operators defined above and @xmath43
. the kernels are @xmath10 matrices .
as mentioned above , the transformation kernels have a nonlocal character .
we also remarked that rose @xcite described this feature quantitatively by calculating the second moment of the fw kernel and showing that its smearing is given by the compton wavelength @xmath44 .
below we evoke his calculations , as it is not easily accessible by now , and compare his result with the corresponding quantities for the mo transformation .
the moments are rarely used to describe quantum mechanical behavior , for this reason we do not limit our subsequent considerations to the moments but also consider other quantities . all the elements of matrices defined in eq .
( [ uk_k ] ) are nonlocal functions of @xmath45 . to estimate the non - locality we calculate the zeroth and second moments @xmath46 of @xmath47 and @xmath48 matrices @xmath49 let us consider first the @xmath50 element of @xmath47 .
for the zeroth moment there is @xmath51 changing the order of integration one has @xmath52 .
then the integration over @xmath53 is trivial and one obtains @xmath54 .
the second moment of the @xmath50 element is @xmath55 integrating twice by parts one obtains @xmath56 d^3{\bm p } \nonumber \\ & = & -\hbar^2 \nabla^2 \left . \left
[ \frac{e_p+mc^2}{\sqrt{2e_p(e_p+mc^2 ) } } \right ] \right|_{{\bm p } = { \bm 0 } } = \frac{3}{4}\lambda_c^2.\end{aligned}\ ] ] in the same way one calculates moments of the remaining matrix elements .
we obtain @xmath57 and @xmath58 , where @xmath59 is the identity matrix .
these are the results of rose @xcite .
thus the @xmath60 kernel is smeared in the @xmath61 space by the amount of @xmath62 in each direction .
the kernel can be considered to be a transformed dirac delta function , cf .
( [ uk_phi ] ) .
hence , one can subscribe to the statement of foldy and wouthuysen that `` a wave function which in the old representation corresponds to a state in which the particle was localized at one point , passes over in the new representation into a wave function which corresponds to the particle being spread out over a finite region '' .
the physical meaning of this result is discussed in refs .
@xcite .
we carry similar calculations for the moss - okninski transformation and calculate the zeroth moment of the @xmath63 element of @xmath39 @xmath64 for the second moment of @xmath63 element of @xmath39 we integrate twice by parts and obtain @xmath65 d^3{\bm p } \nonumber \\ & = & -\hbar^2 \nabla^2 \left .
\left[\frac{imc^2+cp_z}{\sqrt{2}e_p } \right ] \right|_{{\bm p } = { \bm 0 } } = \frac{3i\lambda_c^2}{\sqrt{2}}.\end{aligned}\ ] ] calculating the remaining moments one finally obtains @xmath66 thus the kernels of both fw and mo transformations have finite non - localities of the order of @xmath44 , but they differ somewhat from each other . in this sense
the fw transformation is somewhat more `` compact '' than the mo transformation .
we further investigate properties of fw and mo transformations by studying the transformed functions .
let us consider an initial wave function @xmath67 , where @xmath68 is normalized .
then the transformed function for the mot is @xmath69 , in which @xmath70 where @xmath71 the prime in @xmath72 means that we transform the function @xmath73 , see eq .
( [ up_v ] ) .
if the initial function @xmath74 is transformed according to fwt , there is @xmath75 where @xmath76 .
the asymmetry between two upper and two lower components of @xmath77 arises from the asymmetry of components in the initial wave packet .
let us consider first the initial function in form of the delta function : @xmath78 .
then @xmath79 and the transformed function is obtained from eqs .
( [ fk_psi_1 ] ) and ( [ fk_psi_2 ] ) in terms of four integrals @xmath80 , @xmath81 , @xmath82 , @xmath83 @xmath84 the first integral is @xmath85 where @xmath86 is the modified bessel ( macdonald ) function . to get eq .
( [ fk_d0 ] ) we used identities ( [ ak0 ] ) and ( [ ak1r ] ) in appendix .
the second integral is @xmath87 this integral is divergent , so we separate it into the divergent and convergent parts and obtain @xmath88 \nonumber \\ & = & \frac{2\lambda_c}{z}\delta({\bm \rho } ) + { \cal a}({\bm r}/\lambda_c),\end{aligned}\ ] ] where @xmath89 .
we marked the integral in the second line of eq .
( [ fk_dz ] ) by @xmath90 .
this integral is convergent since the integrand has no singularities and for large @xmath91 it decreases as @xmath92 .
the integral @xmath90 can be expressed in terms of the anger functions and it decays exponentially with @xmath93 with a characteristic length @xmath44 .
the integrals @xmath81 and @xmath82 can be obtained from @xmath83 replacing @xmath94 by @xmath95 and @xmath96 . in conclusion
, it can be seen that the delta function subjected to the mo transformation becomes a function _ having a finite width of the order of @xmath44_. similar calculations can be performed for a function @xmath97 subjected to the fw transformation . the transformed function is a combination of four integrals : @xmath98 , see eq .
( [ fk_phi ] ) . setting @xmath79 and
separating out the divergent part we obtain @xmath99 + 1\right\ } \nonumber \\ & = & \frac{\delta({\bm r})}{\sqrt{2 } } + \int \frac{d^3{\bm k } } { \sqrt{2}(2\pi)^3 } \frac{e^{i{\bm k}{\bm r}}}{\sqrt{1+\lambda_c^2k^2 } } g(k),\end{aligned}\ ] ] where @xmath100 .
we changed @xmath101 to @xmath102 .
the function @xmath103 varies slowly from @xmath104 to @xmath105 .
thus we may approximate @xmath106 by setting @xmath103 to have a constant value @xmath107 .
this gives @xmath108 this final result is very similar to that for the mot , see eqs .
( [ fk_ddelta ] ) and ( [ fk_d0 ] ) .
the integrals @xmath109 , @xmath110 and @xmath111 can be calculated in a similar way to that described for the mo transformation .
thus for both mot and fwt the initial delta function transforms into functions of finite width of the order of @xmath44 . in the above analysis with the initial delta function
the interpretation of results is somewhat difficult because of divergent and singular final integrals . to avoid these problems we take the initial function in the form of a gaussian packet having a finite width @xmath112 @xmath113
the packet is normalized according to @xmath114 .
now there is @xmath115 and all the integrals in eqs .
( [ fk_psi_2 ] ) and ( [ fk_phi ] ) are quickly convergent .
the transformed function is again expressed in terms of four integrals @xmath116 , @xmath117 , @xmath118 and @xmath119 @xmath120 the first integral is [ see eq .
( [ fk_psi_2 ] ) ] @xmath121 integrating over the angular variables in the spherical coordinates we obtain @xmath122 applying the identity : @xmath123 , integrating over @xmath124 with the use of formula ( [ aintsin ] ) , we get @xmath125 the presence of @xmath126 term in eq .
( [ fk_t0eta ] ) reduces the range of integration to @xmath127 . in the limit of @xmath128
we can neglect @xmath129 as compared to @xmath130 and obtain @xmath131 thus , there is no widening of the wave packet in the limit of large widths @xmath112 , and the transformed packet is almost identical to the initial one . in the opposite limit of very narrow packets
: @xmath132 , we may neglect in eq .
( [ fk_t0eta ] ) the @xmath130 term under the integral sign . after the substitution @xmath133 one
gets @xmath134 using identities ( [ aintd0 ] ) in appendix and @xmath135 we finally obtain @xmath136 thus in the limit of narrow packets the transformed function acquires a width of the order of @xmath137 . for all values of @xmath112 the width of the transformed function is larger than the width of the initial packet . for @xmath119 integrals in eq .
( [ fk_dgauss ] ) we have @xmath138 integrals @xmath117 and @xmath118 are obtained the same way .
they have similar properties to @xmath116 integral .
it is seen that @xmath116 of eq .
( [ fk_t0_smalld ] ) differs from @xmath80 of eq .
( [ fk_d0 ] ) by a factor of @xmath139 .
this is due to a different normalizations of the packet @xmath68 and the delta function . if one normalizes the packet according to @xmath140 the results of eqs .
( [ fk_t0_smalld ] ) and ( [ fk_d0 ] ) become the same . for the fw transformation ,
a function @xmath74 is transformed to @xmath141 , where @xmath142 , @xmath143 , @xmath144 and @xmath145 are integrals defined below .
there is [ see eq .
( [ fk_phi ] ) ] @xmath146 where @xmath147 and @xmath148 is defined in eq .
( [ fk_fp ] ) .
the integral @xmath143 is @xmath149 and similarly for @xmath144 and @xmath145 .
integrals @xmath143 , @xmath144 and @xmath145 can be obtained as partial derivatives of an auxiliary integral @xmath150 with respect to @xmath95 , @xmath96 and @xmath94 , respectively .
( mot ) and @xmath77 ( fwt ) calculated for initial packet width @xmath151 .
dotted line @xmath152 the initial gaussian wave packet.,width=302,height=302 ] in figure [ fig1fun ] we plot various elements of the transformed wave functions : the integrals @xmath153 , @xmath154 , @xmath155 and @xmath156 calculated for @xmath151 .
solid lines represent @xmath155 and @xmath156 integrals , while dashed lines represent @xmath153 and @xmath154 .
the factor @xmath157 in front of @xmath153 and @xmath154 integrals is chosen to cancel out the factor @xmath158 appearing in eq .
( [ fk_psi_2 ] ) .
the terms @xmath154 and @xmath159 are related to @xmath160 terms in eqs .
( [ fk_psi_2 ] ) and ( [ fk_phi ] ) .
the integral @xmath155 is very similar the initial gaussian function .
for the mot the difference between the initial and the transformed functions is more pronounced . for larger
@xmath112 , the integrals @xmath155 and @xmath161 tend to the initial gaussian function , so almost no widening occurs .
this agrees with the the large-@xmath112 approximation for @xmath161 in eq .
( [ fk_t0_larged ] ) .
finally , we estimate the spatial extent of functions subjected to the mo and fw transformations by calculating their normalized variances . for the mot we define @xmath162 and similarly for the fwt . functions @xmath72 and @xmath163 are normalized to unity . we assume the initial function in the form of a gaussian packet given in eq .
( [ fk_gauss ] ) . since @xmath164 ,
see eqs .
( [ fk_psi_1 ] ) and ( [ fk_psi_2 ] ) , it is convenient to introduce averages over the first and second parts of @xmath72 .
we define @xmath165 with @xmath166 .
then @xmath167 there is @xmath168 and @xmath169 .
we consider now @xmath170 terms .
introducing @xmath171 we have @xmath172 the dagger denotes hermitian conjugate of the vector and the star its complex conjugate . in the above equation the terms proportional to @xmath14 cancel out .
the terms including @xmath160 are odd functions of @xmath160 , so they vanish after the integration . for the same reasons there is @xmath173 . to calculate @xmath174
we first introduce the fourier transform of @xmath175 @xmath176 where @xmath177 is the column in eq .
( [ fk_psi_2 ] ) or the second column in eq .
( [ v_rn12 ] ) .
then @xmath178 since @xmath179 and @xmath180 , the function @xmath72 is really normalized to unity .
now we calculate @xmath181 @xmath182 integrating twice by parts over @xmath53 and then integrating over @xmath183 and @xmath184 one obtains @xmath185 after some manipulation we find @xmath186,\ ] ] where @xmath187 , and @xmath188 for integer values of @xmath189 and integer or half - integer values of @xmath190 the integrals @xmath191 can be expressed in terms of modified bessel and error functions , see appendix . for the fw transformation
we have similarly @xmath192 where @xmath193 in which @xmath194 .
after some algebra we obtain @xmath195.\ ] ] .
solid line @xmath152 fwt , dashed line @xmath152 mot .
dotted horizontal line @xmath152 normalized variance of the initial gaussian wave packet .
the mot and fwt variances has the small-@xmath112 limit @xmath196 , while for large @xmath112 they reach the gaussian limit @xmath197.,width=302,height=302 ] in figure [ fig2v ] we plot calculated variances of the transformed functions divided by @xmath130 .
for the initial gaussian packet the variance is @xmath198 and it is indicated by the horizontal dotted line .
the other two variances [ mo - transformed function @xmath72 and fw - transformed function @xmath77 ] have very similar dependencies on the packet width @xmath112 .
both variances have the zero - width limit @xmath200 and they decrease with increasing @xmath112 to the gaussian limit @xmath201 .
small-@xmath112 and large-@xmath112 limits of the variances can be obtained analytically using eqs .
( [ va_r2_22 ] ) and ( [ va_r2_fw ] ) .
consider first the variance for @xmath181 from eq .
( [ va_r2_22 ] ) . for small @xmath202 the second term in eq .
( [ va_r2_22 ] ) tends to zero ( see appendix ) , and there is @xmath203 .
the first term is @xmath169 and the total variance is @xmath204 .
similarly , for the fw transformation at small @xmath205 the second term in eq .
( [ va_r2_fw ] ) vanishes and the total variance is again @xmath196 .
the large-@xmath202 limits of @xmath191 in eqs . ( [ va_r2_22 ] ) and ( [ va_r2_fw ] ) are calculated in appendix .
applying these results to eqs .
( [ va_r2_22 ] ) and ( [ va_r2_fw ] ) we find that the large-@xmath112 limits of @xmath206 and @xmath207 are @xmath197 , i.e. they are equal to the variance of the initial gaussian function .
thus the widening of the transformed functions is pronounced for @xmath208 and it is small for @xmath209 , see figure [ fig2v ] .
it can be seen that , again , the fw transformation is more compact than the mo transformation because for a given packet width @xmath112 there is always @xmath210 .
as mentioned in the introduction , rose @xcite demonstrated that a non - locality of the functional kernel generated by the foldy - wouthuysen transformation is extended in the coordinate space over the compton wavelength @xmath44 .
we generalize this result to the transformed functions and another transformation separating positive and negative energy states in the dirac equation ( moss - okninski transformation ) .
in particular , we show that the delta function @xmath211 is transformed into a nonlocal function smeared over the distance @xmath44 .
second - order variances are used to obtain similar results for narrow gaussian wave packets indicating that the non - locality of the order of @xmath44 is a general property of the transformed functions .
our results strongly suggest that all separating transformations for the energies generate coordinate non - localities of this order .
we emphasize that for relativistic - type equations , appearing either in narrow - gap semiconductors or in simulations , the non - localities can extend over hundreds of angstroms .
we first quote formulas for the integrals appearing in the text .
let @xmath212
. then @xmath213d\chi & = & 2\left(\frac{\mu}{2}\right)^{\nu}k_{-\nu}(\mu ) , \ \ \ \label{aintd0}\end{aligned}\ ] ] where @xmath214 , @xmath215 and @xmath216 are the modified bessel ( macdonald ) functions .
next we collect formulas for @xmath191 integrals .
let @xmath217 and @xmath218 be the error function .
then @xmath219}{2 } , \\ a_2
^ 0&=&\frac{\pi \exp(\bar{d}^2)(2\bar{d}^2 - 1)[{\rm erf}(\bar{d})-1]}{4 } + \frac{\sqrt{\pi}\bar{d}}{2}. \ \ \ \ \ \\end{aligned}\ ] ] for @xmath220 there is @xmath221 , because in this limit there is @xmath222 @xcite .
also , @xmath223 and @xmath224 because @xmath225 . using large-@xmath95 expansions for the bessel and error functions we obtain @xmath226 99 p. a. m. dirac , proc .
a * 117 * , 610 ( 1928 ) .
e. schrodinger , sitzungsber .
* 24 * , 418 ( 1930 ) .
schrodinger s derivation is reproduced in a. o. barut and a. j. bracken , phys .
d * 23 * , 2454 ( 1981 ) . j. d. bjorken and s. d. drell , _ relativistic quantum mechanics _ ( mcgraw - hill , new york,1964 )
. j. j. sakurai _ modern quantum mechanics _
( addison - wesley , new york , 1987 ) . w. greiner _
relativistic quantum mechanics _ ( springer , berlin , 1994 ) .
l. l. foldy and s. a. wouthuysen , phys . rev . * 78 * 29 ( 1950 ) .
m. e. rose _ relativistic electron theory _
( wiley , new york , 1961 ) . k. m. case , phys . rev . * 95 * 1323 ( 1954 ) . e. de vries , fortschritte der physik * 18 * 149 , ( 1970 ) . w. y. tsai , phys . rev .
d * 7 * 1945 ( 1973 ) .
d. l. weaver , phys .
d * 12 * 4001 ( 1975 ) .
r. e. moss and a. okninski , phys .
d * 14 * , 3358 ( 1976 ) .
k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , y zhang , s. v. dubonos , i. v. grigorieva and a. a. firsov , science * 306 * , 666 ( 2004 ) .
w. zawadzki , phys .
b * 72 * , 085217 ( 2005 ) .
w. zawadzki and t. m. rusin , j. phys .
matt . * 23 * , 143201 ( 2011 ) .
d. leibfried , r. blatt , c. monroe , and d. wineland , rev .
phys . * 75 * , 281 ( 2003 ) .
l. lamata , j. leon , t. schatz , and e. solano , phys .
* 98 * , 253005 ( 2007 ) .
m. johanning , a. f. varron , and c. wunderlich , j. phys .
b * 42 * , 154009 ( 2009 ) .
t. m. rusin and w zawadzki , phys .
d * 82 * 125031 ( 2010 ) .
r. gerritsma , g. kirchmair , f. zahringer , e. solano , r. blatt and c. f. roos , nature * 463 * 68 ( 2010 ) . i. s. gradshtein and i. m. ryzhik 2007 _ table of integrals , series , and products ( ed .
a jeffrey and d zwillinger 7th edition)_(academic press , new york , 2007 ) . | non - localities of foldy - wouthuysen and related transformations , which are used to separate positive and negative energy states in the dirac equation , are investigated .
second moments of functional kernels generated by the transformations are calculated , the transformed functions and their variances are computed .
it is shown that all the transformed quantities are smeared in the coordinate space by the amount comparable to the compton wavelength @xmath0 . |
dna is best known for its essential role in biology as it carries the genetic information of living beings @xcite . during the last few decades ,
the possibility of using dna in nanotechnology has been gaining attention from chemists , physicists , as well as engineers .
the electronic properties of dna can lead to applications in many areas .
researchers have shown that dna has the potential to be used as circuit components in electrical devices such as double barrier resonant tunneling structures @xcite , spin specific electron conductor @xcite , field effect transistor @xcite and negative differential resistance device @xcite .
in addition , the investigation of the electronic properties of dna can bring new techniques for the third generation sequencing technologies .
current sequencing methods are mainly based on light emission @xcite , which requires polymerase chain reaction ( pcr ) amplification .
an alternative sequencing idea is to measure the transverse current through a single - stranded dna as it translocates a nanogap or nanopore @xcite .
the distinguishable electronic structures of the four dna bases result in different electrical / ionic currents , which can be used in the third generation sequencing technologies .
recent experiments also reveal the possibility of detecting diseases by measuring the electrical conductance of dna strands with methylated cytosine bases @xcite , which may open new perspectives in medical application .
cytosine methylation refers to the replacement of the hydrogen atom by a methyl group at the 5th position of the pyrimidine ring @xcite .
this small modification has been found to be important in changing physical properties of dna .
for example , it can increase the melting temperature of the duplex @xcite , the molecular polarizability of the pyrimidine @xcite and enhance the stability of the double strands .
recent study has also revealed that the methylation of cytosine bases has an essential effect on the separation of dna strand @xcite .
more importantly , it plays a crucial role in many biological and cellular processes such as gene expression regulation @xcite , cellular differentiation @xcite , x - chromosome inactivation @xcite , embryonic development @xcite and genomic imprinting @xcite . as a consequence
, an increasing number of human diseases @xcite , including cancer @xcite , have been found to be associated with the abnormal activities of dna methylation .
targeting and identifying the methylated dna is important for disease diagnosis and drug design .
the main detection techniques used today involve pcr - based methods @xcite , which usually require chemical reactions @xcite to maintain the methylation feature . with the rapid development and progress in single - molecule electrical approach @xcite ,
new techniques utilizing the electronic properties of dna bases have also been used to experimentally detect the methylated strands @xcite .
these pioneering initial works show that it may be possible to detect methylated sequence / base by direct conductance measurement . despite the recent progress in experiment ,
understanding charge transport through a methylated dna sequence / base from a theoretical perspective is lacking . in this work ,
we study charge transport through an eight base - pair methylated dna strand and its native counterpart used in ref . .
we aim to provide an understanding of the effect of cytosine methylation on charge transport through dna strand . to this end
, we use the combination of density functional theory ( dft ) and the landauer - buttiker approach .
dft is a powerful tool to obtain the accurate description of the electronic structure of molecules and materials .
however , it is well known that in dft calculations the results may depend on the choice of exchange - correlation ( xc ) functionals and basis sets .
in addition , the lack of the derivative discontinuity and self - interaction errors in standard functionals like ggas and hybrid - ggas can lead to erroneous single - molecule transport predictions @xcite . in this paper
, we also study the effects of methylation on the eight - base pair dna strand with the b3lyp and cam - b3lyp functionals . the remainder of the paper is organized as follows : we first present the method and model used in this work - we review the extraction of the tight - binding parameters from @xmath1 calculation which help formulate a physical understanding of the charge transport properties .
then , we discuss the green s function approach , landauer - b@xmath0ttiker framework and the implementation of b@xmath0ttiker probes using damoto - pastawski model .
next , we show the results from our calculations .
we first look at the effect of methylation on the tight - binding parameters , including the on - site energies of the guanine and cytosine bases and hopping integrals between the first and second nearest neighboring bases .
we further investigate the influence of cytosine methylation on the electronic properties of dna by studying the transmission , conductance in both phase - coherent and decoherent cases .
we also study the role of xc functionals on dna charge transport .
then , we investigate the effect of contact coupling to understand the effects of the delocalization and self - interaction errors from xc functionals on the transmission . finally , we summarize our findings in the last section .
motivated by the measurement in ref . , we choose the following two dna sequences , each of which has eight base pairs : 1 ) the native strand , 5-gcgcgcgc-3 ( gc8 ) , where g refers to the guanine base and c refers to the native cytosine base ; 2 ) the methylated counterpart , 5-gcmgcmgcmgcm-3 ( gcm8 ) , where cm refers to the methylated cytosine base .
our approach involves two steps : ( 1 ) using dft calculations to obtain the hamiltonian and overlap matrices of dna strands and ( 2 ) using the landauer - b@xmath0ttiker method and green s function approach to compute the transmission and linear response conductance of the strands .
all the dft calculations are carried out with gaussian 09 @xcite software package .
we first generate the atomic coordinates of the b - form dna strands with nucleic acid builder @xcite .
because the conduction through dna strands is mainly caused by the overlap of @xmath2 orbitals in the nitrogenous bases @xcite and our prior work shows that in the vicinity of the highest occupied molecular orbital ( homo ) level the transport is mainly through the bases @xcite , we delete the backbone to reduce the expensive computational cost and terminate the nitrogenous bases with hydrogen atoms .
for the methylated strands , molecular dynamics simulations show that the methyl groups do not have large influence on the watson - crick base pairing for b - form dna @xcite .
thus , we assume that the methylated strand maintains the same b - form as the native one .
we first replace the hydrogen atom in the 5th position with a -ch3 methyl group , then relax the methyl group at b3lyp/6 - 31g(d ) @xcite level to get the optimized structure of the methylated cytosine .
next , we carry out dft calculations using b3lyp xc functional and 6 - 31 g basis set to determine the hamiltonian and overlap matrices @xmath3 and @xmath4 .
after the hamiltonian and overlap matrices are found , we apply l@xmath5wdin transformation @xcite to the system to orthogonalize the basis , @xmath6 each diagonal sub - block of @xmath7 corresponds to the representation of a single base .
we then find the eigenvalues and eigenvectors of the sub - blocks of @xmath7 and perform a unitary transformation of the hamiltonian @xmath7 to @xmath8 via , @xmath9 where the matrix @xmath10 is constructed from the eigenvectors of the diagonal sub - blocks of @xmath7 .
the diagonal sub - blocks of @xmath8 are the eigenvalues corresponding to each dna base and off - diagonal elements represent the interaction between different localized energy levels . the representation of @xmath8 can be truncated to obtain the tight - binding parameters such as the on - site energies of a single dna base and the hopping integrals between two neighboring dna bases for the interested energy levels . here we focus on analyzing the methylation effect on the tight - binding parameters at homo level and the corresponding hamiltonian @xmath11 is defined with the following equation , @xmath12 in the above equation , @xmath13 ( @xmath14 ) is the creation ( annihilation ) operator at the @xmath15th base ( here , i = 1 , 2 , ... 16 ) .
@xmath16 is the hermitian conjugate .
@xmath17 is the on - site energy of the @xmath15th base .
@xmath18 is the intra - strand coupling between two nearest neighboring bases within one strand .
@xmath19 is the inter - strand coupling between two nearest neighboring bases from the two complementary strands ( i. e. , the coupling for two bases within one base pair ) and @xmath20 is the inter - strand coupling between two second nearest neighboring bases from the two strands .
the notation @xmath21 and @xmath22 in the summation depict the two bases in the first and second nearest position , respectively @xcite . in fig .
[ interaction ] , we illustrate the interactions in the tight - binding model with gc8 where the contacts are shown by yellow blocks at the 3 ends as indicated .
the arrows represent the hopping between two neighboring bases both within and between strands , with red color for the nearest intra - strand hopping , purple for the nearest inter - strand hopping , green for the second nearest inter - strand hopping from 5 to 5 ends and blue for the second nearest inter - strand hopping from 3 to 3 ends .
the tight - binding parameters @xmath17 , @xmath18 , @xmath19 and @xmath20 have also been labeled in the corresponding position . as we will see in the later discussion ,
the inter - strand couplings for two second nearest neighboring bases , @xmath20 , depends on the orientation of the strand .
the hamiltonian in eq .
( [ h_tb ] ) can be used to study the properties of a larger system compared to the full hamiltonian approach since it only considers the contribution from the homo level . while this model can capture the essential physics of the system , to get a comprehensive understanding , the full hamiltonian is preferable . in this work ,
we use full hamiltonian shown in eq .
( [ hamiltonian ] ) to compute the transport properties of the dna strands , while the tight - binding hamiltonian in eq .
( [ h_tb ] ) is used to further our understanding of the charge transport mechanism in dna .
the electronic transport properties of the molecule are calculated within the landauer - b@xmath0ttiker framework .
we consider both phase - coherent and decoherent transport . in the latter case
, the effect of decoherence is included using b@xmath0ttiker probes @xcite as implemented by the damoto - pastawski model @xcite .
the b@xmath0ttiker probes are fictitious probes connected to each energy level of the bases , which describes the phase - breaking of electrons due to the interaction with the environment by extracting the electron from the system and then re - injecting it back . at each b@xmath0ttiker probe ,
the net current is zero . after including the b@xmath0ttiker probes ,
the retarded green s function for a dna strand bridging across the two metal contacts is , @xmath23 where @xmath24 is the energy , @xmath8 the hamiltonian of the strand as shown in eq .
( [ hamiltonian ] ) , @xmath25 and @xmath26 the self - energy matrices from the left ( right ) contact and b@xmath0ttiker probes .
we employ the wide - band limit ( wbl ) approximation to evaluate the self - energies , which ignores the real part of the matrices @xcite . within this approximation ,
the self - energy due to the left ( right ) contact is controlled by the broadening matrix @xmath27 , and @xmath28 .
the introduction of @xmath27 is due to the coupling between the dna and the left ( right ) contact .
the self - energy due to each fictitious b@xmath0ttiker probe is treated in a similar manner , @xmath29 , where @xmath30 is the broadening matrix arising from the coupling between the dna and each b@xmath0ttiker probe .
the self - energy matrix @xmath31 can be obtained by summing up the self - energies from all the b@xmath0ttiker probes , @xmath32 . here
@xmath33 is the number of b@xmath0ttiker probes . in the following calculations , we treat the coupling @xmath27 and @xmath30 to be energy - independent diagonal matrices . with the inclusion of b@xmath0ttiker probes , the probability for an electron to transmit from the left ( l ) contact to the right contact ( r ) is , @xmath34 where @xmath35 is the transmission between the @xmath15th site and the @xmath36th site in phase - coherent case , which can be determined by @xmath37 , with @xmath38 being the advanced green s function and @xmath39
particularly , when @xmath40 and @xmath41 , @xmath35 is the phase - coherent component of the transmission between the left and right contacts .
@xmath42 is the inverse matrix of @xmath43 , the elements of which are @xmath44 , where @xmath45 is the reflection probability at probe @xmath15 and @xmath46 .
the small bias ( linear response ) conductance as a function of fermi energy is , @xmath47 where @xmath48 is the fermi function evaluated at 300 k.
in this part , we analyze the tight - binding parameters of the homo level for gc8 and gcm8 , including the on - site energy of each base and hopping integrals between two neighboring bases . on - site energies and hopping
integrals are two important factors in determining the charge transport properties of a dna strand .
these parameters can be extracted from the hamiltonian @xmath8 in eq .
( [ hamiltonian ] ) using the tight - binding representation in eq .
( [ h_tb ] ) . the result with b3lyp
is shown in fig .
[ b3lyp_tb ] , where the numbers shown in black are the on - site energies , the arrows and numbers shown by other different colors correspond to hopping integrals between two neighboring bases both within and between the complementary strands .
our values of the tight - binding parameters are comparable to those in the previous work @xcite .
the most significant differences for the parameters of gc8 and gcm8 at homo level lie in the on - site energies of cs and cms , and the hopping integrals between two nearest neighboring g and c ( or cm ) .
for example , in fig .
[ b3lyp_tb ] we can see that the values of the on - site energies for cs in gc8 are around -5800 mev , while the values for cms in gcm8 are around -5600 mev , which are approximately 200 mev higher than the native cytosine bases .
the absolute values of the intra - strand hopping integrals between the two nearest neighboring g and cm from the 5 to 3 ends and from 3 to 5 ends increase by 30 mev and 15 mev respectively , compared to gc8 , while the absolute values of the inter - strand hopping integrals between the two nearest neighboring g and cm increase by about 8 mev .
it is also interesting to see that the absolute values of the intra - strand hopping between the two nearest neighboring g and c ( or cm ) in the 5 to 3 and 3 to 5 directions are different by almost 9 ( 5 ) times for gc8 ( gcm8 ) , which reveals the orientation - dependent nature of electron hopping in the double helical structure .
the above discussions show that the small modification on cytosine bases can cause a change in tight - binding parameters at the homo level of a dna strand .
to understand the role of methylation in transport properties , we calculate the transmission and conductance of gc8 and gcm8 in both phase - coherent and decoherent cases , using the full hamiltonian from the dft calculation and focus on the energy region which is in the vicinity of the homo level .
we set the coupling strengths to be 100 mev for @xmath27 and 5 mev for @xmath30 .
we first consider the phase - coherent transport in gc8 and gcm8 . the transmission and conductance of the two strands in the homo vicinity using b3lyp
are shown in fig .
[ b3lyp_tran_cond_coherent ] and its inset .
we find that the transmission of gcm8 is very close to that of gc8 nearby the homo level , while away from the homo ( when e @xmath49 -4 ev ) , the transmission of gcm8 is larger than gc8 by about 5 times in the bandgap .
as one moves into the homo band , the transmission of both strands are oscillatory , and the transmission values for gcm8 can be larger than gc8 in some energy windows and smaller in other windows .
the ratio of the transmission between gcm8 and gc8 in the energy window from -5.5 ev to -4 ev ranges from @xmath50 to @xmath51 times . in this work
we only provide a discussion based on the picture for conduction that is used in semiconductors . to proceed , we also show the linear response conductance as a function of fermi energy in the inset of fig .
[ b3lyp_tran_cond_coherent ] , with a shift in the energy axis by @xmath52 for gcm8 .
we note that the conductance of gcm8 is larger than that of gc8 by about 2 times for fermi energy locations nearby the homo level and 5 times away from the homo in the bandgap .
it is also interesting to note that the conductance plots in the insets of both fig .
[ b3lyp_tran_cond_coherent ] points to the fact that if the fermi energy is shifted by as small as 50 mev ( say due to environmental fluctuations , location of contact atoms and so on ) , the conductance can change by as much as 5 - 10 times .
it is of relevance to note that in experiments the conductance fluctuates between various measurements , though the mean conductance for gc8 and gcm8 are well defined .
apart from changes in the coupling between the contact and dna , fig .
[ b3lyp_tran_cond_coherent ] also shows that variations in the location of the fermi level by as small as 50 mev may be responsible for the fluctuations in conductance between different measurements . a calculation that is able to account for the environmental variations which shift the fermi energy while beyond our expertise will be useful and we expect it to yield further insight .
a few groups @xcite have suggested that decoherence due to electron s interaction with the surrounding environment plays a large part in determining dna conductance .
we now study the influence of decoherence on charge transport through gc8 and gcm8 via the b@xmath0ttiker probes approach to see if our conclusions in the above section change .
we first set the decoherence rate to be 5 mev , a typical value in carbon nanotubes at room temperature @xcite for both strands .
the transmission and conductance for gc8 and gcm8 calculated with b3lyp after including decoherence with a rate of 5 mev is shown in blue solid and red dashed curves in fig .
[ b3lyp_tran_cond_deocherent ] .
we find that for both gc8 and gcm8 , after including decoherence , the peaks of the transmission become wider and the transmission values at the peaks are smaller compared to that in phase - coherent transport , due to the broadening of the localized energy levels , while away from the peaks the transmission values increase .
we also note that decoherence can decrease the difference between gc8 and gcm8 for both transmission and conductance .
this effect is more obvious in the homo - lumo ( here , lumo refers to the lowest unoccupied molecular orbital ) gap .
for example , the transmission of gcm8 with 5 mev decoherence is only 1.2 times larger in the homo - lumo gap , compared to 5 times in the coherent transport .
similarly , the conductance of gcm8 with the inclusion of 5 mev decoherence is 1.5 times larger than gc8 throughout the interested energy region .
as has been pointed out by several groups , the addition of the methyl groups in the cytosine bases can increase the molecular polarizability and help stabilize the dna structure @xcite .
using dft calculation with b3lyp/6 - 31 g , we find that the hydrogen bonding for a single methylated g : c base pair is 23 mev stronger than the unmethylated one .
this indicates that the methylated dna has a more rigid structure and undergoes less fluctuations in the presence of the water molecules , counter ions and other impurities , which may result in a smaller decoherence rate than the native dna strand . to investigate the effect of the change in molecular stability
, we study the transport in terms of the decoherence rate for gcm8 . since the precise value of the decohrence rate due to the structural alteration is hard to determine , we treat it as a variable . the transmission and conductance for gcm8 with decoherence rate ranging from 1 mev to 4 mev
are shown in a series of dashed curves in fig .
[ b3lyp_tran_cond_deocherent ] .
interestingly , as the decoherence rate of gcm8 drops below 5 mev , the transmission of gc8 is larger than gcm8 by a few times , which is in qualitative agreement with the experimental observation that the conductance of gc8 is 1.2 times larger than gcm8 on average .
take the energy point at -4 ev for example . as the decoherence rate of gcm8 decreases from 5 mev to 4 mev , the transmission of gc8 is now 1.2 times larger than gcm8 . as the rate further decreases to 3 mev ,
2 mev and 1 mev , the transmission is 2.1 , 4.4 , 15.7 times larger , respectively . for the conductance plots in the inset of fig .
[ b3lyp_tran_cond_deocherent ] , we find the similar trend . with these analysis
, we suggest that the lower conductance for gcm8 in the experimental measurement is due to the enhanced stability of the molecular structure induced by the methyl groups .
note that the method adopted in this paper to include decoherence can qualitatively handle aspects of energy level broadening introduced by electron - vibrational coupling @xcite but can not model energy gain / loss or polaron formation @xcite . on the other hand , @xmath1 calculations of vibrational energies and the strength of their coupling to electrons is a computationally very challenging problem for large aperiodic systems such as dna .
this field has seen some welcome progress using dft inspired tight - binding approaches @xcite , which may eventually lead to a more fundamental treatment of transport by accounting for self - consistency , dna - contact coupling , electron - vibrational coupling and electron - environment coupling . as dft calculations are sensitive to the choice of the xc functionals , it is crucial to investigate this effect on transport .
in addition to b3lyp , we consider cam - b3lyp @xcite based range - separated functional .
the choice of the latter is motivated by recent advances in range - separated dft @xcite , where the range - separation parameters are tuned to obey the koopmans theorem , i.e. making the homo as close as possible to the ionization potential ( ip ) of the neutral molecule . rather than using the default parameters for cam - b3lyp in gaussian 09
, we tune the relevant parameters @xcite to satisfy koopman s theorem , that is , to make the homo and ip of a neutral molecule to satisfy @xmath53homo + ip = 0@xmath53 . in a range - separated functional ,
the exchange term is described with two parts @xcite , @xmath54}{r_{12}}\ ] ] the first term is the short - range interaction , treated with dft exchange and the second term is the long - range interaction , treated with hartree - fock ( hf ) exchange .
@xmath55 and @xmath56 are parameters which control the component of dft and hf exchange , with @xmath57 , @xmath58 , @xmath59 @xcite
. we can find the optimal values of @xmath55 , @xmath56 and @xmath60 by minimizing the following function for @xmath53homo + ip@xmath53 , @xmath61 by varying the values of @xmath55 , @xmath56 and @xmath60 , we find that when @xmath62 , @xmath63 , @xmath64 , the @xmath53homo + ip@xmath53 function in eq .
( [ optimal_ip ] ) for gc8 can be minimized .
this set of parameters is consistent with the choice of pbe - based range - separated hybrid density functional for 3,4,9,10-perylene - tetracarboxylic - dianydride ( ptcda ) and 1,4,5,8-naphthalene - tetracarboxylic - dianhydride ( ntcda ) molecules @xcite . since the structure of gcm8 differ from gc8 only by the methyl groups , we assume this set of parameters also works for gcm8 .
a comparison of the physical quantities , homo , ip and @xmath53ip + homo@xmath53 for gc8 and gcm8 calculated with functionals b3lyp and the tuned cam - b3lyp are shown in table [ b3lyp ] and table [ cam_b3lyp ] , respectively . [ cols="^,^,^,^,^,^,^",options="header " , ]
we can see that with both functionals , the homos of gc8 and gcm8 are very close - only differ by 7 mev , due to the minor structural difference of the two strands .
compared to b3lyp , the effect of the tuned cam - b3lyp is to shift the homo for both strands by around 1.5 ev .
ip of gc8 is also similar to that of gcm8 with the two different functionals .
as expected , @xmath53ip + homo@xmath53 for gc8 improve significantly with the tuned cam - b3lyp functional , which indicates the efficacy of the parameters , @xmath55 , @xmath56 and @xmath60 .
it is also worth noting that the @xmath53ip + homo@xmath53 for gcm8 with the tuned cam - b3lyp is only 15 mev , which is slightly smaller than the one calculated with b3lyp , suggesting that the range - separation parameters determined for gc8 also hold for a different strand with a similar structure .
this establishes the transferability of our work . as a comparison with the results computed with b3lyp
, we also show the tight - binding parameters with tuned cam - b3lyp in fig .
[ cam_tb ] .
similar to the findings with b3lyp , one can see that with the tuned cam - b3lyp , the main influence of cytosine methylation is to change the on - site energies of cytosine bases by about 200 mev and to modify the hopping integrals between the two nearest neighboring g and c ( cm ) with approximately the same amount as that in b3lyp . comparing the quantities computed by b3lyp with the corresponding ones computed by the tuned cam - b3lyp , we find that the main difference between the two sets of the parameters lie in the on - site energies of the g and c ( cm ) bases , shifting by around 1500 mev , while the hopping integrals are of the same order of magnitude , only deviating by a few milli - electron volts .
we also show the transmission and conductance for gc8 and gcm8 in coherent transport computed with the tuned cam - b3lyp in fig .
[ cam_tran_cond_coherent ] .
the main observation in results with b3lyp also holds here with cam - b3lyp - the transmission of gcm8 is close to gc8 nearby the homo level and larger than gc8 away from the homo .
different from b3lyp , the conductance values of gcm8 is now 8 times larger than gc8 nearby the homo , compared to 2 times in the above discussion for b3lyp .
we attribute this finding to the larger difference in ip between gc8 and gcm8 with the tuned cam - b3lyp , as shown in table [ cam_b3lyp ] . comparing the transmission and conductance in fig .
[ b3lyp_tran_cond_coherent ] and fig .
[ cam_tran_cond_coherent ] , we conclude that the transport properties in coherent case calculated by b3lyp and the tuned cam - b3lyp mainly differ in the energy scale .
this is in agreement with the observations in the tight - binding parameters at the homo level shown in fig .
[ b3lyp_tb ] and fig .
[ cam_tb ] that the on - site energies of dna bases with the two different functionals differ by as much as 1500 mev , while the intra- and inter - strands hopping integrals are quite similar . in fig .
[ cam_tran_cond_deocherent ] , we show the transmission and conductance for gc8 and gcm8 calculated by the tuned cam - b3lyp with the inclusion of 5 mev decoherence for gc8 and 1 - 5 mev for gcm8 . we find that the transmission and conductance through gc8 and gcm8 in the decoherent transport calculated with the tuned cam - b3lyp is qualitatively similar to that with b3lyp , except the shift in the energy scale .
the observation that the lower conductance for gcm8 in the experiments may be due to the more stable molecular structure of gcm8 induced by the methyl groups also hold here .
it is worth pointing out that in this case the conductance of gcm8 with the decoherence rate of 5 mev is about 4 times larger than gc8 at the same energy point nearby the homo level , due to the fact that the ip of gcm8 is 24 mev smaller than gc8 with the tuned cam - b3lyp .
based on the observation in sec .
[ transport ] and [ result_cam_b3lyp ] , we find similar transport results with the b3lyp and the tuned cam - b3lyp xc functionals in both phase - coherent and decoherent transports . in this section , we further study the effect of functionals on dna charge transport with various coupling strengths to the contacts .
the broadening of the molecular levels due to the coupling to the contacts plays an important role in determining the electron flow through a molecule @xcite . according to the tight - binding parameters in sec .
[ tight_binding ] and [ result_cam_b3lyp ] , the hopping integral value in a dna molecule ranges from a few milli - electron volts to hundreds of milli - electron volts
. a normal coupling strength due to the electrical contacts should be comparable to the hopping integral .
thus , we choose three values of coupling strengths , @xmath65 1 mev ( weak coupling limit ) , 100 mev ( normal coupling ) and 10 ev ( strong coupling limit ) and assume @xmath66 to study the transmission of gc8 and gcm8 in decoherent transport with a decoherence rate of 5 mev , as shown in fig . [ transport_coupling ] . ,
@xmath67 and @xmath68 .
the energy scale of transmission computed with cam - b3lyp has been shifted by @xmath69.,width=316 ] to quantify our understanding , the energy scale for transmission computed with the tuned cam - b3lyp has been shifted by @xmath69 in fig . [ transport_coupling ] . for gc8 , as the coupling strength increases ,
the transmission increases drastically with both functionals .
this effect is more evident when the coupling strengths are in weak regime .
for example , as the coupling strengths increases from 1 mev to 100 mev , the transmission of gc8 is enhanced by as much as 25 times .
it is reasonable to assume that , in the weak molecule - lead coupling limit , the misaligned molecular levels ( for example , from a b3lyp calculation ) with respect to the leads would result in an incorrect molecular response as charge is transferred from the leads into the molecule which should be observable .
however , our transmission results with both b3lyp and the tuned cam - b3lyp yield almost identical results in the vicinity of the homo irrespective of the strength of the contact coupling even though the delocalization and self - interaction errors are minimized in tuned range - separated xc forms @xcite .
we find similar results for gcm8 .
the natural follow - on question is how large is this effect ? , and if it is observable in realistic systems .
the other factor could be the inherent approximations within the green s function - dft approach itself . in order to quantify these effects , a systematic study of charge transport through well - characterized molecular junctions with different xc potentials and range - separation in the weak and strong coupling limits
is needed .
in summary , we have studied the possibility of detecting a methylated dna strand using its electronic properties from a theoretical perspective , based on dft calculations and the landauer - b@xmath0ttiker approach .
we explore the electronic structures and charge transport properties through an eight base - pair methylated dna strand and its native counterpart used in the experiment of ref . .
we first look at the tight - binding parameters of the homo level for the two strands and find that the methylation has a large influence on the on - site energies of cytosine bases and hopping integrals between two nearest neighboring guanine and cytosine bases .
the on - site energies of the methylated cytosine bases are approximately 200 mev higher than that of native cytosine bases .
the intra - strand hopping integrals between two nearest neighboring guanine and cytosine bases from 5 to 3 ends for a methylated strand are about 30 mev stronger than that of native strand , and 15 mev stronger from 3 to 5 ends .
we then carry out transport calculations in both phase - coherent and decoherent cases .
our calculations show that in phase - coherent transport , the transmission of electrons close to the homo band of the two strands are very similar . at energies away from homo , the transmission of gcm8 is larger than gc8 .
the linear response conductances of gc8 and gcm8 are also close to each other , with the conductance of gcm8 being slightly higher than gc8 .
the calculations also show that the conductance of these molecules can change by 5 - 10 times if there is a small shift ( 50 mev ) in the fermi energy due to environmental variations from experiment to experiment . in the phase - decoherent case , we find that the transmission and conductance trends occuring in the phase - coherent limit also hold in the homo vicinity if the decoherence rates for the two strands are the same . however , if the decoherence rate of the methylated strand is lower , we observe that the transmission and conductance of the methylated strand is also lower than the native strand , which is consistent with the experimental findings .
our results show that the effect of the two different functionals is to alter the on - site energies of the dna bases at the homo level , while the transport properties do not depend much on the two functionals .
the authors thank professor joshua hihath at university of california , davis for extensive discussions on experimental conductance measurements of dna strands .
the authors also thank benjamin e. van kuiken and professor xiaosong li at university of washington , dr .
fernando r. clemente at gaussian inc . for their help with cam - b3lyp .
niranjan govind and m. p. anantram would like to thank the northwest institute for advanced computing ( niac ) workshop at pnnl ( niac day@pnnl in march 2014 ) for stimulating discussions that lead to this collaboration .
m. p. anantram and jianqing qi acknowledge the support from the national science foundation under grant no .
102781 .
r. g. endres , d. l. cox and r. r. p. singh , rev . mod .
* 76 * , 195 ( 2004 ) . c. adessi , s. walch and m. p. anantram , phys .
b * 67 * , 081405(r ) ( 2003 ) . b. g@xmath5hler , v. hamelbeck , t. z. markus , m. kettner , g. f. hanne , z. vager , r. naaman and h. zacharias , science * 331 * , 894 ( 2011 )
. zouti xie , tal z. markus , sidney r. cohen , zeev vager , rafael gutierrez and ron naaman , nano lett .
* 11 * , 4652 ( 2011 ) .
naoto matsuo , shyogo takagi , kazushige yamana , akira heya , tadao takada and shin yokoyama , jpn . j. of appl . phys .
* 51 * , 04dd13 ( 2012 ) .
ning kang , artur erbe and elke scheer , appl .
lett . * 96 * , 023701 ( 2010 ) .
takahito ohshiro , kazuki matsubara , makusu tsutsui , masayuki furuhashi , masateru taniguchi and tomoji kawai , sci
* 2 * , 501 ( 2012 ) .
henk w. ch .
postma , nano lett .
* 10 * , 420 ( 2010 ) .
elizabeth a manrao , ian m derrington , andrew h laszlo , kyle w langford , matthew k hopper , nathaniel gillgren , mikhail pavlenok , michael niederweis and jens h gundlach , nat .
biotechnol . * 30 * , 349 ( 2012 ) .
makusu tsutsui , kazuki matsubara , takahito ohshiro , masayuki furuhashi , masateru taniguchi and tomoji kawai , j. am .
* 133 * , 9124 ( 2011 ) .
joshua hihath , shaoyin guo , peiming zhang and nongjian tao , j. phys . : condes .
matter * 24 * , 164204 ( 2012 ) .
keith d robertson , oncogene * 20 * , 3139 ( 2001 ) .
james e. gill , joseph a. mazrimas and clyde c. bishop , biochim .
acta * 335 * , 330 ( 1974 ) .
lawrence c. sowers , barbara ramsay shaw and w. david sedwick , biochem .
* 148 * , 790 ( 1987 ) .
philip m. d. severin , xueqing zou , hermann e. gaub and klaus schulten , nucl .
acids res . *
39 * , 8740 ( 2011 ) .
walter doerfler , ann . rev
. biochem . * 52 * , 93 ( 1983 ) .
kelly j vining , kyle r pomraning , larry j wilhelm , henry d priest , matteo pellegrini , todd c mockler , michael freitag and steven h strauss , bmc genomics * 13 * , 27 ( 2012 ) .
david a. khavari , george l. sen and john l. rinn , cell cycle * 9 * , 3880 ( 2010 ) .
tetsuya goto and marilyn monk , microbiol .
biol . rev . *
62 * , 362 ( 1998 ) . en li , timothy h. bestor and rudolf jaenisch , cell * 69 * , 915 ( 1992 ) .
hong lei , suk p.
oh , masaki okano , ruth j@xmath0ttermann , kendrick a. goss , rudolf jaenisch and en li , development * 122 * , 1996 ( 3195 ) .
hongshan guo , ping zhu , liying yan , rong li , boqiang hu , ying lian , jie yan , xiulian ren , shengli lin , junsheng li , xiaohu jin , xiaodan shi , ping liu , xiaoye wang , wei wang , yuan wei , xianlong li , fan guo , xinglong wu , xiaoying fan , jun yong , lu wen , sunney x. xie , fuchou tang and jie qiao , nature * 511 * , 606 ( 2014 ) .
zachary d. smith , michelle m. chan , kathryn c. humm , rahul karnik , shila mekhoubad , aviv regev , kevin eggan and alexander meissner , nature * 511 * , 611 ( 2014 ) .
timothy a. hore , robert w. rapkins and jennifer a. marshall graves , trends genet . *
23 * , 400 ( 2007 ) .
keith d. robertson , nature * 6 * , 597 ( 2005 ) .
partha m. das and rakesh singal , j. clin .
* 22 * , 4632 ( 2004 ) .
stephen b baylin , nat .
* 2 * , s4-s11 ( 2005 ) .
lanlan shen , robert a. waterland , curr .
care * 10 * , 576 ( 2007 ) . susan j. clark , janet harrison , cheryl l.paul and marianne frommer , nucl .
acids res . *
22 * , 2990 ( 1994 ) .
christoph bock , eleni m tomazou , arie b brinkman , fabian m@xmath0ller , femke simmer , hongcang gu , natalie j@xmath70ger , andreas gnirke , hendrik g stunnenberg and alexander meissner , nat .
* 28 * , 1106 ( 2010 ) .
bingqian xu , peiming zhang , xiulan li and nongjian tao , nano lett . * 4 * , 1105 ( 2004 ) .
fang chen , joshua hihath , zhifeng huang , xiulan li and n. j. tao , annu .
chem * 58 * , 535 ( 2007 ) .
a. k. mahapatro , k. j. jeong , g. u. lee and d. b. janes , nat . nanotechnol . * 18 * , 195202 ( 2007 ) .
makusu tsutsui , masateru taniguchi , kazumichi yokota and tomoji kawai , nat .
* 5 * , 286 ( 2010 ) . james clarke , hai - chen wu , lakmal jayasinghe , alpeshpatel , stuart reid and hagan bayley , nat . nanotechnol . * 4 * , 265 ( 2009 ) .
emma v. b. wallace , david stoddart , andrew j. heron , ellina mikhailova , giovanni maglia , timothy j. donohoe and hagan bayley , chem . commun . * 46 * , 8195 ( 2010 ) .
elizabeth a. manrao , ian m. derrington , mikhail pavlenok , michael niederweis and jens h. gundlach , plos one * 6 * , e25723 ( 2011 ) . andrew h. laszlo , ian m. derrington , henry brinkerhoff , kyle w. langford , ian c. nova , jenny mae samson , joshua j. bartlett , mikhail pavlenok and jens h. gundlach , pnas * 110 * , 18904 ( 2013 ) .
jiwook shim , gwendolyn i. humphreys , bala murali venkatesan , jan marie munz , xueqing zou , chaitanya sathe , klaus schulten , farhad kosari , ann m. nardulli , george vasmatzis and rashid bashir , scientific reports * 3 * , 1389 ( 2013 ) .
phuc tan bui , tomoaki nishino , hiroshi shiigi and tsutomu nagaoka , chem . commun . * 51 * , 1666 ( 2015 ) .
c. toher , a. filippetti , s. sanvito and kieron burke , phys .
* 95 * , 146402 ( 2005 ) .
aron j. cohen , paula mori - snchez and weitao yang , science * 321 * , 792 ( 2005 ) .
roi baer , ester livshits and ulrike salzner , annu .
chem . * 61 * , 85 ( 2010 ) .
leeor kronik , tamar stein , sivan refaely - abramson and roi baer , j. chem .
theory comput .
* 8 * , 1515 ( 2012 ) .
monika srebro and jochen autschbach , j. chem .
theory comput .
* 8 * , 245 ( 2012 ) .
sivan refaely - abramson , sahar sharifzadeh , niranjan govind , jochen autschbach , jeffrey b. neaton , roi baer and leeor kronik , phys .
rev . lett . * 109 * , 226405 ( 2012 ) .
tamar stein , leeor kronik and roi baer , j. am . chem
* 131 * , 2818 ( 2009 ) .
kenneth lopata and niranjan govind , j. chem .
theory comput .
* 9 * , 4939 ( 2013 ) .
m. j. frisch , g. w. trucks , h. b. schlegel , g. e. scuseria , m. a. robb , j. r. cheeseman , g. scalmani , v. barone , b. mennucci , g. a. petersson , et al .
, gaussian 09 , revision a.1 . ,
gaussian , inc .
, wallingford ct ( 2009 ) .
t. macke , d.a .
case , in : n.b .
leontes , j. santalucia jr .
( eds . ) , molecular modeling of nucleic acids , american chemical society , washington , dc , 1998 , p. 379 .
d. d. eley and d. i. spive , trans .
faraday soc .
* 58 * , 411 ( 1962 ) .
jianqing qi , neranjan edirisinghe , m. golam rabbani and m. p. anantram , phys . rev .
b * 87 * , 085404 ( 2013 ) .
nuri a. temiz , duncan e. donohue , albino bacolla , brian t. luke and jack r. collins , plos one * 7 * , e35558 ( 2012 ) .
axel d. becke , j. chem . phys .
* 98 * , 5648 ( 1993 ) . c. lee , w. yang and r. g. parr , phys .
b * 37 * , 785 ( 1988 ) .
p. j. stephens , f. j. devlin , c. f. chabalowski and m. j. frisch , j. phys .
* 98 * , 11623 ( 1994 ) .
w. j. hehre , r. ditchfeld and j. a. pople , j. chem . phys . * 56 * , 2257 ( 1972 ) .
takeshiz yanai , david p. tew and nicholas c. handy , chemical physics letters * 393 * , 51 ( 2004 )
. m. a. rohrdanz and j. m. herbert , j. chem .
phys . * 129 * , 034107 ( 2008 ) .
perolov l@xmath5wdin , j. chem . phys .
* 18 * , 365 ( 1950 ) .
h. mehrez and m. p. anantram , phys .
b * 71 * , 115405 ( 2005 ) .
m. b@xmath0ttiker , phys .
lett . * 57 * , 1761 ( 1986 ) .
m. b@xmath0ttiker , ibm j. res
. dev . * 32 * , 317 ( 1988 ) .
jorge l. damato and horacio m. pastawski , phys . rev .
b * 41 * , 7411 ( 1990 ) .
carmen herrmann , gemma c. solomon , joseph e. subotnik , vladimiro mujica and mark a. ratner , j. chem .
phys . * 132 * , 024103 ( 2010 ) .
alexander a. voityuk , notker ro1sch , m. bixon and joshua jortner , j. phys .
b * 104 * , 9740 ( 2000 ) .
johan olofsson and sven larsson , j. phys .
b * 105 * , 10398 ( 2001 ) .
alexander a. voityuk , joshua jortner , m. bixon and notker r@xmath5sch , j. chem . phys . * 114 * , 5614 ( 2001 ) .
alexander a. voityuk and notker r@xmath5sch , j. chem . phys . *
117 * , 5607 ( 2000 ) .
kittusamy senthilkumar , ferdinand c. grozema , c@xmath71lia fonseca guerra , f. matthias bickelhaupt , frederick d. lewis , yuri a. berlin , mark a. ratner and laurens d. a. siebbeles , j. am .
soc . * 127 * , 14894 ( 2005 ) .
r. guti@xmath72rrez , s. mandal and g. cuniberti , nano lett . * 5 * , 1093 - 1097 ( 2005 ) .
r. guti@xmath72rrez , s. mandal , g. cuniberti , phys .
b * 71 * , 235116 ( 2005 ) .
r. gutii@xmath72rre , s. mohapatra , h. cohen , d. porath and g. cuniberti , phys .
b * 74 * , 2006 ( 235105 ) . tom@xmath73@xmath74 .
kubar and marcus elstner , j. phys .
b * 112 * , 8788 ( 2008 ) .
mat@xmath75as zilly , orsolya ujs@xmath73ghy and dietrich e. wolf , phys .
b * 82 * , ( 125125 ) .
hiroyuki ishii , nobuhiko kobayashi and kenji hirose , j. vac .
b * 27 * , 882 ( 2009 ) .
christopher k@xmath5hler , tobias watermann and ermin malic , phys .
status solidi b * 249 * , 2483 ( 2012 ) .
christopher k@xmath5hler , tobias watermann and ermin malic , j. phys . : condens .
matter * 25 * , 105301 ( 2013 ) .
shaohui wang and eric t. kool , biochemistry * 34 * , 4125 ( 1995 ) .
j. @xmath76poner , h. a. gabb , j. leszczynski and hobza , biophysical journal * 73 * , 76 ( 1997 ) .
naoki sugimoto , mariko nakano and shu - ichi nakano , biochemistry * 39 * , 11270 ( 2000 ) .
sylvie derreumaux , mounia chaoui , georges tevanian and serge fermandjian , nucleic acids research * 29 * , 2314 ( 2001 ) .
jan norberg and mauno vihinen , journal of molecular structure * 546 * , 51 ( 2001 ) .
lesley r. rutledge , holly f. durst and stacey d. wetmore , phys .
* 10 * , 2801 ( 2008 ) .
luk@xmath73@xmath77fojt , vladim@xmath78r vetterl and th .
doneux , bioelectrochemistry * 75 * , 89 ( 2009 ) .
adam moser , rebecca guza , natalia tretyakova and darrin m. york , theor .
account * 122 * , 179 ( 2009 ) .
carles acosta - silva , vicenc branchadell , joan bertran and antoni oliva , j. phys .
b * 114 * , 10217 ( 2010 ) .
h. morkoc ( eds . ) , electrical conduction through molecules , academic press , 2003 .
danith ly , laurie sanii and gary b schuster , j. am .
soc . , * 121 * , 9400 ( 1999 ) r. guti@xmath72rrez , r caetano , p. b. woiczikowski , t. kuba and g. cuniberti , new j. phys . , * 12 * , 023022 ( 2010 ) | cytosine methylation has been found to play a crucial role in various biological processes , including a number of human diseases .
the detection of this small modification remains challenging . in this work ,
we computationally explore the possibility of detecting methylated dna strands through direct electrical conductance measurements . using density functional theory and the landauer - b@xmath0ttiker method , we study the electronic properties and charge transport through an eight base - pair methylated dna strand and its native counterpart .
we first analyze the effect of cytosine methylation on the tight - binding parameters of two dna strands and then model the transmission of the electrons and conductance through the strands both with and without decoherence .
we find that the main difference of the tight - binding parameters between the native dna and the methylated dna lies in the on - site energies of ( methylated ) cytosine bases .
the intra- and inter- strand hopping integrals between two nearest neighboring guanine base and ( methylated ) cytosine base also change with the addition of the methyl groups .
our calculations show that in the phase - coherent limit , the transmission of the methylated strand is close to the native strand when the energy is nearby the highest occupied molecular orbital level and larger than the native strand by 5 times in the bandgap .
the trend in transmission also holds in the presence of the decoherence with the same rate .
the lower conductance for the methylated strand in the experiment is suggested to be caused by the more stable structure due to the introduction of the methyl groups .
we also study the role of the exchange - correlation functional and the effect of contact coupling by choosing coupling strengths ranging from weak to strong coupling limit . |
the leptonic decay of a charged lepton @xmath0 is described by its michel parameters [ 1 - 4 ] complementary parameters must be introduced @xcite to describe the secondary lepton polarization . the measurements of all the parameters and the cross - section for the @xmath1 reaction allow to put upper limits on all the non - v - a amplitudes @xcite and `` prove experimentally '' the v - a structure of the weak interaction .
in the case of the @xmath2 lepton , this program was followed and successfully completed fifteen years ago @xcite . for the @xmath3 lepton , it is customary [ 7 - 9 ] to present the experimental results , in a similar way , as upper values of the possible coupling constants .
however , the validity of the standard model is today so well established , that the interest of @xmath3 michel parameter measurements is no more to prove the v - a structure but to look for small deviations from its predictions that would be evidence of new physics .
moreover , the additional measurements , which are necessary to complete the experimental proof of the v - a structure , will not be performed in a foreseeable future , while large statistics of @xmath3 pairs will soon be available at b - factories , allowing more precise measurements of the michel parameters .
therefore it seems worthwhile to look for the combinations of the measured parameters which are the most sensitive to non - standard effects .
this is the aim of the present paper .
if only the momentum of the final state charged lepton is measured , the decay of a polarized @xmath3 is entirely described , in its centre - of - mass , by the distribution @xmath4 where @xmath5 is the @xmath3 polarization , @xmath6 the angle between the polarization and the charged lepton momentum and @xmath7 the normalized lepton energy . + taking advantage of the weak interaction short range
, the decay can be represented by the most general four fermion contact interaction [ 1 - 4 ] , written below in the helicity projection formalism [ 6,10 - 12 ] : @xmath8\nonumber & & l = e,\mu~.\end{aligned}\ ] ] @xmath9 is the absolute coupling strength ; the @xmath10 are ten complex coupling constants describing the relative contribution of scalar ( @xmath11 ) , vector ( @xmath12 ) , and tensor ( @xmath13 ) interactions , respectively , for given chiralities @xmath14 of the @xmath3 and the charged decay lepton .
the neutrino chiralities @xmath15 and @xmath16 are uniquely defined for a given set \{@xmath17 } by the chirality selections rules : conservation for vector coupling , reversal for scalar and tensor .
cl@xmath18 & + v & ' '' '' @xmath19 ' '' '' + s & @xmath20 g^s_{lr}=\hspace{1em}2g^t_{lr}\end{array } \hspace{1em } \begin{array}{r } g^s_{rl}\hspace{3em}g^t_{rl}\\[1ex ] g^s_{rl}=\hspace{1em}2g^t_{rl } \end{array } $ ] ' '' '' + ' '' '' t&@xmath21 g^s_{lr}=-6g^t_{lr } \end{array } \hspace{1.3em } \begin{array}{r } g^s_{rl}\hspace{2.5em}g^t_{rl}\\[1ex ] g^s_{rl}=-6g^t_{rl } \end{array}$ ] ' '' '' + + the same matrix element describes also the contributions of possible lepton - number - violating neutral currents @xcite .
the relationship between the neutral currents and the @xmath10 constants , as given by the fierz transformation , is summarized in table [ tab : neutral ] .
+ the matrix element ( [ e : matrix ] ) can be used to compute the functions @xmath22 and @xmath23 of eq .
[ eq : decgencm ] .
they are described by the four real michel parameters : @xmath24 and @xmath25 for @xmath22 , @xmath26 and @xmath27 for @xmath23 . a convenient way to express
the relation between the michel parameters and the @xmath10 coupling constants is to introduce @xcite the six positive parameters @xmath28 they satisfy the relation @xmath29 which means that the normalization is absorbed in the definition of @xmath9 .
the @xmath24 , @xmath26 , and @xmath27 parameters are given by @xmath30 in geometrical terms , the point of coordinates @xmath24 , @xmath26 , and @xmath27 , in the space of the parameters , is the barycentre of six points , @xmath31 , @xmath32 , and @xmath33 with the weights @xmath34 , @xmath35 , and @xmath36 respectively .
since the point @xmath37 lies on the @xmath38 segment , and @xmath39 on @xmath40 , the allowed domain is just the tetrahedron @xmath41 ( fig .
[ fig : posdom ] ) . + ccccc ' '' '' ' '' '' & @xmath42&@xmath43&@xmath44&@xmath45 + ' '' '' ' '' '' maximum&@xmath46&@xmath46&@xmath46&@xmath46 + ' '' '' ' '' '' sm & @xmath46&@xmath47&@xmath47&@xmath47 + ' '' '' ' '' '' @xmath48 & @xmath46&@xmath47&@xmath49&@xmath47 + ' '' '' ' '' '' & @xmath50&@xmath51&@xmath52&@xmath53 + ' '' '' ' '' '' maximum & @xmath54&@xmath54&@xmath54&@xmath54 + ' '' '' ' '' '' sm & @xmath47&@xmath47&@xmath47&@xmath47 + ' '' '' ' '' '' @xmath48 & @xmath54&@xmath47&@xmath54&@xmath47 + ' '' '' ' '' '' & -&@xmath55&@xmath56&- + ' '' '' ' '' '' maximum&- & @xmath57&@xmath57&- + ' '' '' ' '' '' sm & -&@xmath47&@xmath47&- + ' '' '' ' '' '' @xmath48&- & @xmath47&@xmath49&- + the standard model prediction , @xmath58 , implies @xmath59 and is represented geometrically by the point @xmath37 , but @xmath60 does not imply @xmath58 , and the location of @xmath37 on the @xmath38 segment introduces further ambiguities .
the value @xmath60 is also obtained if @xmath61 and all the other constants equal to zero , and there is another possibility to reproduce the parameters predicted by the standard model , namely @xmath62 , @xmath63 , and @xmath64 .
so , with @xmath65 , @xmath66 , @xmath44 can reach an upper value of 1/2 and it is also possible that all the constants vanish but @xmath67 and @xmath68 . in this last case
, we get @xmath69 where @xmath70 is the relative phase of the two amplitudes .
accordingly , the upper possible values of @xmath71 and @xmath72 are 2 and 1/2 respectively .
the upper values of the constants , compatible with the standard model prediction , are given in table [ table : gmax ] .
more stringent experimental limits @xcite are not the consequences of the data but of additional hypotheses or constraints .
as the physically interesting region is the neighbourhood of @xmath58 , no relevant bound on the non - standard @xmath3 left - handed couplings can be extracted from the measurement of @xmath24 , @xmath26 and @xmath27 without additional hypotheses . +
since the standard model prediction is represented by @xmath37 which is located on @xmath38 , a convenient set of parameters is given by two equations of the @xmath38 line and a third variable which specify the position on the segment .
using the equations of the faces @xmath73 and @xmath74 gives expressions which are positive for points inside the tetrahedron .
they are more usefully combined into @xmath75 = \beta^++\alpha^-+\gamma^-\\\nonumber & = & |g^v_{rr}|^2+|g^v_{lr}|^2+\frac{1}{4}\,|g^s_{rr}|^2+\frac{1}{4}\,|g^s_{lr}|^2 + 3|g^t_{lr}|^2~,\\ \label{eq : sdef } { \cal s } _
r^{\tau}&= & \frac{2}{3}\ , [ \rho-\xi\delta]=\beta^++\frac{4}{3}\,\gamma^-\\\nonumber & = & |g^v_{rr}|^2+\frac{1}{4}\,|g_{rr}^s|^2 + \frac{1}{4}\,|g^s_{lr}-2g^t_{lr}|^2~,\end{aligned}\ ] ] which , owing to their normalization , can be interpreted as fractional contributions of @xmath3 right - handed couplings to its leptonic partial width and used to bound the @xmath76 amplitudes . for the third parameter , @xmath26 or @xmath24 can be chosen .
the last michel parameter @xmath25 can be written @xmath77 + 2\,{\mathrm{re}}\,[\,g^{v*}_{lr}\,(g^s_{rl}+6g^t_{rl})/4\,]\\ & ~&+\,{\mathrm{re}}\,[\,g^{v*}_{ll}\,g^s_{rr}/2\ , ] + { \mathrm{re}}\,[\,g^{v*}_{rr}\,g^s_{ll}/2\,]~.\end{aligned}\ ] ] this expression implies the inequality @xmath78 where , @xmath79 , @xmath80 , and @xmath81 is the greatest value of @xmath82 , compatible with the value of @xmath24 .
[ fig : etapos ] shows that @xmath83 is reached for @xmath84 when @xmath85 and for @xmath86 for @xmath87 , leading to the rather weak bound : @xmath88
if only charged vector currents are present ( ) , we have @xmath89 and the allowed domain is the @xmath90 tetrahedron that we call the restricted domain .
it is interesting to remark that , for a very wide class of models and hypotheses , the allowed domain for the michel parameters is this restricted domain and @xmath91 are related to ( neutral ) tensor currents @xcite , and table [ tab : neutral ] . ] and that getting a point outside it requires the conspiracy of scalar and tensor couplings @xcite .
we give below a list of general hypotheses leading to the restricted domain .
we use for that the lorentz covariance of the charged and neutral currents and the properties of the fierz transformation only .
more specific constraints associated with the hypothesis of a single boson exchange can be found in @xcite and @xcite .
when the allowed domain of the parameters is the restricted one , the point @xmath37 , which is the standard model prediction , is one of its vertices , therefore there is a third positive quantity whose vanishing is associated with this point .
we use for it @xmath92 .
its precise physical meaning and the ambiguities in the interpretation of the measurements depend nevertheless on the hypothesis that lead to the restriction .
+ @xmath93 + if all the scalar coupling constants vanish , the condition @xmath94 implies @xmath95 , therefore the complete domain is not allowed . defining the new parametrization @xmath96 we get the relations @xmath97
\rho&=&\frac{3}{4}\,(\beta^++\beta^-)\nonumber + \frac{1}{4}\,(\gamma^+_{\mathrm{ns}}+\gamma^-_{\mathrm{ns}})~,\\ \xi&=&3\,(\alpha^-_{\mathrm{ns}}-\alpha^+_{\mathrm{ns } } ) + ( \beta^--\beta^+)-\frac{5}{3}\,(\gamma^+_{\mathrm{ns}}-\gamma^-_{\mathrm{ns}})~,\\\nonumber \xi\delta&=&\frac{3}{4}\,(\beta^--\beta^+ ) + \frac{1}{4}\,(\gamma^+_{\mathrm{ns}}-\gamma^-_{\mathrm{ns}})~.\end{aligned}\ ] ] the point of coordinates @xmath24 , @xmath26 , and @xmath27 is now the barycentre of @xmath31 , @xmath32 , and two new points @xmath98 whose coordinates are ( @xmath99 , @xmath100 , @xmath101 ) but , since @xmath102 is located on the @xmath103 segment and @xmath104 on @xmath105 ( fig .
[ fig : posdom ] ) , the allowed region is the restricted domain .
the third constraint can be written , in terms of the @xmath106 s , as @xmath107 measurements represented by @xmath37 imply a v - a structure .
the combination of charged , vector and tensor currents belongs to this class .
+ @xmath108 + if @xmath108 , we define @xmath109 and obtain @xmath110\nonumber \rho&=&\frac{3}{4}\,(\beta^++\beta^- + \gamma^+_{\mathrm{nt}}+\gamma^-_{\mathrm{nt}})~,\\ \xi&=&3\,(\alpha^-_{\mathrm{nt}}-\alpha^+_{\mathrm{nt } } ) + ( \beta^--\beta^+)+(\gamma^+_{\mathrm{nt}}-\gamma^-_{\mathrm{nt}})~,\\\nonumber \xi\delta&=&\frac{3}{4}\,(\beta^--\beta^++ \gamma^+_{\mathrm{nt}}-\gamma^-_{\mathrm{nt}})~.\end{aligned}\ ] ] here the points @xmath111 and @xmath112 coincide with the points @xmath37 and @xmath39 respectively .
the allowed domain is then the restricted one .
the third constraint is @xmath113 the vector couplings only are bounded by the measurement of @xmath24 , due to the ambiguity between @xmath32 and @xmath114 .
the combination of charged , vector and scalar currents and neutral , vector currents belongs to this class . + * @xmath115 , @xmath116 * + the restriction of the domain is evident from eq .
[ eq : gammadef ] .
the third constraint can be written @xmath117 measurements represented by @xmath37 imply a v - a structure .
the combination of charged , vector currents and neutral , vector and scalar currents belongs to this class . + * `` v - a plus something '' * + in an especially interesting family of models @xcite the decay is described by the addition of a single , non - standard contribution to the standard model amplitude .
their predictions are perfectly transparent in the present geometrical presentation .
they follow at once from the definitions of the parameters and the properties displayed in table [ tab : neutral ] . -if the non - standard contribution is a charged vector current , all the restricted domain is allowed .
the only new prediction is @xmath118 .
-the hypothesis of an additional charged scalar current belongs to the second of the above defined classes with the further conditions @xmath119 .
the allowed domain is then the @xmath120 segment ( @xmath121 ) , since the points @xmath122 and @xmath123 are identical .
-the contribution of a neutral vector is included in the same class with @xmath124 .
the allowed domain is again the @xmath120 segment .
-the hypothesis of an additional neutral scalar leads to @xmath89 ( third class ) and @xmath125 .
the allowed domain is the two - dimensional one , spanned by the points @xmath126 , @xmath127 , and @xmath37 .
the corresponding condition is @xmath128 and the bound on @xmath25 is stricter : @xmath129 .
in the standard approach described in section [ cha : intro ] , the indicators of new physics constructed with the michel parameters are , besides the @xmath25 parameter itself , the two positive quantities @xmath130 and @xmath131 , defined by eqs . [ eq : pdef ] and [ eq : sdef ] , which bound the coupling constants and can be interpreted as ( non - independent ) contributions of new physics to the @xmath3 decay . using the world - average values of the parameters @xcite , under the hypothesis of @xmath132-@xmath2 universality ,
yields @xmath133 if hypotheses are made that reduce the dimensionality of the domain , the determination of the parameters can be improved by a constrained fit .
for instance , if the domain is the @xmath120 segment , there is only one free parameter .
neglecting the correlations between the measurements and taking advantage of the near equality of the errors on @xmath26 and @xmath134 , this parameter is merely their average @xmath135 and the quantity @xmath130 reduces to @xmath136=\frac{1}{2}\,[\,1-\frac{\xi}{2}-\frac{2}{3}\xi\delta\,]~.\ ] ] it is noteworthy that the same strategy can be followed under the much weaker hypotheses that imply the restricted domain .
this is due to the fact that the measured value of @xmath24 forces the point which represents the measurements in the parameter space to lie on the edge of the domain .
quantitatively , the difference of @xmath26 and @xmath134 is bounded by the inequality @xmath137 from the values of the errors @xcite @xmath138 and @xmath139 , it is clear that there is no physically relevant information in @xmath140 and that even the @xmath3-@xmath141 universality prediction @xcite @xmath142 is better tested by the measurement of @xmath24 than by comparing @xmath26 and @xmath27 . assuming only the restricted domain , the indicators of new physics are @xmath143=0.002\pm 0.011~.\ ] ] they bring a clear improvement of the sensitivity with respect to @xmath144 and @xmath145 .
it must be noted that @xmath146 is not strictly positive but that excursions of @xmath135 beyond 1 are severely limited by the inequality @xmath147 since @xmath148 and @xmath149 .
a complete study of the constraints on the michel parameters and the ambiguities of their interpretation has been presented .
it has been shown that , for a very wide class of hypotheses and models , which cause the same restriction of the parameter domain , the best indicators of new physics are the combinations , @xmath92 and @xmath150 .
compared to the customary estimators , their sensitivities are roughly twice better .
the third available parameter , @xmath140 , is better determined by the geometry of the domain and the value of @xmath24 than by its measurement .
in the numerical exercise above , the error correlations were neglected and the @xmath132-@xmath2 universality was assumed .
we will discuss briefly this two approximations . at low energy , where the @xmath3 s are unpolarized ,
the @xmath24 and @xmath25 parameters are determined by the single - lepton laboratory energy distributions and the @xmath26 and @xmath27 parameters by the spin - correlated @xmath151 @xmath152 decay distributions .
estimates of the covariance matrices for measurements at 4 gev and 10 gev can be found in @xcite . at the z peak ,
the @xmath3 polarization makes ambiguous the interpretation of a single tau leptonic - decay distribution and , since the transverse spin correlations depend on the z couplings , only the helicity correlation of @xmath151 and @xmath152 , which is equal to -1 , is used .
if the decay distribution of a @xmath152 in the channel @xmath153 , is written , @xmath154 the correlated distribution for the @xmath153 and @xmath155 channels reads then @xmath156~,\ ] ] where @xmath5 is the @xmath152 polarization . for a hadronic decay , with the notations of @xcite ,
@xmath157 is the optimal variable @xmath158 and the decay distribution is @xmath159~,\ ] ] where the @xmath160 parameter is equal to 1 in the standard model and @xmath161 , the decay distribution is independent of the @xmath3 charge . in the case of @xmath162 , this last property is true only if no pseudoscalar variable constructed from the @xmath163 momenta is used in the definition of @xmath158 .
a more general analysis is presented in @xcite . ] . for a leptonic decay
, @xmath157 is the normalized energy of the charged lepton , @xmath164 . defining the parameters , @xmath165 which vanish in the standard model , and the functions @xmath166
the decay distribution reads @xmath167h_1(y)+ \tilde\delta p_\tau
h_2(y)+\tilde\eta h_3(y ) \big \}~.\ ] ] the presence of the polarization allows the measurement of a new parameter , @xmath168 , but , as previously mentioned , it also introduces an ambiguity in the interpretation of the first one which is now @xmath169 .
the same kind of ambiguity also arises in the @xmath132-@xmath2 correlation . for @xmath170 ,
only the parameters , @xmath171 , @xmath172 , and the product @xmath173 are measurable @xcite . for @xmath174 , neglecting @xmath175 for the sake of clarity , and keeping only terms of the first order in the standard model violating parameters , @xmath176 , @xmath177 and @xmath178 , the correlated distribution can be written @xmath179h_0(y_\mu)h_1(y_e)}\hspace{4em}\\\nonumber & & + [ \tilde\rho_\mu+p_\tau\xi_\mu]h_0(y_e)h_1(y_\mu ) + [ \xi_\mu+\xi_e+p_\tau(\tilde\rho_e+\tilde\rho_\mu)]h_1(y_\mu)h_1(y_e)\\\nonumber & & + \tilde\delta_e h_2(y_e)[h_1(y_\mu)+p_\tau h_0(y_\mu ) ] + \tilde\delta_\mu h_2(y_\mu)[h_1(y_e)+p_\tau h_0(y_e)]~.\end{aligned}\ ] ] even if @xmath5 is known , there is only three measurements to determine the four parameters @xmath180 , @xmath181 , @xmath182 and @xmath183 . the ambiguity is displaced but not suppressed . using several
known values of the polarization , all the parameters can be determined from single - decay distributions . at the z peak , and/or with polarized beam ,
the @xmath3 polarization is a function of the production angle @xmath6 , hence the michel parameters can be measured by the @xmath6-@xmath184 correlation @xcite .
the other measurements [ 15,22 - 24 ] use all the hadron - hadron , lepton - lepton and hadron - lepton final states to obtain the complete set of parameters up to a global sign ambiguity which is solved , for instance , by the result of @xcite . to calculate the covariance matrix @xmath185 of the michel parameter measurements ,
we assume that , in an ideal experiment at the z peak , similar to [ 15,22 - 24 ] , all the decays into @xmath2 , @xmath132 , @xmath163 , @xmath24 and @xmath186 and their correlations are used to determine the michel parameters , the hadronic , @xmath160 parameters and the @xmath3 polarization .
asymptotically @xmath185 is given by @xmath187 where @xmath188 labels one of the twenty classes of events , @xmath132-@xmath2 , @xmath132-@xmath132 , @xmath132-@xmath163 , etc , and @xmath189 is the number of events in the class .
the computation is made straightforward by the fact that the distributions are quadratic functions of all the parameters , except @xmath25 which appears in the normalizations .
the computed covariance matrix is similar to its estimations @xcite for measurements at lower energy .
the largest correlation coefficients are @xmath197 and @xmath198 .
the numerical values relevant for the analysis of section [ sec : lnp ] are given in table [ tab : covid ] .
for the @xmath199 channel , the inequality @xmath200 is satisfied and the weights of @xmath26 and @xmath134 in their optimal combination are 0.47 and 0.53 respectively .
all the hypotheses made in section [ sec : lnp ] are verified . for the @xmath201 channel ,
the weights are 0.43 and 0.57 but , owing to the correlations with @xmath202 , the error on @xmath177 is slightly smaller than @xmath203 . however , in a more realistic estimation , the inefficiencies in the identification of the various decay channels reduce the statistics for the classes of events with two analysed decays and increase it for the events with only one identified decay which contribute mainly to the measurement of @xmath24 .
therefore , the same analysis scheme remains basically valid .
if the variation of the parameters is limited to the above defined restricted domain , the deviations of the parameters from their standard model values must have the same sign in the @xmath132 and @xmath2 channels .
therefore the universality hypothesis can perhaps reduce the sensitivity to these deviations but complete cancelations are not possible .
10 l. michel , nature * 163 * ( 1949 ) 959 ; proc .
. soc . * 63a * ( 1950 ) 514 . c. bouchiat and l. michel , phys
* 106 * ( 1957 ) 170 . l. okun and a. rudik , sov .
jetp * 5 * ( 1957 ) 520 .
( 1957 ) 844 . c. jarlskog , nucl
* 75 * ( 1966 ) 659 ; nuovo cimento lett .
* iv * ( 1970 ) 377 .
( 1986 ) 102 .
groom et al .
( particle data group ) eur .
j. c * 15 * ( 2000 ) 1 .
( 1990 ) 1544 .
a. stahl , _ physics with tau leptons _
( springer - verlag berlin , heidelberg 2000 ) ( 1978 ) 187 .
f. scheck , _ leptons hadrons and nuclei _ ( north - holland amsterdam 1983 ) k. mursula , m. roos and f. scheck , nucl .
b * 219 * ( 1983 ) 321 ; + k. mursula and f. scheck , nucl .
b * 253 * ( 1985 ) 189 .
p. langacker and d. london , phys .
d * 39 * ( 1989 ) 266 .
a. roug , in _ proc . of 30th rencontres
de moriond : dark matter in cosmology , clocks and tests of fundamental laws , les arcs , france , _
b. guiderdoni et al .
eds , editions frontires 1995 p. 247 . delphi collaboration , p. abreu et al .
j. c * 16 * ( 2000 ) 211 .
a. pich and j .- p .
silva , phys .
d*52 * ( 1995 ) 4006 .
nelson , in tau 94 , nucl .
b ( proc . suppl . )
* 40 * ( 1995 ) 525 .
argus collaboration , h. albrecht et al . ,
b * 349 * ( 1995 ) 576 .
m. davier , l. duflot , f. le diberder and a. roug , phys .
b * 306 * ( 1993 ) 411 .
opal collaboration , k. ackerstaff et al .
phys . j. c * 8 * ( 1999 ) 3 .
collaboration , k. abe et al .
* 78 * ( 1997 ) 4691 .
collaboration , d. buskulic et al .
b * 346 * ( 1995 ) 379 ; * 363 * ( 1995 ) 265 ( erratum ) ; collaboration , _ measurement of the michel parameters and the @xmath204 helicity in @xmath3 lepton decays _
, aleph 98 - 073 ( conf 98 - 042 ) .
collaboration , m. acciarri et al .
b * 438 * ( 1998 ) 405 . | a complete discussion of the constraints on the michel parameters and the ambiguities of their interpretation is presented .
estimators of new physics , optimized for a very wide class of hypotheses and models are proposed .
= -11 mm = -1.5 mm = -1.5 mm + + x - lpnhe 00/01 + |
in the past decade , the `` factor - of - two controversy '' surrounding the value of the hubble constant has been resolved .
however , the uncertainty on h@xmath1 currently remains at the level of a few percent and dominates the error budget when deriving other cosmological parameters such as the equation of state parameter , @xmath2 .
reducing the uncertainty on h@xmath1 will enable us to further constrain the parameters in the concordance cosmological model as @xcite states : `` the single most important complement to the cmb for measuring the dark energy equation of state at @xmath3 is a determination of the hubble constant to better than a few percent . ''
the carnegie hubble program ( chp ) is a warm spitzer " legacy mission ( spitzer exploration program pid 60010 the hubble constant " , p.i .
freedman ) .
it has the primary goal of reducing the systematic uncertainties on the hubble constant to 3% using spitzer data alone , and ultimately 2% with the addition of jwst .
the program uses the @xmath4 and 4.5 @xmath0 m bands of irac @xcite to recalibrate the cepheid distance scale in the mid - infrared , moving out in to the hubble flow by calibrating and applying a revised mid ir tully fisher relation .
a detailed description of the carnegie hubble program is given in @xcite ; an overview of our strategy follows below .
the zero point of the cepheid period luminosity ( pl ) relation ( leavitt law ) will be calibrated using galactic cepheids .
this sample contains the 10 cepheids with trigonometric parallaxes measured using the fine guidance sensors on the _ hubble space telescope _
we also observed 27 cepheids within 4 kpc of the sun close enough for gaia parallax measurements .
seventeen of these are known members of galactic open clusters for which main sequence fitting distances are available .
the milky way program is described in monson et al .
( 2011 , in preparation ) .
the second rung on the distance ladder ( the subject of this work ) involves the long period ( @xmath5 days ) end of the leavitt law .
we will calibrate the slope and dispersion of this period range using a sample from the large magellanic cloud .
the lmc is an important stepping stone for the extragalactic distance scale ; it was used as the base for the key project distance scale @xcite , but uncertainty in its distance eventually became the dominating systematic .
the calibration found here will then be applied to nearby galaxies in the local group , forming a cepheid distance ladder based on a single telescope and instrument .
additionally , by including galaxies such as m33 , the effect of metallicity on the pl slope and zero point will be studied ( e.g. , * ? ? ?
the calibration will then be applied at a step further out , extending to galaxies beyond the local group , such as m81 and ngc 4258 , the well known maser galaxy .
next , we will extend our distance ladder into the hubble flow using the tully fisher relation , itself re - calibrated in the mid infrared .
the dispersion of this relation is expected to drop at irac wavelengths , simply due to the decreased sensitivity to inclination dependent extinction corrections , and will be calibrated using galaxies that contain cepheids .
the distances to over 500 target galaxies in our program will be measured .
finally , we will complete our distance ladder by adding 44 supernova
host field galaxies .
these steps will allow us to measure accurate and consistent distances from the closest cepheids to the most distant galaxies . this work deals with cepheids in the large magellanic cloud ( lmc ) .
the lmc cepheid population has been studied extensively for over half a century .
for example , @xcite presented well sampled light curves for 40 cepheids in the lmc , and study of the population continues to this day with large - scale projects such as ogle iii @xcite .
its proximity means we can readily observe cepheids over a range of periods , from those that are similar to many of the galactic parallax sample ( i.e. @xmath6 days ) , to the long period cepheids with @xmath7 days , which more generally overlap cepheids most easily observed in our most distant targets . until relatively recently
, the majority of distance measurements have been undertaken at optical wavelengths ( e.g. f01 ) .
cepheid studies at optical wavelengths have their shortcomings , the main one being extinction . although the effect can , to first order , be removed by use of the reddening
free wesenheit index @xcite this technique still requires prior knowledge of the extinction law , as well as an assumption that it is universal . by moving to the mid infrared , reddening and extinction
are diminished by around a factor of twenty @xcite , making their absolute contribution and their uncertainties negligible .
in addition to the drop in extinction effects , @xcite describe two other advantages of the infrared over the optical . in the infrared the amplitudes of the cepheids lightcurves decrease , as does the intrinsic width of the instability strip , because these wavelengths are less sensitive to temperature changes .
indeed @xcite demonstrated that the width of the near infrared @xmath8 band period
luminosity relation from _ random _ observations is less than the width from time averaged @xmath9band observations . near
infrared observations of lmc cepheids were more recently obtained by @xcite . by moving to longer wavelengths in the mid infrared , in combination with well phased observations , we can decrease the measured width even further .
the irac imager on _ spitzer _ is a superb instrument for undertaking a recalibration of the cepheid distance scale . like _
hubble _ , _ spitzer _ has the advantage of operating in a stable environment without weather or seeing variations , and with great flexibility in scheduling .
as such , we have been able to obtain precise and deterministically well sampled light curves for 85 cepheids .
the observations of the lmc cepheids were one of the first programs undertaken by post - cryogenic warm spitzer " .
the target selection and observations are described in section [ sec : observations ] and the photometry and calibration are discussed in section [ sec : data_reduction ] .
light and color curves are presented for each cepheid in section [ sec : results ] .
pl relations , including a discussion on their use in determining the tilt of the lmc , are given in section [ sec : pl_relations ] .
the period color relation at mean light has been measured for the first time at these wavelengths ; it is discussed in section [ sec : co_absorption ] .
section [ sec : summary ] provides a summary .
the lmc has a large sample of well studied cepheids to choose from , as light curves have been measured for many thousands .
the ogle studies @xcite , while having a primary goal of observing micro - lensing events , have produced extremely large cepheid catalogs for the lmc and smc in the optical .
for this study , we have chosen a subset of the sample observed by ( * ? ? ? * hereafter p04 ) , who studied 92 lmc cepheids in the infrared @xmath10 bands .
these cepheids also had optical ( @xmath11 ) light curves available .
the majority of them are free from crowding , making them ideal for this study .
we chose the 85 cepheids with periods greater than about 6 days .
this ensures that the period distribution overlaps with both the galactic sample , as well as the most remote galaxies in our program .
the cepheids are distributed over the face of the lmc , allowing any known inclination effects ( c.f . ,
p04 ) to be re - measured and accounted for .
the data were obtained in programs 61000 , 61004 , 61005 , 61006 , and 61007 , executed between 2009 october 3 and 2010 july 18 , at the start of the warm spitzer " mission .
each cepheid was observed at 24 epochs .
for those cepheids with @xmath12 d , the observations were taken over a single cycle , spaced on average by p/24 days .
this produced regularly sampled light curves while preserving the high scheduling efficiency of the spacecraft .
the shortest period cepheids ( p @xmath13 12 d ) could not be observed in such a tightly phased manner ; instead we scheduled 24 observations separated by @xmath14 days .
the appendix explains the choice of 24 evenly - spaced epochs .
each irac observation consisted of ten frames , five in each of the 3.6 @xmath0 m and 4.5 @xmath0 m bands , with a medium - scale , 5-point gaussian dither pattern .
( dithering of exposures is essential for accurate photometry . )
the frame time was 2.0 s , giving an effective per - pixel integration time of 1.2 s in each frame .
irac is known to be susceptible to persistent images ( i.e. latent images of bright , but often unsaturated , stars can remain in later frames ) and this effect is apparently worse in warm spitzer than in the original mission ( see irac handbook , version 2.0 , section 7.2.8 ) . by using a large scale dither pattern , where subsequent observations of the cepheid are spaced by approximately half the chip , we have mitigated the possibility of our photometry being affected by previous observations .
these early warm spitzer " data were some of the first to be reduced with the new warm irac " pipeline .
the images were processed using the spitzer science center pipeline version s18.18.0 , which does not include the artifact correction steps present in the cold mission pipelines .
therefore , in addition to the normal spitzer science center ( ssc ) processing , the bcd ( basic calibrated data ) frames were run through the warm mission column pulldown corrector .
mopex @xcite was used with the standard _ overlap _ and _ mosaic _ name lists for the overlap correction and mosaicking .
the most important objective of this work is to produce an accurate calibration of the cepheid period
luminosity relation .
our photometry is on the same scale as the standard calibration described in @xcite .
they define the flux scale and zero point such that they correspond to a 10-pixel radius aperture , with a sky annulus running from 12 to 20 pixels ( in units of native irac pixels ) with pixel phase and array location corrections applied as necessary .
although the majority of the cepheids in our sample are isolated , several have close neighbors that fall inside the 10-pixel aperture .
these stars could not be measured correctly using aperture photometry . in order to reduce the entire lmc sample in a consistent way , point response function ( prf ) fitting
was used in place of aperture photometry , utilizing the ssc apex software @xcite .
the point response function ( prf ) models the detector response to a point source , and can be thought of as an over - sampled representation of the point spread function . the bcds were mosaicked to obtain the bad pixel maps , but the prf fitting was performed on each bcd individually using the _ apex_1frame _ name list , with two changes .
first , rather than using the prf corresponding to the center of the array , the set of 25 spatially
dependent prfs was used for each channel .
the irac detectors are known to have some spatial response variations ; using the prf map ensured that any variation in flux was properly accounted for .
second , the normalization radius the radius to which the fit is calculated was set to 1000 pixels . as the prf is sampled at 100 times per pixel
, the 1000 pixel normalization radius corresponds to the calibration aperture radius . for each bcd , all the stars in each image were measured .
the construction of the irac prf is a complex task requiring a large number of observations of a stable star .
warm spitzer prf models are not yet available from the spitzer science center ; however , it is still possible to obtain accurate photometry using the cold prf , correcting as necessary for the predicted differences between the cold and warm models .
the most significant difference between the two models is the pixel phase correction ( ppc ) .
the ppc is necessary given that the response of the irac detectors is known to vary on a sub pixel level .
this means that the measured flux of an object not only depends on the position of the star in the frame , but also on the position of the peak of the psf within the pixel .
the fractional position within a pixel is known as the pixel phase " . in the cryogenic mission
the pixel
phase effect was around 4% ( peak to peak ) , but is approximately double that in the warm mission .
in addition to the change in magnitude , whereas the cold mission ppc was represented by a symmetric gaussian centered on the peak response position , the warm mission ppc is described by a gaussian , differing in width in the @xmath15 and @xmath16 directions .
as the prf is an oversampled representation of the pixel s response to light , it includes the cold mission ppc by definition .
to correct the warm mission data , the cold ppc must first be multiplied out of the apex photometry results , and the warm correction then divided in .
although it is possible that the pixel phase effect will cancel out when the results from a dither pattern are averaged together , in our data the ppc was applied to each bcd before the averaging took place .
the stability of irac is of fundamental importance to our program .
we have quantified the stability of the instrumental zero
point by measuring constant stars in our cepheid fields , and monitoring their magnitudes over time .
the photometry method was identical to that used on the cepheid targets . in figure
[ fig : constant_stability ] we show the photometry for constant stars in two fields .
star a was chosen in the hv00872 field , at @xmath17 , @xmath18 , and star b was in the hv02883 field at @xmath19 , @xmath20 . in both cases ,
the dispersion of the measurements is approximately @xmath21 mag , and no systematic trend with time is seen .
the cepheids in the fields have very different periods ( 29.82 d vs. 108.97 d ) , hence the time baseline and the cadence of the observations differ .
this confirms both the short and long term stability of irac .
as mentioned above , the irac magnitude system is defined by the flux measured within a 10-pixel aperture . to verify that the photometry for this work is on the calibrated system
, the data for the cepheid hv00872 were also processed using the ssc idl routine _ photo_for_star.pro _ ( ssc 2010 , private communication ) .
this routine performs aperture photometry on a single star using two aperture sizes a 3 pixel aperture with a 37 pixel sky annulus , and the 10 pixel aperture with a 1220 pixel sky annulus used in the calibration definition given by @xcite .
it also includes the latest pixel phase corrections and array location corrections .
the apex photometry was compared to the results from both aperture sizes .
no systematic trends with magnitude were found .
the difference between the apex and the 3-pixel aperture fluxes were consistent with the published aperture correction for transformations between the 3 and 10 pixel apertures .
both the apex and small aperture results were then compared with the large aperture values . as was expected , the scatter in the large aperture results was much larger than the other two methods , but no systematics were found .
aperture corrections were derived for each band , and agreed with the cold mission values ( 1.124 and 1.127 at 3.6 @xmath0 m and 4.5 @xmath0 m respectively ) to within 1% . to put the prf photometry on the same system as the standard aperture photometry
the following correction was used : @xmath22 where @xmath23 is the flux in the standard system , @xmath24 is the flux output by apex , @xmath25 is the aperture correction ( 1.03 and 1.02 for channels 1 and 2 respectively ) and @xmath26 and @xmath27 are the cold and warm pixel phase corrections .
the fluxes were converted to magnitudes using the standard zero magnitude flux densities of 280.9 jy ( [ 3.6 ] ) and 179.7 jy ( [ 4.5 ] ) .
we first refined the available periods as follows .
light curves were produced in the 3.6 and 4.5 @xmath0 m bands .
the data points from @xcite , taken as part of the sage project @xcite , were also phased in .
the periods given in p04 were taken as initial values ; these were improved by iterating on the period and re - phasing the data until a minimum dispersion in the light curves was obtained in all available wavebands ( @xmath28,[4.5],[5.8],[8.0]$ ] , but note that not all wavelengths were available for all cepheids ) . on average , the periods changed by less than 1% . in some cases it was clear that the period of the cepheid had changed in the time interval between the optical and infrared observations .
this is to be expected , as we utilized as much archival data as possible ; the time baseline for some cepheids was of the order of many decades , which is comparable to the time many cepheids have been found to significantly change their periods ( for a discussion of period changes in a sample of lmc cepheids , see * ? ? ?
the practice of fitting all wavebands simultaneously makes it easy to find objects for which this is the case , as the period that fits the later observations will result in an incorrectly phased light curve for the archival data , and vice versa .
when this was the case the irac and @xmath29 data alone were used to refine the period and to phase the data .
the finally adopted periods are given in table [ tab : mag_results ] .
once the periods had been adjusted , the irac light curves were plotted using just the chp data .
this avoids issues such as slight period changes and larger photometric uncertainties in the archival data that could affect the irac fit .
the resulting light curves are shown in figure [ fig : light_curves ] in order of increasing period . the gloess program ( described in section 3.2 of p04 )
was used to interpolate between the data points and to thus draw the smooth curves .
gloess was also used to determine time averaged ( intensity mean ) magnitudes and uncertainties .
the results are given in table [ tab : mag_results ] .
for the @xmath30 $ ] ( top ) and @xmath31 $ ] ( middle ) panels , the statistical uncertainties on the individual measurements are comparable to the size of the plotted points .
table [ tab : example_photometry_hv872 ] gives a sample of the individual data points for hv00872 ; all data for this program are available with the online version of the article .
the bottom panels of the plots in figure [ fig : light_curves ] show the @xmath30-[4.5]$ ] color variation through the cepheid pulsation cycle . as light curves of this quality
have not previously been available in the near - ir , this cyclical effect has not been noted before . the effect is almost certainly due to absorption by the 4.65 @xmath0 m co bandhead in the 4.5 @xmath0 m filter , as first suggested by @xcite .
the color variations are discussed below in section [ sec : color_curves ] .
period luminosity ( pl ) relations for the two irac bands were fitted using the weighted least squares technique described in @xcite .
this takes into account the intrinsic width of the pl relation , giving a more robust result than a regular weighted
squares fit .
the pl relations take the form @xmath32 and were fit using four different period cuts .
the results of the different fits are given in table [ tab : pl_relations ] .
the relations derived from the two phase points available from sage @xcite are also given for comparison . in table
[ tab : pl_relations ] , @xmath33 is the number of cepheids in the fit , @xmath34 and @xmath35 were both found using the @xcite weighted least squares method , @xmath36 and @xmath37 are their respective uncertainties , and @xmath38 is the standard deviation of the fit .
the finally adopted relations are plotted in figure [ fig : pl_relations ] .
the solid line shows the adopted pl relations ; these were the fits that used only cepheids in the range @xmath39 d @xmath40 d. note the systematic deviation of the longest period cepheids , which are well known to be significantly fainter than the pl relation at virtually all wavelengths .
consistent with other studies ( such as f01 ) we have excluded objects with @xmath41 days from the fitting .
the sage pls ( given in table [ tab : pl_relations ] ) are entirely consistent with the values found in this work .
although we find no evidence in our data to suggest a break or change in slope in the @xmath42 days region ( c.f .
, ) , we have also cut the low end of the period distribution , again consistent with f01 . in both cases ,
the dispersion around the relation has fallen dramatically in comparison with previously published determinations .
the previous measurements from sage gave standard deviations of 0.135 and 0.141 mag in [ 3.6 ] and [ 4.5 ] respectively , whereas the scatter around our final relations are 0.108 and 0.115 mag .
this is because the magnitudes used in the @xcite pl relation are not fully phase averaged , but rather they are the simple mean of two random phase points ( see the appendix ) . the precision of our phase averaged light curves is best demonstrated by comparing the residuals from the [ 3.6 ] pl relation to those from the [ 4.5 ] relation .
this is shown in figure [ fig : pl_residuals ] . in this figure
the small inset plot shows the entire data set , with the main plot showing just the objects with @xmath43 mag @xmath44 , which encompassed 80 of the 85 cepheids in the sample .
it is clear that the residuals are highly correlated .
that is , if a star lies above the mean pl relation in one band , it will tend to be above the mean in the other band as well . the same effect in the @xmath45 and @xmath46 bands
was found by p04 . excluding two outliers
, the range in @xmath43 mag is @xmath47 mag .
this spread is a combination of two effects : depth and tilt effects in the structure of the lmc , and the intrinsic magnitude width of the instability strip .
the residual dispersion in this correlation is only 0.02 mag , suggesting that we are seeing the photometric limit of the data .
figure [ fig : colour_mag_residuals ] shows how the residuals of the @xmath48-[4.5])$ ] color correlate with the pl residuals of the @xmath30 $ ] magnitude .
as we know from the previous plot that any dispersion in figure [ fig : colour_mag_residuals ] arises from the intrinsic properties of the cepheids rather than photometric uncertainties , the width of the instability strip can be measured from this figure .
assuming that the instability strip is represented by a rectangular distribution , i.e. it is uniformly filled and has hard limits at the blue and red edges , the dispersion around the mean is related to the width by the relation @xmath49 , where @xmath50 is the full width of the distribution .
the color dispersion of the data is @xmath51 mag , leading to a color width of the instability strip of 0.062 mag . by combining our knowledge of the cepheid s location within the lmc with its position in the period
luminosity plane , the tilt of the galaxy with respect to the plane of the sky can be constructed , as was done in p04 . in figure
[ fig : tilt_plot ] we show the positions of the cepheids within the lmc in projection on the sky .
the bottom left panel shows the position of each of the cepheids relative to the center of the lmc . solving for pl coefficients @xmath34 and @xmath35 , but now including @xmath52 and @xmath53 ( distances from the center of the lmc in ra and dec ) as free parameters , we can obtain the tilt and the angle of the line of nodes if the cepheids on average lie in a plane .
the top left and bottom right plots show the distribution of the points when viewed from orthogonal directions in the plane of the sky .
empty circles represent the cepheids that are brighter than the average ( i.e. they lie above the mean pl relation ) , and filled circles show those that are fainter than the average ( below the mean pl relation ) . in this figure
the reference coordinate system has been rotated by 50@xmath54 so that the line of nodes of the lmc is parallel with the @xmath15axis . in this plot , it is clear that the cepheids in the bottom half tend to be fainter than the pl , hence further away , while those in the top half tend to be systematically brighter , hence closer than average .
the simplest interpretation is that the lmc is in fact tilted with respect to the plane of the sky .
by tilting the lmc by @xmath55 relative to the plane of the sky we can minimize the width of the histogram in the top right corner of fig .
[ fig : tilt_plot ] .
this is in good agreement with the results of p04 , and will be the subject of a more complete discussion of the data at a later time .
we believe that the variations in the @xmath30-[4.5]$ ] color , both at mean light and throughout the pulsation cycle , are due to co absorption in the 4.5 @xmath0 m band .
we do not see any evidence of extended emission surrounding the cepheids in our images .
this is consistent with the results of @xcite and @xcite who found evidence for circumstellar emission in the longer wavelength irac bands , but nothing at 3.6 or 4.5 @xmath0 m .
@xcite notes that the @xmath30-[4.5]$ ] color of the cepheids is indistinguishable from their non variable control stars , and is therefore caused by an intrinsic feature of stars of this spectral type , regardless of whether or not they are cepheids .
this rules out circumstellar envelopes and mass loss as the source of the feature , leaving only the co bandhead .
the possibility of a period color ( pc ) relationship at mid
infrared wavelengths was correctly noted by @xcite .
however , their magnitudes ( and therefore colors ) were measured at only a single epoch , introducing large scatter into the pc relation , as is shown in their figure 5 .
figure [ fig : pc_relation ] shows the relationship between period and @xmath30-[4.5]$ ] color for our sample .
the period color relation was fit using cepheids with @xmath56 , and is found to be @xmath57 - [ 4.5 ] = -0.087 ( \pm 0.012 ) ( \log p - 1 ) + 0.005 ( \pm 0.005)\text{. } \label{eqn : pc_relation}\ ] ] the lmc cepheids clearly exhibit a period color relation in this period range .
the standard deviation around the relation is @xmath58 mag .
the relation appears to invert for the longest
period ( @xmath59 ) cepheids . ignoring the @xmath59 sample ,
the long period cepheids are brighter and bluer in @xmath30-[4.5]$ ] than their short period counterparts . at longer periods
the cepheids are at cooler average temperatures , and so the carbon and oxygen are situated in co molecules .
as discussed previously , the co absorbs some of the 4.5 @xmath0 m flux , making the star appear bluer .
as we move to shorter periods the average temperature rises , the co is gradually dissociated , and the average colors become redder . the period color relation reverses at @xmath60 days .
we do not know what causes this reversal , but the cepheids at these periods are deviant in their luminosities and colors at almost all wavelengths . the period - color relation is closely related to the variations in color shown in figure [ fig : light_curves ] , as they are both due to co affecting the 4.5 @xmath0 m flux . in a static stellar atmosphere , with a temperature and surface gravity corresponding to an f type supergiant , the 4.5 @xmath0 m flux is suppressed . in the case of cepheids , however , we are not dealing with static atmospheres .
as the cepheid contracts its temperature increases .
as the star becomes hotter , co in the atmosphere dissociates , causing less absorption at 4.5 @xmath0 m , making the star then appear redder than average .
once the cepheid begins to expand again its temperature drops , allowing the c and o to recombine .
as more co is present , the 4.5 @xmath0 m flux is once again suppressed , making the star bluer , in the particular color combination ( @xmath30-[4.5]$ ] ) discussed here .
inspection of figure [ fig : light_curves ] shows that the phasing of the light and color curves exhibits precisely this behavior .
there also appears to be a relation between period and the amplitude of the color variation , such that the short period cepheids ( @xmath6 days ) tend to have flat color curves .
the cyclical variation appears to turn on " around @xmath42 days , increasing in amplitude with period .
again , this is a temperature effect .
the short period cepheids are so hot that the co is always dissociated , whereas the longest period cepheids have lower maximum temperatures and so never dissociate all their co , thus have the largest color amplitude .
the dependence of color on co implies that the 4.5 @xmath0 m band should be subject to a metallicity effect , effectively ruling it out as a distance indicator .
fortunately , the bandhead is confined to the [ 4.5 ] filter alone ; no co or other molecular feature is present at 3.6 @xmath0 m , and so it can still be confidently used to measure cepheid distances .
the physics behind the cyclical co variation , along with the effect of metallicity on cepheid s irac colors will be described in detail in scowcroft et al .
( 2012 , in preparation ) .
we now return to consideration of the slopes of the pl at different wavelengths .
figure [ fig : pl_slopes ] shows the dependence of the slope on wavelength , a plot that is repeated from figure 4 of @xcite .
the @xmath30 $ ] and @xmath31 $ ] slope values have been updated using the table 3 values in this paper .
the slope of the @xmath31 $ ] pl relation is clearly deviant .
we can again understand this as due to co as follows : as we move from shorter to longer periods we are moving to cooler stars ; these have more co absorption on average .
more co means depressed 4.5 @xmath0 m flux and thus a shallower pl slope . therefore , this plot gives further evidence that the color curves are due to co absorption that affects only @xmath31 $ ] ; the @xmath30 $ ] magnitude is unaffected .
a smaller magnitude effect is also seen in the @xmath61 band where a shallower co bandhead is present .
however , this effect can be mitigated using the @xmath46 band , which excludes the longest wavelength section of the bandpass the part that happens to include the co bandhead .
we have used the irac camera on warm spitzer " to measure [ 3.6 ] and [ 4.5 ] light curves for 85 cepheids spread across the face of the lmc and covering a 6 to 140 day period range .
each star has been measured at 24 phase points .
our data agree well with the random phase data of the sage project , the latter taken for different purposes .
the light curves ( shown in figure [ fig : light_curves ] ) allow us to measure the mean magnitudes and color of each cepheid to a precision never achieved before in the mid - infrared .
these lead to our best pl relations based on the measurements of the [ 3.6 ] and [ 4.5 ] magnitudes for 67 stars with @xmath62 days .
the pl relations have been measured to such high precision that their scatter directly correlates with the front - to - back geometry of the sample of cepheids in the lmc .
this correlation recalls a similar effect presented in p04 .
using the residuals of the period luminosity relation , we have measured the tilt of the galaxy as 28@xmath54 , with respect to the plane of the sky .
the dense and uniform sampling of our cepheid light curves reveals a cyclical color variation which we interpret as being due to the effect of temperature sensitive molecular absorption of co in the cepheid s atmosphere .
when the cepheid is cool , the co present in its atmosphere absorbs at 4.67 @xmath0 m , suppressing flux exclusively in the irac 4.5 @xmath0 m band , making the star appear blue in the @xmath30-[4.5]$ ] color . as the cepheid contracts , and so heats up , the co molecules begin to dissociate , ending the suppression of the @xmath63 m flux and reddening the color .
the temperature dependence of the co absorption also explains the period color relation we observe ( see figure [ fig : pc_relation ] ) : the cooler , longer period cepheids appear bluer ( more highly absorbed by co at 4.5 @xmath0 m ) than their short period counterparts .
the implications of this temperature sensitive effect will be studied in detail in scowcroft et al .
( 2012 , in preparation ) .
the period luminosity relations measured here will be used in combination with those from our galactic sample ( monson et al .
2011 , in preparation ) , to establish the bottom rung of the extragalactic distance ladder .
they will then be combined with parallel observations of cepheids in the local group and beyond , to establish the distance scale to a systematic accuracy not achievable before now .
we will also be testing for the effect of metallicity on the mid ir period
color relation , by combining this sample with galactic and smc cepheids , ( scowcroft et al .
2012 , in preparation ) .
this work is based in part on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
support for this work was provided by nasa through an award issued by jpl / caltech .
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 .
we would like to thank the staff of the spitzer science center , and in particular nancy silbermann , for their assistance with scheduling such a large and complex project . without their help , this project would not have been possible .
the observing plan we adopted using _ spitzer _ to optimally follow up known cepheids is very differently constrained from what one would or could attempt from the ground .
ground - based observations are obviously restricted by 24-hour resonances with the day - night observing cycle .
they are often additionally modulated by the lunar cycle ( especially for optical observations ) .
they are stochastically interrupted by weather , instrument availability , technical problems , and time assignments . randomly spaced observations of known cepheids
may thus be the only way to fully sample light curves and to determine their mean values . in the case of space - based observations of known cepheids
randomly sampled data may be the only kind available ( e.g. data from the sage project @xcite , which was mined for cepheids by @xcite and @xcite ) .
however , the 24-hour availability of a solar - orbiting telescope opens up a unique opportunity to design the observing program to be optimally scheduled on a star - by - star basis .
let us first consider the case of random sampling . in randomly sampling any given distribution function the error on the derived mean value goes down in inverse proportion to the square root of the number n of randomly timed observations ( i.e. , @xmath64 ) .
what is not widely appreciated is that convergence of the error on the mean can be very significantly accelerated by deterministic ( non - random ) sampling . in the case of an individual cepheid where the time distribution of luminosities ( i.e. , its amplitude ) is bounded ( but not _ a priori _ known ) , and where the cycling time ( i.e. , its period , but not necessarily its phase ) is also known , then it is possible ( see * ? ? ?
* ) to construct a simple observing strategy where the errors on the mean decrease in direct inverse proportion with the number of observations ( i.e. @xmath65 ) .
not surprisingly , for a specified number of observations n , and a known period of variation p , the optimal separation of data points in time is the uniform distribution p/(n+1 ) .
the distribution function in this case is closely approximated by a rectangular probability density distribution .
the equivalent sigma for such a distribution is @xmath66 , where @xmath25 is the total range of the distribution function ; in the case of a cepheid , @xmath25 is its amplitude . in the mid - ir
the amplitudes of cepheids can reach 0.8 mag ( e.g. hv00883 ) , which converts to an equivalent sigma of @xmath67 mag . if the required precision on a determination of the mean magnitude were to be set at 0.01 mag ( i.e. , one percent ) then random sampling of the light curve would require over 500 observations ! to reach the same precision using deterministic ( uniform ) sampling a grand total of 23 observations are required . to compensate for slight variations away from exact scheduling we requested and obtained 24 observations per lmc cepheid .
the light curves shown in figure [ fig : light_curves ] are a testament to the care and efficiency with which _ spitzer _ was scheduled for this program .
akritas m. g. , & bershady m. a. 1996 , , 470 , 706 barmby p. , marengo m. , evans n. r. , et al .
2011 , , 141 , 42 benedict g. f. , mcarthur b. e. , feast m. w. , et al .
2007 , , 133 , 181 bird j. c. , stanek k. z , & prieto j. l. 2009 , , 695 , 874 fazio g. g. , hora j. l. , allen l. e. , et al .
2004 , , 154 , 10 freedman w. l. , hughes s. m. , madore b. f. , et al .
1994 , , 427 , 628 freedman w. l. , madore b. f. , gibson b. k. , et al .
2001 , , 553 , 47 ( f01 ) freedman w. l. , madore b. f. , rigby j. , persson s. e. , & sturch l. 2008 , , 679 , 71 freedman w. l. , & madore b. f. 2010 , araa , 48 , 673 freedman w. l. , madore b. f. , scowcroft v. , et al .
2011 , aj , submitted grieve g. r. , madore b. f. , & welch d. l. 1985 , , 194 , 513 hu w. 2005 , in asp conf .
339 , observing dark energy , ed .
s. c. wolff , & t. r. lauer ( san francisco , ca ; asp ) , 215 madore b. f. 1976 , rgob , 182 , 153 madore b. f. & freedman w. l. 1998 , in stellar astrophysics for the local group : viii canary islands winter school of astrophysics , ed a. aparicio , a. herrero , & f. snchez , cambridge university press , 263 madore b. f. , & freedman w. l. 2005 , , 630 , 153 madore b. f. , freedman w. l. , rigby j. , et al .
2009 , , 695 , 988 makovoz d. , & khan i. 2005 , in asp conf . ser . 347 , astronomical data analysis software and systems xiv , ed .
p. shopbell , m. britton , & r. ebert , 81 makovoz d. , & marleau f. r. 2005 , , 117 , 1113 marengo m. , evans n. r. , barmby p. , et al . 2010a , , 709 , 120 marengo m. , evans n. r. , barmby p. , et al . 2010b , 725 , 2392 mcgonegal r. , mclaren r. a. , mcalary c. w. & madore b. f. 1982 , , 257 , l33 meixner m. , gordon k. d. , indebetouw r. , et al .
2006 , , 132 , 2268 ngeow c. , & kanbur s. m. 2008 , , 679 , 76 persson s. e. , madore b. f. , krzemiski w. , et al .
2004 , , 128 , 2239 ( p04 ) pietrukowicz p. 2001 , aca , 51 , 247 reach w. t. , megeath s. t. , cohen m. , et al .
2005 , , 117 , 978 rieke g. h. , & lebofsky m. j. 1985 , , 288 , 618 riess a. g. , macri l. , casertano s. , et al .
2011 , , 730 , 119 sandage a. , tammann g. a. , & reindl b. 2004 , , 424 , 43 scowcroft v. , bersier d. , mould j. r. , & wood p. r. 2009 , , 396,1287 shapley h. , & nail v. m. 1940 , anhar , 90 , 253 soszynski i. , poleski r. , udalski a. , et al .
2008 , aca , 58 , 163 spergel d. n. , bean r. , dor o. , et al .
2007 , , 170 , 377 udalski a. , szymanski m. , kaluzny j. , kubiak m. , & mateo m. 1992 , aca , 42 , 253 udalski a. , kubiak m. , & szymanski m. 1997 , aca , 47 , 319 limits .
the dispersion in both cases is approximately 0.01 mag .
the time baselines differ as the cepheids in the fields were observed with different cadences .
no systematic trends are seen.,width=566 ] @xmath68{f2_1.eps } & \includegraphics[width=50mm , angle=270]{f2_2.eps } & \includegraphics[width=50mm , angle=270]{f2_3.eps } \\ \includegraphics[width=50mm , angle=270]{f2_4.eps } & \includegraphics[width=50mm , angle=270]{f2_5.eps } & \includegraphics[width=50mm , angle=270]{f2_6.eps } \\ \includegraphics[width=50mm , angle=270]{f2_7.eps } & \includegraphics[width=50mm , angle=270]{f2_8.eps } & \includegraphics[width=50mm , angle=270]{f2_9.eps } \\ \end{array}$ ] @xmath68{f2_10.eps } & \includegraphics[width=50mm , angle=270]{f2_11.eps } & \includegraphics[width=50mm , angle=270]{f2_12.eps } \\ \includegraphics[width=50mm , angle=270]{f2_13.eps } & \includegraphics[width=50mm , angle=270]{f2_14.eps } & \includegraphics[width=50mm , angle=270]{f2_15.eps } \\ \includegraphics[width=50mm , angle=270]{f2_16.eps } & \includegraphics[width=50mm , angle=270]{f2_17.eps } & \includegraphics[width=50mm , angle=270]{f2_18.eps } \\ \end{array}$ ] @xmath68{f2_19.eps } & \includegraphics[width=50mm , angle=270]{f2_20.eps } & \includegraphics[width=50mm , angle=270]{f2_21.eps } \\ \includegraphics[width=50mm , angle=270]{f2_22.eps } & \includegraphics[width=50mm , angle=270]{f2_23.eps } & \includegraphics[width=50mm , angle=270]{f2_24.eps } \\ \includegraphics[width=50mm , angle=270]{f2_25.eps } & \includegraphics[width=50mm , angle=270]{f2_26.eps } & \includegraphics[width=50mm , angle=270]{f2_27.eps } \\ \end{array}$ ] @xmath68{f2_28.eps } & \includegraphics[width=50mm , angle=270]{f2_29.eps } & \includegraphics[width=50mm , angle=270]{f2_30.eps } \\ \includegraphics[width=50mm , angle=270]{f2_31.eps } & \includegraphics[width=50mm , angle=270]{f2_32.eps } & \includegraphics[width=50mm , angle=270]{f2_33.eps } \\ \includegraphics[width=50mm , angle=270]{f2_34.eps } & \includegraphics[width=50mm , angle=270]{f2_35.eps } & \includegraphics[width=50mm , angle=270]{f2_36.eps } \\ \end{array}$ ] @xmath69{f2_37.eps } & \includegraphics[width=50mm , angle=270]{f2_38.eps } & \includegraphics[width=50mm , angle=270]{f2_39.eps } \\ \includegraphics[width=50mm , angle=270]{f2_40.eps } & \includegraphics[width=50mm , angle=270]{f2_41.eps } & \includegraphics[width=50mm , angle=270]{f2_42.eps } \\ \includegraphics[width=50mm , angle=270]{f2_43.eps } & \includegraphics[width=50mm , angle=270]{f2_44.eps } & \includegraphics[width=50mm , angle=270]{f2_45.eps } \\ \end{array}$ ] @xmath68{f2_46.eps } & \includegraphics[width=50mm , angle=270]{f2_47.eps } & \includegraphics[width=50mm , angle=270]{f2_48.eps } \\ \includegraphics[width=50mm , angle=270]{f2_49.eps } & \includegraphics[width=50mm , angle=270]{f2_50.eps } & \includegraphics[width=50mm , angle=270]{f2_51.eps } \\ \includegraphics[width=50mm , angle=270]{f2_52.eps } & \includegraphics[width=50mm , angle=270]{f2_53.eps } & \includegraphics[width=50mm , angle=270]{f2_54.eps } \\ \end{array}$ ] @xmath68{f2_55.eps } & \includegraphics[width=50mm , angle=270]{f2_56.eps } & \includegraphics[width=50mm , angle=270]{f2_57.eps } \\ \includegraphics[width=50mm , angle=270]{f2_58.eps } & \includegraphics[width=50mm , angle=270]{f2_59.eps } & \includegraphics[width=50mm , angle=270]{f2_60.eps } \\ \includegraphics[width=50mm , angle=270]{f2_61.eps } & \includegraphics[width=50mm , angle=270]{f2_62.eps } & \includegraphics[width=50mm , angle=270]{f2_63.eps } \\ \end{array}$ ] @xmath68{f2_64.eps } & \includegraphics[width=50mm , angle=270]{f2_65.eps } & \includegraphics[width=50mm , angle=270]{f2_66.eps } \\ \includegraphics[width=50mm , angle=270]{f2_67.eps } & \includegraphics[width=50mm , angle=270]{f2_68.eps } & \includegraphics[width=50mm , angle=270]{f2_69.eps } \\ \includegraphics[width=50mm , angle=270]{f2_70.eps } & \includegraphics[width=50mm , angle=270]{f2_71.eps } & \includegraphics[width=50mm , angle=270]{f2_72.eps } \\
\end{array}$ ] @xmath68{f2_73.eps } & \includegraphics[width=50mm , angle=270]{f2_74.eps } & \includegraphics[width=50mm , angle=270]{f2_75.eps } \\
\includegraphics[width=50mm , angle=270]{f2_76.eps } & \includegraphics[width=50mm , angle=270]{f2_77.eps } & \includegraphics[width=50mm , angle=270]{f2_78.eps } \\
\includegraphics[width=50mm , angle=270]{f2_79.eps } & \includegraphics[width=50mm , angle=270]{f2_80.eps } & \includegraphics[width=50mm , angle=270]{f2_81.eps } \\ \end{array}$ ] and [ 4.5 ] period luminosity relations , with short period ( @xmath72 days ) cepheids as red open stars , long period ( @xmath73 days ) as blue open triangles , and the pl fitting sample as black filled circles .
the inset plot shows the whole sample ; the main plot is cut at @xmath74 mag to show an expanded view of the details .
the dashed line indicates @xmath75 = \delta[4.5]$ ] , the solid line is a fit to the black points .
the deviation from a one - to - one fit is believed to be due to co pushing the [ 4.5 ] points further from the adopted pl . , width=566 ] dispersion . excluding the longest period ( @xmath59 ) stars , the color dispersion around the [ 3.6 ] ridge line is 0.018 mag .
assuming that the instability strip has a regularly filled , rectangular distribution , the width will be @xmath76 .
this gives the color width of the instability strip as 0.062 mag .
, width=566 ] and @xmath53 components .
the top right plot shows the histogram of the size of the pl deviations .
the dispersion in the pl is minimized by assuming a tilt of @xmath55 in @xmath53 relative to the plane of the sky .
, width=566 ] ) , p04 ( @xmath29 ) , this work ( [ 3.6 ] , [ 4.5 ] ) and @xcite ( [ 5.8 ] , [ 8.0 ] ) .
there is an asymptotic trend towards @xmath79 , which is consistent with the theoretical prediction of a slope of @xmath80 for the period radius relation .
however , the @xmath31 $ ] ( and to a smaller extent , @xmath61 ) slope is clearly discrepant .
we attribute this to the co bandhead that suppresses flux in the 4.5 @xmath0 m .
the [ 3.6 ] slope is not affected it follows the trend of the other wavelengths and is therefore appropriate to use for distance determinations . , width=566 ]
l c c c c c c c hv00872 & 29.820 & 11.320 & 0.005 & 11.381 & 0.005 & -0.061 & 0.002 + hv00873 & 34.426 & 10.872 & 0.006 & 10.903 & 0.006 & -0.029 & 0.002 + hv00875 & 30.338 & 11.094 & 0.003 & 11.077 & 0.003 & 0.016 & 0.001 + hv00876 & 22.716 & 11.558 & 0.006 & 11.592 & 0.007 & -0.033 & 0.002 + hv00877 & 45.155 & 10.701 & 0.004 & 10.790 & 0.004 & -0.089 & 0.002 + hv00878 & 23.306 & 11.531 & 0.007 & 11.559 & 0.007 & -0.027 & 0.002 + hv00879 & 36.840 & 10.923 & 0.007 & 10.976 & 0.007 & -0.052 & 0.002 + hv00881 & 35.740 & 10.945 & 0.007 & 10.996 & 0.007 & -0.051 & 0.002 + hv00882 & 31.786 & 11.070 & 0.006 & 11.086 & 0.006 & -0.016 & 0.002 + hv00883 & 133.400 & 9.523 & 0.011 & 9.479 & 0.010 & 0.044 & 0.001 + hv00885 & 20.703 & 11.368 & 0.003 & 11.395 & 0.003 & -0.026 & 0.001 + hv00886 & 23.983 & 11.346 & 0.006 & 11.373 & 0.006 & -0.027 & 0.002 + hv00887 & 14.486 & 12.113 & 0.004 & 12.123 & 0.004 & -0.010 & 0.001 + hv00889 & 25.797 & 11.374 & 0.006 & 11.432 & 0.005 & -0.057 & 0.003 + hv00891 & 17.272 & 11.861 & 0.005 & 11.873 & 0.005 & -0.012 & 0.002 + hv00892 & 15.990 & 12.185 & 0.006 & 12.191 & 0.006 & -0.004 & 0.001 + hv00893 & 21.115 & 11.638 & 0.006 & 11.673 & 0.006 & -0.035 & 0.002 + hv00899 & 31.047 & 11.132 & 0.007 & 11.185 & 0.007 & -0.051 & 0.002 + hv00900 & 47.515 & 10.510 & 0.006 & 10.547 & 0.006 & -0.037 & 0.002 + hv00901 & 18.470 & 11.989 & 0.006 & 12.020 & 0.005 & -0.033 & 0.002 + hv00902 & 26.340 & 11.269 & 0.006 & 11.281 & 0.006 & -0.012 & 0.002 + hv00904 & 30.405 & 11.174 & 0.007 & 11.216 & 0.007 & -0.042 & 0.002 + hv00909 & 37.566 & 10.795 & 0.005 & 10.808 & 0.005 & -0.012 & 0.001 + hv00911 & 13.915 & 12.290 & 0.005 & 12.297 & 0.005 & -0.007 & 0.002 + hv00914 & 6.879 & 13.122 & 0.011 & 13.088 & 0.009 & 0.032 & 0.005 + hv00932 & 13.281 & 12.300 & 0.004 & 12.304 & 0.005 & -0.005 & 0.002 + hv00953 & 48.048 & 10.220 & 0.005 & 10.249 & 0.005 & -0.025 & 0.002 + hv00955 & 13.700 & 12.173 & 0.004 & 12.177 & 0.004 & -0.004 & 0.001 + hv00971 & 9.297 & 12.681 & 0.015 & 12.680 & 0.018 & -0.001 & 0.005 + hv00997 & 13.145 & 12.356 & 0.005 & 12.367 & 0.005 & -0.012 & 0.002 + hv01002 & 30.472 & 10.990 & 0.006 & 11.017 & 0.007 & -0.026 & 0.002 + hv01003 & 24.339 & 11.295 & 0.005 & 11.300 & 0.005 & -0.004 & 0.001 + hv01005 & 18.716 & 11.922 & 0.007 & 11.897 & 0.006 & 0.023 & 0.001 + hv01006 & 14.217 & 12.162 & 0.005 & 12.172 & 0.005 & -0.010 & 0.001 + hv01013 & 24.131 & 11.392 & 0.006 & 11.458 & 0.005 & -0.064 & 0.002 + hv01019 & 13.659 & 12.252 & 0.003 & 12.253 & 0.003 & 0.001 & 0.001 + hv01023 & 26.559 & 11.386 & 0.006 & 11.441 & 0.006 & -0.052 & 0.002 + hv02244 & 13.977 & 12.219 & 0.004 & 12.230 & 0.005 & -0.011 & 0.002 + hv02251 & 27.887 & 11.124 & 0.007 & 11.166 & 0.007 & -0.043 & 0.002 + hv02257 & 39.395 & 10.722 & 0.006 & 10.784 & 0.006 & -0.062 & 0.002 + hv02260 & 12.987 & 12.633 & 0.005 & 12.620 & 0.005 & 0.012 & 0.001 + hv02262 & 15.832 & 12.035 & 0.004 & 12.067 & 0.004 & -0.034 & 0.002 + hv02270 & 13.625 & 12.399 & 0.004 & 12.424 & 0.004 & -0.025 & 0.001 + hv02279 & 6.895 & 13.077 & 0.013 & 13.050 & 0.014 & 0.023 & 0.002 + hv02282 & 14.677 & 12.218 & 0.005 & 12.221 & 0.005 & -0.003 & 0.002 + hv02291 & 22.317 & 11.667 & 0.006 & 11.706 & 0.006 & -0.036 & 0.002 + hv02294 & 36.552 & 10.663 & 0.006 & 10.700 & 0.006 & -0.036 & 0.002 + hv02324 & 14.466 & 12.205 & 0.005 & 12.221 & 0.004 & -0.017 & 0.002 + hv02337 & 6.863 & 13.247 & 0.016 & 13.212 & 0.015 & 0.034 & 0.005 + hv02338 & 42.200 & 10.580 & 0.007 & 10.640 & 0.007 & -0.057 & 0.002 + hv02339 & 13.880 & 12.139 & 0.004 & 12.143 & 0.004 & -0.005 & 0.001 + hv02352 & 13.632 & 12.282 & 0.003 & 12.267 & 0.003 & 0.014 & 0.001 + hv02369 & 48.377 & 10.297 & 0.006 & 10.355 & 0.006 & -0.058 & 0.003 + hv02405 & 6.924 & 13.309 & 0.011 & 13.283 & 0.012 & 0.025 & 0.005 + hv02432 & 10.918 & 12.446 & 0.015 & 12.427 & 0.016 & 0.018 & 0.005 + hv02447 & 118.350 & 9.440 & 0.006 & 9.438 & 0.005 & 0.003 & 0.002 + hv02463 & 13.965 & 12.142 & 0.004 & 12.134 & 0.004 & 0.006 & 0.001 + hv02527 & 12.950 & 12.452 & 0.005 & 12.456 & 0.005 & -0.004 & 0.002 + hv02538 & 13.869 & 12.246 & 0.003 & 12.266 & 0.003 & -0.020 & 0.002 + hv02549 & 16.220 & 11.885 & 0.004 & 11.891 & 0.004 & -0.006 & 0.001 + hv02579 & 13.428 & 12.121 & 0.003 & 12.114 & 0.004 & 0.006 & 0.001 + hv02580 & 16.918 & 11.879 & 0.006 & 11.910 & 0.005 & -0.031 & 0.001 + hv02733 & 8.722 & 12.887 & 0.011 & 12.855 & 0.010 & 0.029 & 0.005 + hv02749 & 23.105 & 11.625 & 0.004 & 11.682 & 0.004 & -0.058 & 0.002 + hv02793 & 19.220 & 11.734 & 0.006 & 11.767 & 0.005 & -0.033 & 0.002 + hv02827 & 78.802 & 9.744 & 0.004 & 9.824 & 0.003 & -0.081 & 0.002 + hv02836 & 17.528 & 11.988 & 0.005 & 12.008 & 0.005 & -0.019 & 0.002 + hv02854 & 8.635 & 12.810 & 0.016 & 12.785 & 0.013 & 0.024 & 0.005 + hv02883 & 108.970 & 9.953 & 0.010 & 9.958 & 0.009 & -0.006 & 0.002 + hv05497 & 99.100 & 9.345 & 0.004 & 9.376 & 0.003 & -0.030 & 0.002 + hv05655 & 14.211 & 12.308 & 0.005 & 12.325 & 0.005 & -0.015 & 0.002 + hv06065 & 6.838 & 13.301 & 0.011 & 13.282 & 0.013 & 0.014 & 0.005 + hv06098 & 24.237 & 11.224 & 0.003 & 11.218 & 0.003 & 0.006 & 0.001 + hv08036 & 28.369 & 11.329 & 0.006 & 11.374 & 0.006 & -0.044 & 0.002 + hv12452 & 8.736 & 12.837 & 0.016 & 12.832 & 0.015 & 0.003 & 0.005 + hv12471 & 15.863 & 12.178 & 0.004 & 12.208 & 0.004 & -0.028 & 0.002 + hv12505 & 14.389 & 12.327 & 0.005 & 12.338 & 0.005 & -0.013 & 0.001 + hv12656 & 13.400 & 12.269 & 0.002 & 12.248 & 0.002 & 0.020 & 0.001 + hv12700 & 8.153 & 12.975 & 0.011 & 12.953 & 0.012 & 0.022 & 0.005 + hv12717 & 8.843 & 12.877 & 0.015 & 12.842 & 0.015 & 0.034 & 0.005 + hv12724 & 13.744 & 12.419 & 0.004 & 12.444 & 0.005 & -0.025 & 0.002 + hv12815 & 26.135 & 11.212 & 0.006 & 11.257 & 0.006 & -0.042 & 0.002 + hv12816 & 9.108 & 12.890 & 0.011 & 12.869 & 0.011 & 0.019 & 0.002 + hv13048 & 6.853 & 13.165 & 0.011 & 13.152 & 0.013 & 0.010 & 0.005 + p71-u01 & 22.558 & 11.661 & 0.007 & 11.702 & 0.007 & -0.038 & 0.002 + l c c c c 55133.665 & 11.538 & 0.013 & 11.633 & 0.015 + 55134.756 & 11.568 & 0.013 & 11.605 & 0.014 + 55136.253 & 11.510 & 0.013 & 11.498 & 0.013 + 55136.953 & 11.485 & 0.012 & 11.471 & 0.014 + 55138.621 & 11.396 & 0.013 & 11.372 & 0.013 + 55139.561 & 11.378 & 0.012 & 11.330 & 0.013 + 55140.963 & 11.334 & 0.011 & 11.296 & 0.013 + 55141.986 & 11.268 & 0.011 & 11.276 & 0.012 + 55143.049 & 11.244 & 0.012 & 11.255 & 0.012 + 55144.443 & 11.205 & 0.011 & 11.235 & 0.013 + 55146.159 & 11.182 & 0.011 & 11.232 & 0.012 + 55146.916 & 11.162 & 0.011 & 11.235 & 0.012 + 55148.340 & 11.165 & 0.011 & 11.243 & 0.012 + 55149.495 & 11.167 & 0.011 & 11.267 & 0.014 + 55151.036 & 11.171 & 0.011 & 11.290 & 0.012 + 55151.989 & 11.187 & 0.011 & 11.292 & 0.012 + 55153.002 & 11.220 & 0.011 & 11.315 & 0.013 + 55155.321 & 11.266 & 0.011 & 11.394 & 0.013 + 55155.906 & 11.282 & 0.011 & 11.378 & 0.013 + 55156.963 & 11.302 & 0.011 & 11.419 & 0.013 + 55158.231 & 11.357 & 0.012 & 11.460 & 0.013 + 55159.318 & 11.389 & 0.012 & 11.528 & 0.014 + 55160.793 & 11.439 & 0.012 & 11.523 &
0.014 + 55161.870 & 11.472 & 0.012 & 11.584 & 0.014 + | the carnegie hubble program ( chp ) is designed to improve the extragalactic distance scale using data from the post - cryogenic era of spitzer .
the ultimate goal is a determination of the hubble constant to an accuracy of 2% .
this paper is the first in a series on the cepheid population of the large magellanic cloud , and focusses on the period luminosity relations ( leavitt laws ) that will be used , in conjunction with observations of milky way cepheids , to set the slope and zero point of the cepheid distance scale in the mid infrared . to this end , we have obtained uniformly sampled light curves for 85 lmc cepheids , having periods between 6 and 140 days . period
luminosity and period color relations are presented in the 3.6 @xmath0 m and 4.5 @xmath0 m bands .
we demonstrate that the 3.6 @xmath0 m band is a superb distance indicator .
the cyclical variation of the [ 3.6][4.5 ] color has been measured for the first time .
we attribute the amplitude and phase of the color curves to the dissociation and recombination of co molecules in the cepheid s atmosphere .
the co affects only the 4.5 @xmath0 m flux making it a potential metallicity indicator . |
transiting extrasolar planets of bright host stars provide a unique opportunity for physical characterization in ways impossible for non - transiting planets that are also too close to their host stars for direct imaging . currently , the transit and radial velocity methods dominate the growth in planet discoveries both overall and for planets suitable for detailed physical characterization .
this paper addresses the question , `` of these two methods , transits or radial velocities , which is more efficient at discovering planets that transit bright stars ? ''
this is a complex question given the astrophysical effects ( e.g. planet mass - radius relation , stellar mass - radius relation , stellar mass - luminosity relation , radial velocity jitter , stellar activity , etc . ) and technical challenges ( e.g. instrumentation , telescope aperture , available observing time , costs of operation , available funding opportunities , etc . )
that influence the planet yield of a survey . to compare the performance of a radial velocity survey to that of a transit survey
, we adopt a parametric model for each . with some simplifying assumptions
, we quantify the precision of a radial velocity survey that will achieve the same level of search completeness as a transit survey using the same amount of observing time .
for any transit survey , the vast majority of observations of any particular star can not contribute directly to the detection of a transit , simply because they are made when the planet is not transiting . to be successful ,
a transit survey must make a sufficiently large number of independent observations in order that the probability @xmath3 of witnessing @xmath4 or more transit(s ) matches or exceeds a design requirement .
for example , the four - year kepler mission was designed to witness @xmath5 transits of earth analogs with orbital periods @xmath6 year @xcite . for a specific set of observation times of a transit survey , the _ window function _ ,
@xmath7 , describes the visibility of the repetitive transit signal of any given period , i.e. the planet s orbital period , @xmath8 @xcite .
a survey s window function can be computed numerically with high fidelity , either before or after the actual observations , by monte carlo simulations @xcite . however , analytic approximations to the window function are quicker and more convenient to compute than monte carlo simulations @xcite . in addition , an analytic approximation is readily inverted to predict the minimum observing time necessary to exceed a given probability @xmath7 of witnessing @xmath4 or more transits of a given @xmath8 .
we show in the appendix that the hypergeometric distribution provides a high fidelity statistical approximation to the window function and its derivation gives physical insight into the design of a transit survey or interpretation of results from them .
most importantly we show how the transit survey window function is primarily dependent upon the number of independent observing epochs , which enables us to compare the performance of a transit survey to a radial velocity survey for detecting transiting planets . for a radial velocity survey ,
the situation is quite different : nearly every measurement may contribute to the planet detection because the star s radial velocity can be measured at any time and is constantly changing in response to the orbiting planet .
( we write `` nearly '' every measurement because multiple observations made at the same epoch could be redundant if one would suffice at that particular orbital phase . ) to estimate the detection sensitivity of a radial velocity survey , we adopt the prescription outlined by @xcite .
our analysis ends at the phase of a survey when the significance of the detection reaches a suitable level such that follow - up observations to refine the ephemeris and confirm the detection are warranted .
thus , we neglect budgeting for the additional observational effort required to follow - up on a significant signal as the effort required depends very strongly on the nature of the signal and remaining ambiguities of the signal that must be ruled out . in the case of a transit survey a detection from a single to few events
leads to ambiguity in the orbital period @xcite that can be difficult to reconcile with low signal to noise detections .
the numerous sources of astrophysical false positives to the transiting planet signal need to be investigated @xcite along with statistical validation @xcite or mass measurement confirmation @xcite . in the case of a radial velocity detection , in addition to the photometric confirmation of a transit signal @xcite ,
typically additional radial velocity observations are required to refine the ephemeris to a suitable level for scheduling the photometric observations @xcite .
also , radial velocity detections must contend with the astrophysical false positives induced by stellar variability @xcite .
we apply our prescription to address two representative case studies .
first , we compare our models of the performance of a particular transiting - planet finder , _ mearth _ and its southern extension , _
mearth_-south @xcite , to that of a particular radial velocity instrument , _ harps _ @xcite and its northern twin , _
@xcite , for detecting a super - earth transiting at the inner edge of the habitable zone ( hz ) as a function of stellar mass . in the second case
, we study the relative performances of these planet finding techniques for finding earth - sized planets transiting at the inner edge of the hz of m dwarf stars . in each case
, we use the inner edge of the hz simply because planets there will have shorter periods than any other planets in their respective hzs and hence will be more easily detected by either method .
this paper is organized as follows . in section [ sec : transit ] the transit survey window function is introduced , and in the appendix we derive analytic approximations to it .
this leads to simple approximations for the observing time required to reach a specified level of completeness of a transit survey .
the detection sensitivity of a radial velocity survey is inverted in section [ sec : rv ] in order to determine the observing time required for it to reach an appropriate level of completeness to compete with the transit survey .
the implications of this study as applied to the detection of a transiting super - earth and an earth analog at the inner edge of the hz in section [ sec : disc ] and its conclusions are given in section [ sec : conclusion ] .
we can approach the statistical modeling of a transit survey window function in two ways .
we can ask , _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \a ) each time the planet transits , were we observing ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ or _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \b ) each time we observe , is the planet transiting ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for our statistical model , we want _ independent _ trials , so to form a trial , we group observations over an appropriate time interval .
previous work ( e.g. eq . a2 of * ? ? ?
* ) adopted approach a , where each independent planetary _ orbital period _
interval is a binomial trial with probability of our witnessing it equal to the average fraction of time spent observing @xmath9 . in approach b ( see appendix ) , the appropriate independent time interval is the _ transit duration _ , @xmath10 } , \label{eq : tdur}\ ] ] for arbitrary stellar mass , @xmath11 , radius , @xmath12 , and orbital period , @xmath8 .
we have assumed the average relative chord length , @xmath13 and eccentricity @xmath14 @xcite .
effects of non - zero eccentricity on transit survey yield are addressed by @xcite and @xcite .
approach b has the advantage of being readily related to individual observations parceled into lengths of the transit duration .
the window function is determined by the number of unique observing epochs , @xmath15 , of duration @xmath16 .
a set of observation exposures at a cadence shorter than the transit duration will improve the signal to noise ratio of the detection , but count as a single parcel increment to @xmath15 for the purposes of the window function .
thus , @xmath15 is not necessarily the number of exposures ( see section [ sec : hyperg ] ) .
regardless of approach ( a or b ) , the physical situation is the same : in the time span , @xmath17 , between the first and last observations of a particular star with a transiting planet of orbital period , @xmath8 , the expected number of transits is@xmath17/p@xmath18 or @xmath19@xmath17/p@xmath20 transits where @xmath21 is the truncate to nearest integer operation . for uniform and random relative phase of the planetary transits with respect to the set of observation epochs ,
the expectation value of @xmath22@xmath17/@xmath8 ] @xmath24 from the probability , @xmath25 , of witnessing exactly @xmath4 transits ( derived in the appendix ) , the probability , @xmath26 , of witnessing @xmath4 or more transits is given by @xmath27 a transit survey may require detection of @xmath28 transits in order to know the period and hence predict future transits , or it may require @xmath29 transits in order to validate the detection , i.e. confirm that the times separating the transits are all consistent with a particular orbital period . in the appendix we have isolated the mathematics of approximating the window function analytically .
figure [ fig : beattycomp ] shows the window function for a hypothetical transit survey that operates in a mode dedicated to a single field for 30 consecutive days with observations for 8 hours each night .
the specifications for the hypothetical transit survey are the same as presented in figure 1 of @xcite to demonstrate their analytical representation of a transit survey and is typical for the first generation of ground - based hot jupiter transit surveys ( _ hat _
@xcite ; _ kelt _ @xcite ; _ ogle _ @xcite ; _ qatar _ @xcite ; _ tres _
@xcite ; _ wasp _
@xcite ; _ xo _ @xcite ; and others ) .
the window function presented in this study and others ( e.g. * ? ? ? * ; * ? ? ?
* ) provide convenient approximations to the average performance of a transit survey
. these approximations may be useful for design purposes and for real time updates to a transit survey in progress .
also , this study articulates that transit survey completeness , in terms of the window function , derives primarily from the number of distinct observing epochs , @xmath15 , and only secondarily from the time span @xmath17 ( see figure [ fig : alternatobs ] and discussion in section [ sec : numericalcomp])@xmath30 the number of exposures ( see section [ sec : hyperg ] ) ] .
that is , observations can be spread over an observing season or years rather than concentrated in a single observing run .
a consequence is that a longitudinally - distributed network ( e.g. hat - net ) is not necessarily more observationally efficient at finding planets per unit of observing time than a single site ( e.g. wasp ) , all other factors being equal .
however , a longitudinally - distributed network of n telescopes can retire a particular field of view of targets in a smaller number of days than could be accomplished by the same n telescopes located at a single site at the same latitude as the telescopes of the network .
also , continuous observations avoid the ambiguities ( e.g. of orbital period ) inherent in interpreting observations with many gaps where transits can hide .
figure [ fig : winfunc ] shows the window function for a transit survey that operates in a mode cycling between targets such as employed by the _
mearth _ project @xcite .
_ mearth _ surveys the brightest m dwarfs by cycling between targets to obtain several observation per target per night .
the observation span is an observing season for each target .
we simulate this mode of operation through specifying a time series with @xmath15@xmath31 observations that is expected for a target from a single observatory over the course of an observing season including losses due to weather . as a proxy for an observing time
series we begin with a continuous pwv time series downloaded from the noaa earth system research laboratory ground - based gps meteorology program @xcite for @xmath329 months starting september , 1 2008 with a 30 minute cadence from flagstaff , az .
we use the celestial coordinates ra 7h dec + 30 , optimized for the middle of the observing season and observatory latitude , to find the airmass for each pwv observations .
only values when the airmass @xmath33 1.8 , pwv @xmath33 10 mm ( to crudely simulate when the weather conditions would be too poor for observations ) , and sun altitude @xmath33 -5 deg are used to simulate times when observations could be made .
the window function assumes stellar parameters appropriate for a mid m dwarf star ( @xmath34=0.2 @xmath35 @xmath36@xmath34@xmath37 $ ] ; @xmath38=0.26 @xmath36@xmath39@xmath37 $ ] ) .
in addition to the window functions assuming at least @xmath40 1 , 2 , 3 , and 5 transit events are observed ( solid lines from top to bottom , respectively ) , figure [ fig : winfunc ] also shows the window function ( solid orange line ) which corresponds to the triggering mode of operation adopted by the _ mearth _
project @xcite .
real - time analysis of the data identifies a potential flux decrement for a target , and the observing scheduler interrupts cycling through targets and enters a high cadence follow - up and stare mode in order to get a high sensitivity detection of the transit egress .
this has the potential of detecting a planet with a single transit and relaxed photometric precision , which extends the detectability to longer orbital periods and enables sampling more targets .
however , in practice for m dwarfs , the transit duration @xmath41 hr and the transit must be observed during the first portion of the transit event to allow sufficient time to commence the high cadence staring mode .
thus , @xmath16 must be reduced to account for the triggering mode delay . throughout , we reduce @xmath16 by @xmath42=30 min when simulating the triggering mode . for the mid m dwarf example shown in figure [ fig : winfunc ] , the triggering mode window function is superior to the @xmath43 window function for @xmath44 day .
in this section we summarize our metrics for the performance of a radial velocity survey .
we follow the procedure by @xcite to determine the limiting semi - amplitude threshold , @xmath45 , of a radial velocity survey for a given @xmath15 observations , each of which has radial velocity precision , @xmath46 .
we use the performance of the lomb - scargle periodogram for detecting planetary orbits . for a given completeness probability , @xmath47 ,
we invert equation 24 of @xcite to determine the argument of the error function , @xmath48 the error function argument , @xmath49 , relates the expectation of the periodogram in the presence of a signal , @xmath50 where @xmath51@xmath15@xmath52 is the degrees of freedom , to the detection threshold of the experiment , @xmath53 where @xmath54 is the number of independent frequencies searched and @xmath55 is the false alarm probability of the experiment .
the relationship between equations [ eq : gamma ] , [ eq : exps ] , and [ eq : detthresh ] is given by @xmath56 equations [ eq : gamma]-[eq : rvall ] are taken directly from @xcite .
we numerically solve equation [ eq : rvall ] to derive the limiting @xmath57 that a radial velocity achieves for a given @xmath47 and @xmath15 .
we adopt parameters following @xcite that the radial velocity survey has a @xmath58 and that @xmath59 @xmath15 @xcite .
an additional assumption for the validity of these results is that the baseline of observations is longer than the orbital period of the planet . in this study
we ignore the impact of uncorrected stellar activity for setting a noise floor on radial velocity precision @xcite , and the additional observations necessary to refine the orbital parameters with sufficient precision to make an accurate prediction for the epoch of transit @xcite .
when analyzing the identification of transiting planets from a radial velocity detected sample of planets in this paper , we assume a purely geometric prior transit probability rather than the potentially 10 - 25% higher posterior transit probability for planets first detected with radial velocities @xcite .
with the results from the previous sections , we can evaluate a transit survey designed to detect a transiting planet and quantify the equivalent radial velocity survey that is expected to have the same detection capability for the same total amount of observing time . with the framework described in this paper , we assume both the radial velocity survey and the transit survey achieve @xmath15 observations .
we then determine which method using @xmath15 ( independent of how long it takes to achieve @xmath15 observations ) achieves a higher survey completeness level for the planet of interest .
one opportunity provided by _
jwst _ will be an unprecedented capability for measuring the atmospheric constituents of potentially habitable planets @xcite .
taking advantage of the _ jwst _
opportunity requires identifying the most viable transiting planets for follow up by pushing the limits of photometric and radial velocity precision .
we first compare the performances of the _ mearth _ transiting survey @xcite and the _ harps _
m dwarf radial velocity survey @xcite for detection of super - earth planets orbiting the brightest m dwarfs .
then we show general results for the design of future surveys pushing the detection limits down to earth - sized planets .
the steps for the comparison begin with determining the @xmath15 necessary for a transit survey to reach a given percent completeness for an inner hz planet .
we then calculate the radial velocity precision required to detect the same planet with the same amount of observing time . as an example we calculate the performance for a survey modeled after the _ mearth _
transit survey @xcite .
the _ mearth _ survey is focused on identifying transiting super - earths orbiting the brightest m dwarfs employing the triggering mode of detection along with ex post facto period - folding search for multiple transit events . since _ mearth
_ is observing the brightest m dwarfs , which are far apart on the sky , it typically can observe only a single target per telescope pointing . in order to specify @xmath15 required for a transit survey
, we outline the properties of the planets and their stellar hosts in table [ tab : results ] . for a given stellar mass , @xmath35 , the stellar radius , @xmath38 , stellar surface gravity @xmath60 , and stellar effective temperature , @xmath61 ,
are estimated by a spline fit to the calibration of mk spectral types table 15.7 and 15.8 of @xcite .
for the given stellar parameters , we determine the orbital period corresponding to the inner edge of the hz , @xmath62 , from the results of @xcite . empirically for m dwarfs ,
we find that an equilibrium temperature @xmath63 equal to 273 k occurs at the inner edge of the @xcite hz for a planetary bond albedo , @xmath64 , and full redistribution of the stellar flux , @xmath65 .
although a warmer @xmath66 k is a better match to the inner hz edge of @xcite for higher mass stars , @xmath35@xmath67 @xmath68 , we use the cooler @xmath69 k to define the inner hz for all stellar types throughout this work . with the stellar parameters and
@xmath62 given , the additional quantities follow ( @xmath70 , transit probability , @xmath71 , radial velocity semi - amplitude , @xmath45 , and transit depth , @xmath72 ) for an earth - mass and earth - radius analog as well as a super - earth mass and radius analog to gj 1214 @xcite .
table [ tab : results ] shows @xmath73 necessary for a transit survey to achieve a completeness probability @xmath74 of witnessing a transit at least once in the triggering mode of operation at @xmath62 .
we assume @xmath17=200 day or @xmath17=@xmath75 whichever is greater .
focusing on the 0.2 @xmath11 star , @xmath76 are required with sufficient precision to detect the corresponding transit depth , @xmath77 ppm .
if every star has a planet orbiting with a period of approximately , @xmath62 , then with @xmath78 , @xmath79 such targets need to be observed .
at least @xmath80 observations are required to expect to find one transiting planet at the inner hz orbiting a 0.2 @xmath11 m dwarf .
cycling between targets with a cadence of 5 min and 8 hours of observations , reaching 34000 observations requires 355 clear nights .
the above results are first - order guidance assuming unit planet occurrence .
they can be conveniently scaled to the measured occurrence rates of planets orbiting m - dwarfs @xcite .
fortunately , the targets are independent , so this amount of observing time can be achieved with multiple telescopes working in parallel ( _ mearth _ employs eight telescopes and is commissioning eight more in the southern hemisphere ) or observing fainter stars with multiple targets per field .
of course , if the fraction of stars with such a planet is less than unity , the required observing time will be more .
it is interesting to note that @xmath73 is not monotonically increasing with @xmath11 because the increasing hz period is offset by the increase in @xmath70 with increasing @xmath11 .
in addition , the trigger mode delay becomes an appreciable fraction of @xmath70 for the lowest mass stars . for @xmath81 , @xmath82 ,
thus in principle similar number of observations are needed to find a hz planet orbiting a late k as late m star in the triggering mode of operation . however , the minimum number of clear nights necessary to achieve the requisite number of observations @xmath83 , assuming @xmath84 hr observing each night is a monotonically increasing function of @xmath11 .
in addition , the depth of the transit for fixed planet size is @xmath856 times smaller for late k hosts than late m hosts placing more stringent requirements on the photometric precision for observations of late k hosts . for an individual target
we determine the radial velocity precision necessary to reach the same level of survey completeness for making a super - earth @xmath86 detection at @xmath62 given the same amount of observing epochs as the transit survey .
in this case the goal of the radial velocity survey is to make a sufficient number of @xmath86 detections in order to identify at least one that also transits .
for the example at 0.2 @xmath11 the circular orbit semi - amplitude of a super - earth at @xmath62 is @xmath875.2 m s@xmath1 . to complete the analysis
, we must relate @xmath73 of a transit survey to an equivalent @xmath88 for a radial velocity survey .
the _ mearth _ survey cycles between targets with a 5 min cadence on average and employs eight independent telescopes , and the _ harps _
m dwarf survey had a fixed exposure time of 15 min .
thus , for the super - earth survey comparison , we assume @xmath89 , where @xmath90 , implying 24 observing epochs over independent targets are obtained in the transit survey for each radial velocity measurement . in the case of a transit survey with multiple targets per field , the conversion factor can be significantly larger ( @xmath91 is @xmath924 orders of magnitude larger for a survey like _ kepler _ ; * ? ? ?
using the same completeness probability @xmath74 for the radial velocity survey with a false alarm probability of @xmath58 , @xmath93 , and @xmath94 , we use equation [ eq : rvall ] to estimate the limiting semi - amplitude precision ratio , @xmath95 . hence , for the 0.2 @xmath11 case , @xmath96 m s@xmath1 . to summarize
, a radial velocity survey will have reached the same survey completeness probability for a super - earth orbiting at the inner hz for a late m dwarf as a simulated trigger - mode transit survey if it obtains @xmath88 with a precision of @xmath96 m s@xmath1 .
if the radial velocity precision is @xmath97 m s@xmath1 , then the radial velocity survey will not have reached the same level of survey completeness as the transit survey and is not as efficient for surveying @xmath86 detections that also transit their host . for comparison , the mode of the radial velocity precision from the _ harps _
m dwarf survey sample is approximately equal to @xmath98 m s@xmath1@xcite . in the above example of a super - earth at the inner hz of @xmath99 @xmath34 host star
, the inequality , @xmath100 , implies that formally the efficiency for transiting planet detection for _ mearth _
is slightly greater than that for transiting planet detections among the _ harps _
m dwarf survey @xmath86 detections .
the transiting planet yield not only depends upon relative efficiency , but also total observing time . the advantage for _ mearth _
may even be larger given that the _ mearth _ telescopes are dedicated to the survey whereas _ harps _
is a shared instrument .
the above results depend on the definition of the inner - hz @xcite .
to first order , the results of section [ sec : onetransit ] show that for a transit survey for fixed detection completeness @xmath15@xmath101 . the expected scaling with @xmath8 for a radial velocity survey maintaining the
same detection completeness @xmath102 ( see equation 26 of @xcite ) is the same .
numerical calculations confirm this expected weak dependence on @xmath8 .
for the 0.2 @xmath11 case , @xmath46=1.9 m s@xmath1 and 1.7 m s@xmath1 for a cooler , @xmath103 k , and hotter , @xmath104 k inner hz definition , respectively .
the discovery of gj 1214 did not rely on the triggering mode for detection , but was detected after phasing two potential transit events @xcite .
figure [ fig : rvprecse ] summarizes the radial velocity precision required in order to reach the same level of completeness for a single target using the same amount of observing time for the transit survey design of requiring @xmath40 1 , 2 , 3 , and 5 events as the solid lines from bottom to top , respectively .
the dash line shows results for the triggering mode of operation .
the achieved precision of the _ harps _ m - dwarf survey @xmath98 is also shown ( horizontal dotted line ) demarcates the stellar hosts where a 8-telescope system like _ mearth _ is more ( below ) and less ( above ) efficient . in the triggering discovery mode ( dashed line )
, @[email protected]@xmath34 is the stellar host regime where the 8-telescope _ mearth _ system in the triggering discovery mode is more efficient than _ harps _ for detecting transiting super - earth planets .
for a 3-transit requirement detection mode , the detection efficiency equivalency is at @[email protected]@xmath34 .
the results shown in figure [ fig : rvprecse ] and the previous discussion assume the goal of the survey is to find a hz planet that actually transits . given the small probability to transit , the radial velocity survey will achieve a large number of detections @xmath105 for every single hz planet that transits .
if the goal of the radial velocity survey is just to detect a single @xmath86 super - earth mass object , then it is more appropriate to compare to the time necessary for a transit survey to result in a single detection , @xmath106 . the requisite radial velocity precision for a single radial velocity detection of a super - earth with the equivalent observing time for a single transiting planet detection using the 8-telescope _ mearth _ system is shown in figure [ fig : rvpreconedetse ] .
for measuring the occurrence rate of super - earth @xmath86 planets @xcite , the _ harps _ m dwarf survey is predicted to have superior efficiency for measuring super - earth population statistics compared to the _ mearth _ survey . however , a transit survey observing fainter targets can achieve a much larger conversion factor , @xmath91 . a transit survey requiring @xmath107 transits for detection with a multi - plexing conversion factor of @xmath108
has equivalent detection efficiency to a radial velocity survey with @xmath109 m s@xmath1 for inner hz super - earth planets around late m dwarfs .
one caveat to the discussion is the need for @xmath110 @xmath86 detections at the inner hz for late m dwarfs before one expects to find one that also transits .
both the _ harps _ and _ mearth _ surveys target the brightest m dwarfs available , but due to the geometric transit alignment one must survey a larger volume and fainter magnitudes for the @xmath86 detections that also transit . for uniform stellar density , the brightest transiting planet at the inner hz for late m dwarfs will typically be @xmath111 mag fainter than the brightest @xmath86 nontransiting detection . at the median brightness of the _ mearth _
m dwarf sample ( v=15.2 , j. irwin , _ private communication _ ) the _ harps _ instrumental radial velocity precision is predicted to be @xmath112 m s@xmath1@xcite . thus at the median brightness of the _ mearth _ m dwarf sample , _ mearth _
is much more efficient at finding transiting super - earth planets at the inner hz than if attempting to survey the _ mearth _
m dwarf sample for transiting planets with _ harps_. we repeat the analysis of the previous section down to the regime of truly earth - size and earth - mass planets at the inner hz of the brightest m dwarfs .
the _ tess _
mission @xcite is one possibility for identifying transiting hz earth analogs orbiting bright stars , but here we briefly explore the possibility of a scaled - up _ mearth _ survey and the associated equivalent radial velocity precision . from the perspective of window functions ,
the required @xmath113 is the same as for the detection of super - earths at the inner hz .
however , the photometric precision must be improved by a factor @xmath857 from the super - earth survey to detect the shallower earth - analog transit depth . in principle , the requisite photometric precision for enabling the trigger mode of detection for earth - size planets orbiting m dwarfs has been demonstrated from the ground .
scintillation limited performance using 4 m class telescopes has achieved @xmath114=250 ppm min@xmath1 photometric noise rate on v@xmath8512 for fields with a comparable brightness and spectral type companion in the optical @xcite . @xcite
achieve 1250 ppm min@xmath1 photometric noise rate on v@xmath8513
late k and early m dwarfs with a 2 m class telescope with an innovative target cycling approach .
thus , the practical limitation of a scaled up _ mearth _
survey is available observing time on larger telescopes rather than a physical limit imposed by the earth s atmosphere .
carrying through the above calculation for an earth analog at the inner hz , we find @xmath115 m s@xmath1 is needed for comparable efficiency to the triggering mode transit survey .
for this result we assume @xmath116 , which would be expected given the efficiency difference between a high resolution spectrograph for precision radial velocity measurements compared to single - target photometry . since radial velocity instrumentation is typically on larger aperture telescopes , assuming more efficient radial velocity observations , @xmath117 , results in @xmath118 ms@xmath1 .
the noise floor for m dwarf precision radial velocities was @xmath119 m s@xmath1 for the _ harps _
m dwarf survey @xcite .
results for additional transit survey designs of 2 , 3 , or 5 transits for detection are shown in figure [ fig : rvprec ] , and the precision needed to reach comparable efficiency of an @xmath86 detection is shown in figure [ fig : rvpreconedet ] .
pushing the ground - based photometric and radial velocity techniques provides an opportunity for detecting true earth - sized planets at the inner hz orbiting the brightest late m dwarfs .
for example a scaled up _ mearth _
survey could detect hz earth - sized transiting planets among the brightest late m dwarfs .
we predict that for the same amount of observing time , a radial velocity survey requires precision @xmath120 m s@xmath1 in order to more efficiently detect a @xmath86=1 @xmath121 transiting planet at the inner hz for mid m dwarf . from space ,
the _ tess _ mission @xcite plans to survey early m and k dwarfs for such earth analogs .
there is great impetus for finding the rare earth analogs that also transit among the brightest stars to take full advantage of the follow - up potential of _ hst _ and the upcoming opportunities provided by _
jwst_. _ jwst _ will usher in an unprecedented capability for measuring the atmospheric constituents of potentially habitable planets @xcite .
we demonstrate that the window function for a transit survey can be expressed in terms of the hypergeometric probability distribution .
an analytic approximation to the window function simplifies the design of a transit survey and provides an accurate estimate to the observing time necessary to achieve a given level of completeness for detection of a transiting planet .
furthermore , to first order , the resulting window function for a transit survey depends only on the total number of independent observations and secondarily depends upon the distribution of observation times or baseline over which the observations are obtained .
we employ these results regarding the transit survey window function model in order to determine the comparable radial velocity survey precision that achieves the same completeness level given the same number of observations . for the brightest m dwarfs , @xmath122
, we predict _ harps _ and the _ mearth _ surveys have approximately the same efficiency for detecting super - earth planets that transit and orbit at the inner hz . however , at the median brightness of the _ mearth
_ sample , @xmath123 , the _ mearth _ survey is predicted to have much higher efficiency at detecting transiting super - earths than _ harps _ due to the degraded radial velocity precision for fainter stars . the higher efficiency is predominately driven by the small geometric transit probability at the inner hz for late m dwarfs .
_ harps _ is predicted to have much higher efficiency in @xmath86 detections than _ mearth
_ has for transiting planets , thus making _ harps _
efficient at determining super - earth @xmath86 population statistics .
however , complementary planet population studies are possible from wide field transit surveys that observe many targets per observing epoch ( @xmath124 in the case of _ kepler _ ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
prior work ( e.g. eq . a2 of * ? ? ?
* ) adopted approach a of section [ sec : transit ] , `` each time the planet transits , were we observing ? '' for this statistical model of the window function each orbital period time interval represents an independent trial .
specifically , the orbital period time interval represents a binomial trial with probability of our witnessing it equal to the average fraction of time spent observing @xmath9 .
thus , the probability of witnessing exactly @xmath4 transits is @xmath125 recalling that the expectation value for the number of trials @xmath126@xmath17@xmath127 ( eq . [ eq : m ] ) and that the probability of witnessing @xmath4 or more transits @xmath26 is given by eq .
[ eq : probn ] .
approach a also can be addressed , in appropriate limits , specifically large m and small @xmath9 ( e.g. appendix 4 of lyons 1996 ) , using the poisson approximation to the binomial distribution , @xmath128 where @xmath129 . even though the poisson distribution is defined for all @xmath130 ,
the maximum number of transits in time span @xmath17 is m = @xmath21@xmath17/@xmath131 .
the physical interpretation of the poisson process would be that transits are observed at random times at an average rate of @xmath132 transits in @xmath17 , which approaches @xmath133 in the appropriate limit of large @xmath134 .
of course , transits do not occur at random times , but the approximation works if observations are sparsely and randomly distributed throughout the time span @xmath17 . @xcite
recommend that the poisson approximation gives good results so long as @xmath135 and @xmath134 is large . while it is the least accurate ,
the poisson approximation is the simplest to compute ; we recommend it for hand calculations ( see section [ sec : onetransit ] ) and to validate more complicated window function implementations . in approach b , `` each time we observe , is the planet transiting ? ''
, the statistical model of the window function adopts the transit duration time interval as the relevant independent trial .
we make an analogy between searching for a transiting planet and randomly drawing colored balls from an urn without replacement ( figure [ fig : cartoon ] ) .
each ball represents a time interval equal to the transit duration @xmath16 and also a potential observation of a transit .
the parameter @xmath15 in the hypergeometric distribution is the number of unique observation intervals of length @xmath16 that have sufficient sensitivity to detect a transit if it had occurred during the interval . parceling the observing span @xmath17 into intervals of @xmath16 results in a set of @xmath136@xmath17/@xmath16 time intervals , or by analogy @xmath137 balls in the urn . in the illustration ,
the orbital period @xmath8 is six times the transit duration @xmath16 . in general , for circular orbits and central transits , the ratio @xmath8/@xmath16@xmath138 , where @xmath139 is the orbital semimajor axis and the formula neglects the planetary radius .
the illustration shows a contiguous set of @xmath136@xmath17/@xmath16@xmath140 balls , where @xmath141 balls are red ( in transit ) , and the @xmath142 remaining balls are colored white ( out - of - transit ) .
there are @xmath15=5 balls drawn with solid outlines ( observations ) and @xmath143@xmath15@xmath144 balls with dashed outlines ( observation gaps or observation intervals with insufficient sensitivity to detect a transit even if it had occurred in the particular interval ) .
thus the the probability of having @xmath4 observations occur during transit among the @xmath15 observations is given by the hypergeometric distribution , @xmath145 which is appropriate for this case of sampling without replacement .
in the nomenclature of statistics , @xmath146 a limitation common to the binomial and hypergeometric distributions is that their input parameters formally are required to be integers , because of the combinatorics operator `` choose . '' for example , each equation contains @xmath134 choose @xmath4 ; that @xmath4 is an integer is not a problem , but if @xmath134 is forced to also , then the window function can have a stair - step appearance .
this is readily mitigated by using implementations of the operators that can handle real - valued inputs via the gamma function , e.g. @xmath147 one may wish to use stirling s formula or a similar approximation for the factorials , or instead of the gamma function itself , one may wish to use a log(gamma ) function if available , to avoid numerical overflow .
it is possible to show the numerical similarities of the binomial model of @xcite with the hypergeometric model of this study by expressing @xmath148 , then the binomial model can be written as @xmath149 .
the expectation of the binomial model of @xcite , @xmath150=f_{\rm cov}m = f_{\rm cov}t_{\rm obs}/p = n\tau_{\rm dur}/p$ ] , is the same as the expectation of the hypergeometric model of this study , @xmath151=nm /
n = n\tau_{\rm dur}/p$ ] . the approaches higher moments are not identical ; for example , the variance ratio between the models , @xmath152/var[h(n , m , n)]=(1-\tau_{\rm dur}/t_{\rm obs})/(1-\tau_{\rm dur}/p)$ ] .
however , since @xmath153 , for parameters generally applicable to a transit survey , the variances are nearly equal .
the probability of witnessing at least one transit ( from equation [ eq : probn ] ) is @xmath154 associating @xmath155 , equation [ eq : poione ] is readily inverted : @xmath15@xmath156 .
this provides a convenient benchmark for planning purposes as well checking how probable a potential planetary discovery is in a transit search . in figure
[ fig : beattycomp ] we compare the resulting window function given by equation [ eq : probn ] to the exact numerical window function for a hypothetical ground - based transit survey that observes for 30 consecutive days with observations for 8 hours each night . the specified transit survey is the same as presented in figure 1 of @xcite to demonstrate their analytical representation of a transit survey ( solid blue line ) .
the exact numerical window function ( dash black line ) shows the substantial alias structure that is typical of ground - based transit surveys due to the day - night observing cycle , and which is not modeled by the simpler analytic approximations .
the analytic expressions approximate well the average window function .
to demonstrate this we show in figure [ fig : beattycomp ] a moving average of the exact numerical function using a 1 day window ( solid black line ) .
the left and right panels provide results for requiring @xmath28 and @xmath29 transit events for detection , respectively .
we show the hypergeometric window function model from this study ( equation [ eq : hyperg ] ) as the solid red line .
solar values for the stellar parameters are assumed .
the poisson window function model ( equation [ eq : poisson ] , dash red line ) also does a reasonable job in this scenario with @xmath158 .
figure [ fig : winfunc ] shows the window function for a transit survey that operates in a low duty cycle mode obtaining several observations per night over the course of an observing season .
we simulate this mode of operation through specifying a time series with @xmath15@xmath31 that is expected for a target from a single observatory over the course of an observing season including losses due to weather and stellar parameters appropriate for a mid m dwarf star .
the exact numerical window functions assuming at least @xmath40 1 , 2 , 3 , and 5 transit events are observed ( solid lines from top to bottom , respectively ) .
the hypergeometric model ( solid red line ) and its poisson model approximation ( dash blue line ) agree extremely well as expected for a low duty cycle transit survey .
figure [ fig : keplerwin ] shows the window function for two typical targets observed by _
using quarters 1 through 16 ( @xmath853.6 yr ) of nearly continuous coverage @xcite . the rightmost set of black solid , red solid , and blue solid lines correspond to the coverage of _
kepler_-11 , host to six transiting planets @xcite .
this target was observed for all 16 _ kepler _ quarters . for the time series
, we employ barycentric julian date mid - cadence times stamps that have valid data as determined by the _ kepler _
pipeline @xcite .
the hypergeometric model ( solid red line ) requiring @xmath403 transit events being observed agrees well with the exact numerical calculation ( solid black line ) also requiring @xmath107 . as expected the analytical window model function for continuous observations with no gaps of ( table 2 of
* ) , overestimates the window function due to the reality of gaps in _ kepler _ data .
the leftmost set of solid black and solid red lines correspond to the coverage of the host to _ tres_-2b ( a.k.a .
kepler-1b ; * ? ? ?
* ; * ? ? ?
this target is impacted by the loss of a detector in quarter four @xcite , causing missing data every fourth _ kepler _ quarter .
the detector loss impacts @xmath8519% of _ kepler _ targets having reduced amount of data available .
the dash lines are the same as above , but they require @xmath159 transits for detection .
the hypergeometric model does not fully capture the details present in the exact numerical window functions .
the assumptions of the hypergeometric model begin to break down in observing scenarios consisting of high duty cycle blocks interspersed with long periods of no data .
in addition to the sharp , isolated aliases at the diurnal cycle and its harmonics prevalent in ground - based transit survey window functions ( which we readily average over ) , there are `` broad - band '' alias structures present across all orbital periods that can result in the exact numerical window function to deviate from the hypergeometric model . in figure [
fig : alternatobs ] we demonstrate the principal dependence of a transit survey window function on @xmath73 and the weak dependence on @xmath17 .
we show the hypergeometric model ( solid lines ) and the moving average of the exact numerical window function ( dash lines ) for three scenarios that have the same @xmath73 but vary in @xmath17 .
the left and right panels vary by requiring @xmath40 2 and 3 transit events for detection , respectively .
the baseline transit survey consists of 7 hours of observations each day for 30 consecutive days ( black lines ) .
we simulate a longitudinally - distributed network survey by considering a transit survey with half @xmath17 ( 15 consecutive days ) of the baseline case while doubling the observing time each day to conserve @xmath73 ( red lines ) .
the window function does show an increase in detection probability at intermediate periods at the cost of removing the low detection probability tail toward longer periods of the baseline scenario .
the third @xmath73-conserving scenario consists of splitting the 30 consecutive days of observations from the baseline case into two 15 consecutive day groups separated by a year ( blue lines ) giving a @xmath17=380 day .
in contrast to the @xmath73-conserving cases , we show a transit survey window function resulting from doubling @xmath73 while keeping @xmath17 the same as the baseline case ( magenta lines ) .
these scenarios demonstrate that the detection probability of a transit survey are very sensitive to @xmath73 .
@c@c@c@c@c@c@c@c@c@c@c@c@c@c@c@ @xmath35 & @xmath38 & @xmath61 & @xmath62 & @xmath160 & @xmath161 & @xmath162 & @xmath70 & @xmath15 & @xmath163 & @xmath15@xmath164 & @xmath165 & @xmath166 & @xmath167 & @xmath168 + @xmath36@xmath34@xmath37 $ ] & @xmath36@xmath39@xmath37 $ ] & @xmath36k@xmath37 $ ] & @xmath36day@xmath37 $ ] & & & @xmath36ppm@xmath37 $ ] & @xmath36ppm@xmath37 $ ] & @xmath36hr@xmath37 $ ] & & & @xmath36 m s@xmath1@xmath37 $ ] & @xmath36 m s@xmath1@xmath37 $ ] & @xmath36 m s@xmath1@xmath37 $ ] & @xmath36 m s@xmath1@xmath37 $ ] + 0.15 & 0.20 & 3077 & 10.47 & 0.0188 & 2167 & 15801 & 1.18 & 492 & 73 & 26139 & 6.83 & 2.69 & 1.035 & 0.81 + 0.17 & 0.22 & 3109 & 11.62 & 0.0184 & 1724 & 12568 & 1.29 & 472 & 76 & 25606 & 6.07 & 2.29 & 0.920 & 0.70 + & * 0.26 * & * 3155 * & * 13.61 * & * 0.0179 * & * 1266 * & * 9233 * & * 1.46 * & * 449 * & * 82 * & * 25086 * & * 5.17 * & * 1.85 * & * 0.783 * & * 0.58 * + 0.22 & 0.28 & 3185 & 15.13 & 0.0176 & 1052 & 7671 & 1.59 & 436 & 87 & 24825 & 4.68 & 1.62 & 0.709 & 0.51 + 0.25 & 0.32 & 3230 & 17.75 & 0.0171 & 818 & 5967 & 1.82 & 421 & 96 & 24655 & 4.08 & 1.35 & 0.618 & 0.44 + 0.27 & 0.35 & 3261 & 19.75 & 0.0168 & 703 & 5127 & 1.99 & 413 & 102 & 24653 & 3.74 & 1.21 & 0.566 & 0.40 + 0.30 & 0.38 & 3310 & 23.14 & 0.0163 & 572 & 4171 & 2.26 & 404 & 114 & 24849 & 3.30 & 1.04 & 0.501 & 0.34 + 0.35 & 0.44 & 3404 & 29.86 & 0.0154 & 426 & 3110 & 2.75 & 404 & 139 & 26278 & 2.74 & 0.86 & 0.415 & 0.28 + 0.40 & 0.50 & 3520 & 37.76 & 0.0144 & 336 & 2449 & 3.26 & 418 & 170 & 29071 & 2.32 & 0.76 & 0.351 & 0.25 + 0.45 & 0.55 & 3661 & 46.32 & 0.0133 & 278 & 2025 & 3.69 & 442 & 204 & 33257 & 2.00 & 0.70 & 0.303 & 0.22 + 0.50 & 0.59 & 3811 & 55.12 & 0.0123 & 239 & 1745 & 4.06 & 473 & 240 & 38559 & 1.76 & 0.67 & 0.267 & 0.20 + 0.55 & 0.63 & 3952 & 64.07 & 0.0114 & 213 & 1555 & 4.38 & 503 & 276 & 44098 & 1.57 & 0.63 & 0.238 & 0.19 + 0.60 & 0.66 & 4105 & 74.87 & 0.0106 & 192 & 1402 & 4.75 & 538 & 319 & 50889 & 1.41 & 0.61 & 0.213 & 0.18 + 0.70 & 0.75 & 4591 & 117.39 & 0.0085 & 148 & 1078 & 5.95 & 657 & 489 & 77718 & 1.09 & 0.56 & 0.166 & 0.15 + 0.80 & 0.86 & 5198 & 189.37 & 0.0066 & 114 & 834 & 7.49 & 827 & 775 & 125410 & 0.85 & 0.53 & 0.129 & 0.14 + 0.90 & 0.90 & 5509 & 231.34 & 0.0059 & 103 & 749 & 8.15 & 924 & 941 & 157421 & 0.73 & 0.49 & 0.112 & 0.13 + 1.00 & 1.03 & 5794 & 307.04 & 0.0053 & 80 & 581 & 9.78 & 1011 & 1235 & 190520 & 0.63 & 0.44 & 0.095 & 0.11 + 1.10 & 1.16 & 6063 & 404.58 &
0.0048 & 63 & 459 & 11.76 & 1097 & 1613 & 226371 & 0.54 & 0.40 & 0.081 & 0.10 + [ tab : results ] burke , c. j. 2010 , in towards other earths : perspectives and limitations in the elt era , ed .
n. c. santos , c. melo , l. pasquini , a. glindemann ( porto : centro de astrofisica da universidade do porto ) , session 1 , http://www.astro.up.pt/investigacao/conferencias/toe2009/dvd/toe/index.html | transit and radial velocity searches are two techniques for identifying nearby extrasolar planets to earth that transit bright stars .
identifying a robust sample of these exoplanets around bright stars for detailed atmospheric characterization is a major observational undertaking . in this study
we describe a framework that answers the question of whether a transit or radial velocity survey is more efficient at finding transiting exoplanets given the same amount of observing time . within the framework we show that a transit survey s window function can be approximated using the hypergeometric probability distribution .
we estimate the observing time required for a transit survey to find a transiting earth - sized exoplanet in the hz with an emphasis on late type stars .
we also estimate the radial velocity precision necessary to detect the equivalent hz earth - mass exoplanet that also transits when using an equal amount of observing time as the transit survey .
we find that a radial velocity survey with @xmath00.6 m s@xmath1 precision has comparable efficiency in terms of observing time to a transit survey with the requisite photometric precision @xmath2300 ppm to find a transiting earth - sized exoplanet in the hz of late m dwarfs . for super - earths ,
a @xmath02.0 m s@xmath1 precision radial velocity survey has comparable efficiency to a transit survey with @xmath22300 ppm . |
in the last year a large excitement has arisen in the heavy ion physics community , related to the possibility of quark gluon plasma ( qgp ) already been obtained at sps energies @xcite . in particular several signals
were mentioned , which point out to the existence of qgp .
putting aside the abnormal @xmath4 suppression and the excess of dileptons found , there are three signals related to baryon and strangeness production , namely the large enhancement of the ( anti)hyperon yields ( @xmath5 , @xmath6 , @xmath7 ) in pb - pb collisions compared to p - pb , observed by the wa97 @xcite and the na49 @xcite collaborations data done by the na49 collaboration gives yields at midrapidity which are in much closer agreement to the wa97 @xcite results than the previous analysis of na49 @xcite . ] ; the linear increase of the inverse exponential slope of the @xmath0 distributions ( temperature ) in pb - pb collisions with the mass of the observed particle , except for @xmath7 ( for pions there is also a small departure from the straight line ) @xcite ; and the different behaviour of the temperature between p - p and a - a collisions . concerning this last feature , while the temperature in p - p collisions remains flat as a function of particle mass , in heavy ion collisions it increases as the mass increases .
furthermore , for a given mass , the heavier the colliding system , the higher the temperature @xcite .
these last two characteristics have been interpreted as the existence of an intrinsic freeze - out temperature and a collective hydrodynamical flow which is gradually developed : firstly , for s - s collisions , and , in a more clear way , in pb - pb collisions .
we would like to study these three signals , together with other important related data as @xmath8 production @xcite , different particle ratios @xcite and stopping power @xcite , in a monte carlo code based on the string fusion model ( sfm ) @xcite .
compared to previous versions , the model has been improved , including minijet production and rescattering of the produced secondaries in a standard way ( an extensive description of these new features will be presented elsewhere @xcite ) .
summarizing , hard collisions have been included using pythia @xcite , in order to reproduce the non - exponential tail of the @xmath0 spectrum experimentally measured in p - p collisions at spps and tevatron ; this is crucial for the applicability of the model at rhic and lhc .
rescattering is introduced as @xmath9 collisions @xcite between mesons and baryons , taking into account both strangeness conserving and strangeness exchange reactions , and the possibility of the inverse processes as required by detailed balance .
only three cross sections are used : the first one for reactions involving @xmath7 and @xmath10 , the second one for reactions related to @xmath11 annihilation , and the third one for all other reactions . in the string fusion model
it is assumed that strings fuse when their transverse positions come within a certain interaction area . in the code
we consider only fusion of two strings but there is a probability of fusion of more than two . an effective way of taking this into account is to increase the cross section for the fusion of two strings , for which we will take @xmath12 mb .
this value is crucial to reproduce the strangeness enhancement in central s - s and s - ag collisions at sps @xcite .
the fusion can take place only when the rapidity intervals of the strings overlap .
it is formally described by allowing partons to interact several times , the number of interactions being the same both for projectile and target .
the quantum numbers of the fused strings are determined by the interacting partons and their energy - momentum is the sum of the energy - momenta of the ancestor strings .
the colour charges of the fusing string ends sum into the colour charge of the resulting string ends according to the su(3 ) composition laws .
the breaking of each fused string is due to the production of two ( anti)quark complexes with the same colour charges @xmath13 and @xmath14 as those at the ends of the string @xcite .
the probability rate is given by the schwinger formula @xcite @xmath15 where @xmath16 is the string tension for the @xmath17 su(3 ) representation , proportional to the corresponding quadratic casimir operator @xmath18 @xcite . in our case @xmath19 ,
@xmath20 and @xmath21 .
therefore , the @xmath22 and @xmath23 fused strings have a higher string tension , giving rise to larger heavy flavour and baryon / antibaryon production .
hard strings are not fused , their area being proportional to @xmath24 .
some effect of the fusion of such strings could appear at lhc energies where , for instance , in central pb - pb collisions they amount for 35 % of the binary nucleon - nucleon collisions @xcite . in principle
, the fusion of strings means nothing related to a phase transition . on the contrary , percolation of strings
@xcite is a non - thermal second order phase transition . in this case , the key parameter is @xmath25 , which is the density of strings @xmath26 ( number of strings @xmath27 produced in the overlapping area of the collision , @xmath28 for central collisions ) times the transverse size of one string @xmath29 .
the critical point for percolation is @xmath30 depending on the profile function of the colliding nuclei @xcite . with @xmath31 fm
, this critical value means @xmath32 strings/@xmath33 .
the value of 9 is reached in central pb - pb collisions at sps , in central ag - ag collisions at rhic and in central s - s collisions at lhc .
we expect for @xmath34 around or greater than @xmath35 , that the approximation of fusion of just two strings fails .
in fig . 1 we show our results for @xmath7 , @xmath6 and @xmath5 yields for p - pb , and central pb - pb collisions at sps with four different centralities , together with the experimental data . in order to disentangle the different processes contributing , in fig .
2 it is shown the results of the code for central ( @xmath36 fm ) pb - pb collisions without string fusion and rescattering , with string fusion , and with string fusion and rescattering . a reasonable agreement with data
is obtained . only the @xmath7 s
are a little below the data ( 40 % ) .
similar results have been obtained in the relativistic quantum molecular dynamics model @xcite by a mechanism of colour ropes which consider fusion of strings ; also in the ultra relativistic quantum molecular dynamics model @xcite and in the hijing model @xcite by using an ad hoc multiplicative factor in the string tension . also the dual parton model @xcite ,
considering the possibility of creation of diquark - antidiquark pairs in the nucleon sea , together with the inclusion of diagrams which take into account baryon junction migration @xcite , can reproduce the experimental data ( for @xmath7 s some rescattering has still to be added ) .
the string fusion is the main ingredient to obtain a good agreement with @xmath37 experimental data and also to reproduce the @xmath6 data .
however the rescattering is fundamental to get enough @xmath7 s .
notice also that there is not any jumping between s - s and pb - pb .
the most pronounced enhancement takes place between p - pb and s - s . about @xmath37 ,
our results are higher than the wa97 data , its production being mainly determined by string fusion and hardly affected by rescattering .
this fact makes that our results for pb - pb are really an extrapolation in the model from the value for @xmath37 production in central s - s collisions by the na35 collaboration , which was used to fix the fusion cross section @xmath38 @xcite .
so , from the point of view of our model , there exists either a large @xmath37 annihilation or a conflict between na35 data for s - s and wa97 data for pb - pb . in fig .
3 we plot the inverse exponential slopes of the @xmath0 distributions for different particles , together with the wa97 experimental data .
a reasonable agreement is obtained .
in particular it can be seen that the @xmath7 slope is below the straight line both in the model and in data . also our predictions for rhic are presented .
we see that the @xmath7 is now on the straigth line , so the decrease at sps energies is due to energy - momentum conservation . about @xmath8 enhancement ,
our results without fusion , with fusion , and with fusion and rescattering are 3.55 , 4.20 and 5.35 respectively , in satisfactory agreement with experimental data , @xmath39 @xcite . in fig .
4 the stopping power is shown , i.e. the @xmath40 rapidity distributions for central pb - pb collisions at sps , compared with the experimental data @xcite , together with the predictions for rhic and lhc energies .
this quantity is essentially determined by the string fusion mechanism and rescattering only plays a minor role .
as discussed for strangeness enhancement , it has been pointed out that baryon junction migration @xcite will enhance the stopping power due to diagrams additional to the usual ones of the dual parton model .
the inclusion of these diagrams also explains the sps data .
we have not taken into account such diagrams to avoid double counting , because in the fusion of strings they are partially included in an effective way . in fig .
5 the antiproton rapidity distribution in central pb - pb collisions is presented and compared to the experimental data @xcite ; a great suppression of the antiproton yield is seen , due to rescattering . in table
i our results for the ratios between different particles are compared with the experimental data @xcite for pb - pb central collisions at sps .
we include our predictions for rhic and lhc in table ii .
some comments are in order : first , we observe an overall , rough agreement with the sps data .
second , our results are not very different to those of statistical models @xcite . however , the saturation of the strangeness enhancement in our case has nothing to do with thermal and/or chemical equilibrium . in string fusion
the enhancement of the different strange particle yields is similar to a threshold behaviour : first there is a pronounced rise and afterwards , a saturation .
the main difference in the predictions for rhic and lhc between the string fusion model and statistical models is the overall charged multiplicity , which is respectively 950 and 3100 for sfm and 1500 and 7600 for statistical models @xcite ( assuming initial temperatures of 500 and 1000 mev for rhic and lhc respectively ) .
let us emphasize that we obtain a reasonable agreement with the experimental data in three of the features advocated as signals of qgp production .
we are only below data in the @xmath7 production by less than a factor 2 .
it is possible to improve this agreement arranging the corresponding rescattering cross sections .
however , we think that the usefulness of any monte carlo is in the overall picture and not in obtaining detailed fits which require a fine tuning . for this reason we do not change our rescattering cross sections . nevertheless , related to the @xmath7 production , it is tempting to speculate about percolation of strings as a possible explanation of the discrepancy .
it is just around pb - pb central collisions at sps energies where the critical point @xmath35 , which corresponds to @xmath32 strings/@xmath33 , is placed .
close to this point , the formation of at least one cluster of strings through the whole reaction area is possible .
therefore it is expected that the string fusion model does not work there .
this point can be checked at rhic . there , the density of strings formed in central ag - ag collisions is also around 9 and therefore the number of @xmath7 s would be above the string fusion prediction . as a last point , we compare our results with the preliminary data of the phobos collaboration at rhic @xcite . for charged particles
we obtain @xmath41 and 585 for the 6 % more central au - au collisions at @xmath42 and 130 gev per nucleon respectively , to be compared with @xmath43 and @xmath44 .
our prediction for @xmath45 gev per nucleon with the same centrality criterium is @xmath46 . in conclusion , the agreement of the model with baryon and strangeness experimental data at sps is satisfactory , so the sfm code can be useful as a tool for simulations of the forthcoming experiments at rhic and lhc energies .
we thank m. a. braun and e. g. ferreiro , who participated in early stages of this work .
we also thank a. capella , a. b. kaidalov , c. a. salgado and yu .
m. shabelski for useful discussions , and j. stachel for comments on the predictions of the statistical models for rhic and lhc .
n. a. , c. p. and d. s. acknowledge financial support by cicyt of spain under contract aen99 - 0589-c02 and n. s. a. by academy of finland under grant number 48477 .
n. a. thanks departamento de fsica de partculas of the universidade de santiago de compostela , and d. s. the alice collaboration at cern , for hospitality during stays in which part of this work was completed .
wa97 collaboration : e. andersen _
et al . _ ,
b433 * ( 1998 ) 209 ; _ ibid . _ * b449 * ( 1999 ) 401 .
na49 collaboration : h. appelshuser _ et al .
_ , phys . lett .
* b444 * ( 1998 ) 523 ; eur .
j. * c2 * ( 1998 ) 661 .
na49 collaboration : r. a. barton _
et al . _ , in _ proceedings of the strangeness 2000 conference _ ( berkeley , usa , july 20th-25th 2000 ) .
wa97 collaboration : f. antinori _
et al . _ ,
a661 * ( 1999 ) 481c ; eur .
j. * c14 * ( 2000 ) 633 .
na44 collaboration : n. xu _
et al . _ ,
a610 * ( 1996 ) 175c .
na50 collaboration : n. willis _
et al . _ ,
nucl . phys .
* a661 * ( 1999 ) 534c .
na49 collaboration : g. hhne _
et al . _ ,
* a661 * ( 1999 ) 485c ; s. v. afanasev _
et al . _ ,
b491 * ( 2000 ) 59 .
wa97 collaboration : i. krlik _
et al . _ ,
a638 * ( 1998 ) 115c .
na49 collaboration : g. e. cooper _ et al .
_ , nucl . phys .
* a661 * ( 1999 ) 362c .
n. s. amelin , m. a. braun and c. pajares , phys . lett .
* b306 * ( 1993 ) 312 ; z. phys . * c63 * ( 1994 ) 507 .
n. armesto , m. a. braun , e. g. ferreiro and c. pajares , phys . lett .
* b344 * ( 1995 ) 301 ; e. g. ferreiro , c. pajares and d. sousa , phys . lett . *
b422 * ( 1998 ) 314 .
n. s. amelin , n. armesto , c. pajares and d. sousa , in preparation .
t. sjstrand , comput .
physics commun .
* 82 * ( 1994 ) 74 .
p. koch , b. mller and j. rafelski , phys .
* 142 * ( 1986 ) 167 .
j. schwinger , phys .
rev . * 82 * ( 1951 ) 664 ; e. brezin and c. itzykson , phys
* d2 * ( 1970 ) 1191 .
g. s. bali , nucl .
* 83 * ( 2000 ) 831 ; preprint hub - ep-99 - 67 ( hep - ph/0001312 ) .
n. armesto , m. a. braun , e. g. ferreiro and c. pajares , phys .
* 77 * ( 1996 ) 3736 .
a. rodrigues , r. ugoccioni and j. dias de deus , phys .
b458 * ( 1999 ) 402 . h. sorge , phys . rev .
* c52 * ( 1995 ) 3291 .
s. a. bass _ et al .
_ , nucl . phys .
* a661 * ( 1999 ) 205c .
m. bleicher , m. belkacem , s. a. bass , s. soff and h. stcker , phys . lett . *
b485 * ( 2000 ) 213 .
m. bleicher , w. greiner , h. stcker and n. xu , hep - ph/0007215 .
s. e. vance , in _ proceedings of the strangeness 2000 conference _
( berkeley , usa , july 20th-25th 2000 ) , nucl - th/0012056 .
a. capella and c. a. salgado , phys . rev .
* c60 * ( 1999 ) 054906 ; preprint lpt - orsay-00 - 66 ( hep - ph/0007236 ) ; a. capella , e. g. ferreiro and c. a. salgado , phys . lett .
* b459 * ( 1999 ) 27 .
s. e. vance and m. gyulassy , phys .
lett . * 83 * ( 1999 ) 1735 .
a. capella and b. z. kopeliovich , phys
b381 * ( 1996 ) 325 ; b. z. kopeliovich and b. povh , phys .
b446 * ( 1999 ) 321 ; d. kharzeev , phys . lett . * b378 * ( 1996 ) 238 ; f. w. bopp , hep - ph/0002190 .
na49 collaboration : j. bachler _
et al . _ ,
a661 * ( 1999 ) 45c .
braun - munzinger , i. heppe and j. stachel , phys . lett . *
b465 * ( 1999 ) 15 .
p. braun - munzinger , nucl . phys .
* a661 * ( 1999 ) 261c .
j. stachel , in _ proceedings of the xxixth international symposium on multiparticle dynamics _
( providence , usa , august 9th-13th 1999 ) , to be published by world scientific .
n. armesto and c. pajares , int .
a15 * ( 2000 ) 2019 .
phobos collaboration : b. b. back _ et al .
* 85 * ( 2000 ) 3100 . | strangeness and baryon enhancement in heavy ion collisions are discussed in the framework of the string fusion model .
the monte carlo version of this model is shown to reasonably reproduce three of the features that have been pointed out as evidences of the finding of the quark gluon plasma . namely hyperon / antihyperon enhancement , dependence of the slope of the @xmath0 distributions on the mass of the produced particles and on the centrality of the collision , are described .
predictions for rhic and lhc energies are presented . 0.6 cm n. s. amelin@xmath1 , n. armesto@xmath2 , c. pajares@xmath3 and d. sousa@xmath3 0.6 cm _ @xmath1 department of physics , university of jyvskyl , fin-40351 jyvskyl , finland _ 0.3 cm _ @xmath2 departamento de fsica , mdulo c2 , planta baja , campus de rabanales , _ _ universidad de crdoba , e-14071 crdoba , spain _ 0.3 cm _ @xmath3 departamento de fsica de partculas elementales , _ _ universidade de santiago de compostela , e-15706 santiago de compostela , spain _ 0.6 cm |
the equation for the high temperature superconducting state has the form @xcite ( see also @xcite and @xcite ) : @xmath30 where the quantities @xmath14 and @xmath15 denote the effective pairing potentials for the electron - phonon and electron - electron - phonon interaction , respectively .
the symbol @xmath31 is the amplitude of the @xmath13-wave order parameter ; @xmath32 $ ] .
the function @xmath33 is given by the expression : @xmath34 where @xmath35 denotes the electron band energy : @xmath36 ; @xmath37 is the hopping integral and @xmath38 $ ] . the quantity @xmath39 has been defined as : @xmath40 , where @xmath41 is the boltzmann constant .
the equation ( [ ra1 ] ) has too much a complicated form to be solved in the analytical way .
for this reason , it has been analyzed by using the numerical methods .
additionally , we have assumed that : @xmath42 , where @xmath43 is the heaviside function .
the normalization constant is given by the following : @xmath44 ; the symbol @xmath45 represents the characteristic phonon frequency , which is of the order of debye frequency .
the energy gap amplitude at the temperature of zero kelvin is defined as : @xmath46 .
[ cols="^,^,^,^,^,^,^,^,^,^,^ " , ] | the model for the cuprates based on the modified electron - phonon pairing mechanism has been tested . for this purpose ,
the superconductors with high value of the critical temperature have been taken into consideration .
in particular : @xmath0 , @xmath1 , @xmath2 , and @xmath3 .
it has been shown that the dependence of the ratio @xmath4 on the doping ( @xmath5 ) can be properly predicted in the framework of the presented theory ; the symbol @xmath6 denotes the energy gap amplitude at the temperature of zero kelvin , and @xmath7 is the critical temperature .
the numerical results have been supplemented by the formula which describes the function @xmath8 .
# 1([#1 ] ) # 1fig.[#1 ] # 1[#1 ] in the solid state physics , the issue of the correctly determining the pairing mechanism in the high - temperature superconductors ( cuprates ) is very controversial @xcite .
essentially , there is a clash of the two views .
first of them is based on the original frhlich idea indicating that the interaction of the electron gas with the phonons is responsible for the formation of the superconducting state @xcite .
the second view favors the clean electron correlations , which in the simplest way can be modeled by the one - band hubbard hamiltonian @xcite .
it is true that in the cuprates , the classical electron - phonon interaction is over one level weaker than the electron correlations @xcite .
hence , it may seem that the limitation to the correlations in the electronic subsystem is correct .
unfortunately , in the framework of the hubbard model with the positive values of the potential ( @xmath9 ) it is very difficult to describe the thermodynamic properties of the high - temperature superconducting state .
what is even worse , there is no convincing evidence that the superconducting state can exist at the sufficiently high temperature @xcite . on the other hand , in the case of the negative value of the interaction potential with the same intensity of @xmath9 , proving
the existence of the superconducting state with the high critical temperature is a simple matter .
additionally , it is easy to prove that the pairing mechanism based on the classical electron - phonon interaction is insufficient to fully describe the superconducting state in the cuprates @xcite , @xcite . in the presented paper
, we have tested the pairing mechanism , which combines two opposing approaches .
the new mechanism has been proposed recently in @xcite ( see also @xcite and @xcite ) .
it is based on three postulates : ( i ) _ in the superconductivity domain of the cuprates the fundamental role is played by the electrons on the @xmath10 planes . _
( ii ) _ in the cuprates there exists the conventional electron - phonon interaction , which does not have to be strong . _
( iii ) _ in the cuprates there exist strong electronic correlations , but the electron - electron scattering in the superconductivity domain is inseparably connected with absorption or emission of the vibrational quanta .
_ and @xmath11 as the functions of the doping in ybco .
the empty circles represent the averaged values of @xmath11 .
the filled circles and squares are the experimental data : ( a)- kabanov _ et al . _
@xcite , ( b)- daou _ et al . _
@xcite , ( c)- solovjov _ et al . _
@xcite , ( d)- obolenskii _ et al . _
@xcite , ( e)- prokofev _ et al . _
@xcite , ( f)- timusk _ et al . _
@xcite , ( g)- chaban @xcite , ( h)- vignolle _ et al . _
@xcite , ( i)- lin _ et al . _
@xcite , ( j)- rice _
et al .
_ @xcite .
+ ( b ) the dependence of the ratio @xmath12 on the doping for ybco .
the empty circles represent the numerical results .
the line has been prepared on the basis of the formula .
the stars and the triangles are the experimental data : ( a ) - sutherland _ et al . _
@xcite , ( b ) - nakayama _ et al . _
@xcite , ( c ) - kaminski _ et al . _
@xcite , ( d ) - plate _
et al . _
@xcite , ( e ) - morr _ et al . _
@xcite , fong _ et al . _
@xcite , ( f ) - yeh _ et al . _
@xcite , ( g ) - born _ et al . _
@xcite , ( h ) - murakami _ et al . _
@xcite , ( i ) - edwards _ et al . _
@xcite , ( j ) - edwards _ et al . _
@xcite , ( k ) - tsai _ et al . _
@xcite , ( l ) - schrieffer _ et al . _
@xcite , ( m ) - maggio - aprile _ et al . _
@xcite , @xcite , ( n ) - koinuma _ et al . _
@xcite , ( o ) - koyanagi _ et al . _
@xcite , ( p ) - kugler _ et al . _
@xcite , ( r ) - nantoh _ et al . _
@xcite , @xcite , ( s ) - shibata _ et al . _
@xcite , ueno _
et al . _
@xcite , @xcite , ( t ) - kirtley _ et al . _
@xcite , ( u ) - hoevers _ et al . _
@xcite , ( w ) - tanaka _ et al .
_ @xcite . ] and @xmath11 as the functions of the doping in hg1201 .
the empty circles represent the averaged values of @xmath11 .
the filled circles and squares are the experimental data : ( a ) , ( b)- yamamoto _ et al . _
@xcite , ( c)- bandyopadhyay _ et al . _
@xcite , ( d)- yamamoto _ et al . _
@xcite , ( e)- honma _ et al .
_ @xcite .
+ ( b ) the dependence of the ratio @xmath12 on the doping for hg1201 .
the empty circles represent the strict numerical results .
the line has been prepared on the basis of the formula .
the stars and the triangles are the experimental data : ( a)- wei _ et al . _
@xcite , ( b)- hasegawa _ et al . _
@xcite , ( c)- yang _ et al . _
@xcite , ( d)- guyard _ et al . _
@xcite , ( e)- inosov _ et al .
_ @xcite . ]
the first postulate emphasizes the importance of the quasi two - dimensionality of the system .
the second one refers to the classical pairing mechanism given by frhlich .
the third postulate states that the strong electron correlations in the cuprates are inseparably coupled with the phonon subsystem .
the first two postulates define the van hove scenario @xcite , @xcite .
the third postulate requires further discussion because it is far more subtle .
in particular , it should be noted that the postulated electron correlations generalize the hubbard approach ; the classical two - body interaction is replaced with the three - body interaction ( electron - electron - phonon ) . it should be pointed out very clearly that the third postulate does not require the additional phonon channel to be as strong as the electron channel . by analyzing the set of the given
postulates in the terms of the classical theory of the superconductivity @xcite , @xcite , it can be noticed that the first and second theory is based on the same philosophy .
namely , the physical system is constantly treated in a holistic manner , without the artificial division into the electron and phonon subsystem .
the hamiltonian corresponding to the presented postulates has been derived in the paper @xcite .
then , using the canonical transformation and the formalism of the thermodynamic green functions one can get the equation determining the properties of the @xmath13-wave superconducting state .
a detailed form of the equation is presented in the appendix a. the considered model has two main input parameters : the electron - phonon pairing potential @xmath14 and the electron - electron - phonon potential @xmath15 .
the first parameter can be calculated based on the value of the critical temperature ( @xmath7 ) , the second parameter has been chosen in such a way that it reproduces the value of the temperature at which the pseudogap appears ( @xmath11 ) .
the values of two other input parameters ( the hopping integral and the characteristic phonon frequency ) have been taken from the literature .
on the basis of the calculated input parameters , we have determined the ratio of the energy gap amplitude at the temperature of zero kelvin to the critical temperature ( @xmath4 ) .
the obtained results have been compared with the experimental data . in particular , the following compounds have been studied : @xmath0 ( ybco ) , @xmath1 ( hg1201 ) , @xmath2 ( hg1201-zn ) , and @xmath3 ( hg1223 ) . it should be noted that the considered compounds are the examples of the superconductors with the highest critical temperature ( the optimal doping ) . and
@xmath11 as the functions of the doping in hg1201-zn .
the diagram has been made on the basis of the paper by yamamoto _
et al .
_ @xcite .
+ ( b ) the dependence of the ratio @xmath12 on the doping .
the empty circles represent the strict numerical results for hg1201-zn .
the line has been prepared on the basis of the formula .
the green filled circles show the results achieved for hg1201 . ] and @xmath11 as the functions of the doping in hg1223 .
the diagram has been formed on the basis of the following experimental data : ( a)- liu _ et al .
_ @xcite and ( b)- lam _
et al .
_ @xcite .
+ ( b ) the dependence of the ratio @xmath12 on the doping for hg1223 .
the empty circles represent the strict numerical results .
the line has been prepared on the basis of the formula .
the experimental data : ( a)- jeong _ et al . _
@xcite , ( b)- rossel _ et al .
_ @xcite . ] in the first step , the ratio @xmath12 for the superconductor ybco has been calculated . in ( a )
there are presented the experimental courses of @xmath16 and @xmath17 , where @xmath5 denotes the concentration of the holes ( the doping ) .
the function @xmath16 has been plotted on the basis of the data included in the paper @xcite ; the values @xmath17 have been determined averaging the large number of the experimental data .
then , on the basis of the equation ( 1a ) , the pairing potentials @xmath14 and @xmath15 have been calculated .
the achieved results and the detailed descriptions of the experimental data used in this paper have been summarized in the appendix b. next , we have prepared the dependence @xmath8 ( ( b ) ) .
it has been found that the theoretical results correctly reproduce the experimental data .
let us notice that the dependence of the critical temperature on the doping for the cuprates has been parametrized with the help of the expression : @xmath18 @xcite .
the shape of the function @xmath8 can be reproduced in a similar way . in the considered case ,
the following result has been obtained : @xmath19_{\rm dbcs},\ ] ] where @xmath20 is the value of the concentration of the holes at which the superconducting state appears ; @xmath21 is the fitting parameter . for ybco
, it has been assumed that : @xmath22 and @xmath23 .
the symbol @xmath24_{\rm dbcs}$ ] denotes the value of the ratio @xmath12 calculated in the framework of the @xmath13-wave bcs model : @xmath24_{\rm dbcs}=4.28 $ ] @xcite .
the form of the analytical results has been presented in ( b ) . in the same way the course of the dependence of @xmath12 on @xmath5 in the family of the high temperature superconductors containing mercury has been submitted .
the results have been presented in - ; similarly for analysis as in the case of ybco , in the appendixes ( c ) , ( d ) and ( f ) , we have collected the values of the pairing potentials and the detailed descriptions of the experimental data . in the case of the superconductor hg1201 ,
the shape of the function @xmath16 has been obtained on the basis of the papers @xcite , @xcite , and @xcite ; the averaged values of @xmath17 have been determined with the help of the experimental data presented in ( a ) .
the theoretical results and the experimental data have been collected in ( b ) .
it has been found that the theoretical predictions agree with the experimental data .
then , by using the formula , the course of the function @xmath8 has been reproduced ; the following parametrization has been selected : @xmath25 and @xmath26 .
also in this case the formula correctly reproduces the dependence of the ratio @xmath12 on the concentration of the holes .
the superconductor hg1201-zn represents the case for which no experimental values of @xmath8 have been found in the literature . however , there are well defined functions @xmath16 and @xmath17 ( ( a ) ) .
thus , the course of the dependence @xmath8 can be calculated theoretically .
the strict numerical data have been collected in ( b ) .
in addition , there have been presented the results obtained on the basis of the formula , for which it has been assumed that : @xmath27 and @xmath23 .
it is easy to notice that for @xmath28 , in the relation to the course obtained for hg1201 , the disorder induced by zinc does not exert the significant influence on @xmath8 .
much bigger deviations can be noticed for the lower values of the doping .
the discussed effect can be relatively easily verified by the experimenters - we strongly encourage to do so .
the superconductor hg1223 is characterized with the highest observed critical temperature for the optimal doping ( @xmath29 k ) @xcite .
the full courses of @xmath16 and @xmath17 for hg1223 have been determined in the paper @xcite and @xcite ( ( a ) ) . using the already discussed method of the analysis
, we have obtained the theoretical form of the function @xmath8 .
the results have been presented in ( b ) .
we have found that the theoretical predictions reproduce the existing experimental data very well . additionally , in ( b )
there have been plotted the analytical values of @xmath8 ; we have assumed that : @xmath22 and @xmath23 .
to summarize : in the presented paper , we have calculated the dependence of the ratio @xmath12 on the doping for four superconductors characterized by high values of the critical temperature . in the case of ybco , hg1201 and hg1223 ,
the theoretical results reproduce the experimental data very well . for the superconductor hg1201-zn
only the theoretical results have been provided due to the fact that the relevant experiments have not been undertaken yet .
finally , we have shown that the shape of the function @xmath8 can be reproduced by the analytical formula .
the presented theory can be used to explore all other thermodynamic parameters of the cuprates .
it seems that one of the most fundamental issues is associated with the interpretation of the experimental data obtained by using the arpes method .
such studies are currently underway and , in a short time , the results will be presented .
the authors wish to thank prof .
k. dziliski , the head of the institute of physics at czstochowa university of technology , for providing excellent working conditions and the financial support .
additionally , we would like to thank our colleagues : d. szczniak , m.w .
jarosik , e.a .
drzazga , and t. mila for their kindness and the technical support offered during the preparation of this paper .
the calculations have been conducted on the czstochowa university of technology cluster , built in the framework of the platon project , no .
poig.02.03.00 - 00 - 028/08 - the service of the campus calculations u3 . |
incompressible quantum liquids ( iqls)@xcite underlie the fractional quantum hall effect ( fqhe ) discovered by tsui , stormer , and gossard.@xcite the charge carriers in these liquids are anyons , i.e. , quasiparticles ( quasielectrons and quasiholes ) carrying fractional charges@xcite and obeying fractional statistics.@xcite historically the main experimental discoveries in this field were done by magnetotransport experiments .
however , the role of spectroscopic methods is continuously increasing since they provide an indispensable tool for investigating spectra of elementary excitations .
fine structure specific for different electronic phases was discovered in the spectra of radiative photoemission.@xcite these spectra were used to measure gaps in the energy spectra of iql s , for investigation of phase transitions between the iql and wigner solid phases , etc .
the frequency of long - wave neutral elementary excitations of iql s was measured in raman scattering experiments.@xcite a challenging problem in physics of iql s is a direct observation of the charge fractionalization . between different exciting approaches to this problem
the spectroscopic approach seems to be one of the most promising .
indeed , spectroscopy permits one to observe properties of iql s in the bulk where the effect of the impurities and edges is reduced to the minimum .
the intrinsic spectroscopy of iql s is the magnetospectroscopy of excitons .
however , properties of excitons reflect the spectrum of the elementary excitations of an iql ( quasielectrons and quasiholes , magnetorotons,@xcite etc . ) and can be treated in terms of them only when the separation @xmath2 between electron and hole confinement planes is sufficiently large . indeed , when @xmath13 , where @xmath4 is the magnetic length , the filling factor @xmath14 of the liquid strongly deviates in the vicinity of the hole from the quantized value @xmath15 .
under these conditions the properties of the iql can not be treated in terms of its quasiparticles .
the spectroscopy of a remote hole has been discussed from different standpoints in a number of papers.@xcite despite the fact that experiments in the @xmath16 region are rather complicated , the separations up to @xmath17 were achieved in experiments on extrinsic radiative photoemission.@xcite of special importance might be experiments performed for a fixed filling factor @xmath18 and a variable dimensionless electron - hole separation @xmath19.@xcite the first experiments of this kind were reported recently.@xcite the investigation of excitons is also important from the different standpoint . in the theory of iql
s the statistical properties of the system of free anyons are usually discussed . in an exciton
the anyons exist in a bound state because of the attracting potential of a hole .
it was shown@xcite that the energy spectra of excitons of @xmath20 and @xmath21 iql s are closely related to the statistical and dynamical properties of quasielectrons of these iql s .
therefore , the exciton problem can be really treated as a few - anyon problem .
excitons represent a wide class of few - anyon systems .
it includes excitons , anyon ions@xcite ( which can be also treated as quasiparticle - exciton complexes@xcite ) , some impurity centers , etc . in what follows ,
we consider excitons as bound few - anyon composites and investigate their properties in some detail .
there are two approaches to the theory of excitons in iql s .
the first approach is based on the exact diagonalization for finite - size systems in the spherical geometry.@xcite having in mind workable system sizes , it provides reliable results for @xmath22 , at least as applied to the @xmath20 and @xmath21 iql s .
the second approach is based on the anyon exciton model ( aem ) proposed by the present authors.@xcite in the framework of this approach an exciton is considered as a neutral composite quasiparticle consisting of several anyons and a hole .
this model is exact only when the exciton size which is about @xmath2 is large compared to the size of anyons which is about @xmath4 , i.e. , for @xmath23 .
therefore , the two approaches are complementary and one can expect that they match when @xmath24 . it is the main statement of the aem that excitons of iql s possess a multiple branch spectrum .
indeed , a charged particle at the lowest landau level possesses a single degree of freedom and a single quantum number .
an exciton being a neutral entity possesses in a magnetic field a vector momentum @xmath5 absorbing two degrees of freedom.@xcite therefore , an exciton consisting of @xmath25 anyons and a hole possesses @xmath26 internal degrees of freedom . for an ordinary magnetoexciton @xmath27 and the spectrum
consists of a single branch . for @xmath28
an exciton acquires internal quantum number(s ) and multiple branch spectrum .
this prediction of the aem permitted apalkov _
et al._@xcite to represent the energy spectra found by finite - size computations for @xmath29 and @xmath1 iql s as a system of exciton branches and to assign to these branches internal quantum numbers .
the latter determine the values of the exciton angular momentum , @xmath6 , in the @xmath7 states .
zang and birman@xcite and chen and quinn@xcite also inferred existence of several exciton branches in their finite - size data .
in the range accessible for finite - size studies , @xmath22 , there are two types of excitons in the lower part of the energy spectrum , anyon excitons and tight excitons .
_ anyon excitons _ are loose entities with a pronounced anyon - density dip at the center .
they are generically related to the quantum states making up the low - energy ( anyon ) sector@xcite of the electron subsystem .
each anyon exciton is a bound state of a three - quasielectron complex from the anyon sector and a hole .
this finding establishes a connection between the spectroscopy of excitons and the low - energy physics of the fqhe .
there is a striking difference in the quantum numbers of the low - energy anyon excitons of the @xmath0 and @xmath1 iql s .
it is related to the difference in the energy spectra of the three - quasielectron complexes originating due to the different populational statistics of the quasielectrons of these liquids .
_ tight excitons _ are dense entities .
a sharp density maximum is achieved either in the center of an exciton or in a close vicinity of the center .
these excitons are not related to the low - energy sector of the electron subsystem .
the @xmath30 tight exciton is the bottom state of the exciton spectrum in all the region @xmath22 .
therefore , the @xmath3 region seems to be of the most importance for the study of the anyon substructure of excitons and for the spectroscopic observation of the charge fractionalization .
we suppose everywhere in this paper that the spin - polarized background is stable with respect to formation of charged spin - textures ( skyrmions ) .
it was predicted recently@xcite that spin - textures develop in the ground state of a @xmath31 exciton for large values of @xmath19 ; the critical value of @xmath19 is about 1.5 for gaas .
nevertheless , we feel that the spin - polarized ground state of a @xmath20 magnetoexciton is stable because skyrmions can exist only at very low magnetic fields when @xmath32.@xcite spin - depolarized excitons should be seen in the spectrum of excited states . in this paper
we develop the theory of excitons in the framework of the aem as applied to excitons consisting of three quasielectrons and a hole .
we use halperin pseudo - wave - functions@xcite in a boson basis and investigate the energy spectrum , electron density distribution , and anyon correlations in an exciton .
we make comparison with the finite - size data@xcite on the excitons of the @xmath0 and @xmath1 iql s and conclude that excitons of the @xmath21 iql are described rather well by the aem because of the bosonic populational statistics and the narrow form - factors of the quasielectrons of this liquid .
we also relate the difference in the energy spectra of the anyon excitons of the @xmath0 and @xmath1 iql s to some specific features in the exciton shape found in the framework of the aem . for both iql s
the bottom exciton states are made by tight @xmath30 excitons for @xmath22 and by a succession of anyon excitons for @xmath3 .
this succession consists of hard core excitons with the angular momenta @xmath6 which are integers of 3 and increase as @xmath33 with @xmath2 .
the general outline of the paper is as follows . in sec . [
sec : functions ] we construct a full basis of the exciton wave functions for the aem using the translational symmetry and permutation symmetry arguments . in sec .
[ sec : hamilt ] we develop a technique for calculating different matrix elements entering the schroedinger equation . in secs .
[ sec : energy ] , [ sec : density ] , and [ sec : correlation ] we obtain energy spectra , electron density distributions , and the density correlation functions , respectively .
the latter functions unveil the anyon substructure of excitons both in the finite - size data and in the aem . in sec.[sec :
comparison ] we make a comparison of the results obtained in the framework of the aem with finite - size data of refs . and .
we propose that a striking difference in the finite - size data for the @xmath0 and @xmath1 iql s originates from the difference in the populational statistics of the quasielectrons of these liquids .
let us consider an exciton consisting of a valence hole with a charge @xmath34 and three qe s with electrical charges @xmath35 and statistical charges @xmath36 .
such an entity provides the aem description of the anyon excitons of the @xmath0 and @xmath1 iql s . for an @xmath20 iql the statistical charge equals @xmath37 , while for an @xmath21 iql the statistical charge has the same value , @xmath38 , as for quasiholes in an @xmath20 iql .
@xcite in comparison , @xmath39 for bosons and @xmath40 for fermions . in the strong magnetic field limit ,
when the coulomb energy @xmath41 , where @xmath42 is the cyclotron frequency and @xmath43 is the dielectric constant , it is convenient to employ dimensionless variables scaled in units @xmath44 , @xmath4 and @xmath45 .
we use the symmetric gauge , @xmath46 , where @xmath47 is a unit vector perpendicular to the confinement plane . instead of the hole , @xmath48 , and anyon , @xmath49 ,
coordinates it is convenient to introduce the following two - dimensional ( 2d ) coordinates : @xmath50 @xmath51 has a meaning of the center - of - mass of the exciton coordinate .
the coordinates @xmath52 and @xmath53 are the internal variables which are not affected by the translational motion of an exciton .
complex coordinates @xmath54 , as well as @xmath53 , are not independent . indeed , @xmath55 despite the fact that the constraint ( [ eq1.2 ] ) results in some complications , the introduction of the variables @xmath53 enables one to develop the theory in a form symmetric in all anyons and , therefore , finally simplifies the equations .
anyons and hole live in two different parallel planes separated by the distance @xmath2
. nevertheless , only the 2d coordinates of eq .
[ eq1.1 ] enter into the exciton wave function .
the separation @xmath2 enters only into the hamiltonian of the anyon - hole interaction derived in the sec .
[ subsec : anyhole ] . the most general form of the pseudo - wave - function of an anyon exciton meeting all general requirements is as follows : @xmath56 where the pair of indices , @xmath57 , takes the values specified in eq .
( [ eq1.1 ] ) , @xmath58 is the normalization area , and @xmath59 is a homogeneous polynomial in coordinates @xmath60 of the degree @xmath6 .
the basic properties of the functions @xmath61 can be checked by inspection .
( i ) since the exciton is a neutral entity , it possesses an in - plane momentum @xmath5,@xcite and @xmath61 satisfies the equation of magnetic translations : @xmath62 \psi_{l,{\bf k}}({\bf r } - { \bf a } , { \bbox \rho } , \ { { \bar{z}_{jl}}\ } ) = { \rm e}^{- i{\bf k}\cdot { \bf a } } \psi_{l,{\bf k}}({\bf r } , { \bbox \rho } , \ { { \bar{z}_{jl}}\})~. \label{eq1.4}\ ] ] the parameter @xmath63 is related to the dipole moment of the exciton , @xmath64 .
( ii ) the function @xmath61 belongs to the lowest landau level .
indeed , the nonanalytic factor of it can be shown to have the form @xmath65 , whereas the other factors are analytic functions of @xmath66 and @xmath67 .
( iii ) the function @xmath68 is a homogeneous polynomial of the degree @xmath6 which is symmetric in all coordinates @xmath69 .
these polynomials form a boson basis , and the effect of the fractional statistics is taken into account by the factor @xmath70.@xcite ( iv ) a system of four charged particles in a magnetic field possesses four quantum numbers .
two of them are absorbed in the 2d momentum @xmath5 .
two others determine the form of the polynomial @xmath59 and are internal quantum numbers of an anyon exciton .
the operator @xmath71 of the @xmath72 projection of the angular momentum commutes with the hamiltonian and the square of the momentum , @xmath73 , but it does not commute with the projections of @xmath74 , i.e. , with @xmath75 and @xmath76 .
therefore , the function @xmath61 chosen in the @xmath77 representation is simultaneously an eigenfunction of @xmath71 only for @xmath78 , and in this limit @xmath79 .
therefore , the quantum numbers of an anyon exciton include the 2d momentum @xmath5 and the projection of the angular momentum , @xmath79 , of the exciton with @xmath78 . the angular momentum @xmath6 numerates branches of the exciton spectrum . the fourth quantum number , which will be specified in what follows , numerates branches with coinciding values of @xmath6 .
the multiple branch structure of the anyon exciton spectrum is a direct consequence of the charge fractionalization which results in the appearance of the internal degrees of freedom of an anyon exciton and of the related internal quantum numbers . by definition ,
the polynomial @xmath59 is symmetric in coordinates @xmath80 . to establish the symmetry of it in the symmetric coordinates @xmath60
, one can start with a monomial @xmath81 , apply to it all operations of the permutation group , and take the sum over the group .
this transformation results in the polynomial @xmath82 which has different properties depending on the parity of @xmath83 .
if @xmath6 is even , the polynomial is a permanent , and , therefore , is symmetric in the coordinates @xmath60 .
however , when @xmath6 is odd , the polynomial is a determinant which is obviously antisymmetric in the coordinates @xmath60 and is non - equal to zero only for @xmath84 .
for example , for the lowest possible value of @xmath6 , @xmath85 , this determinant turns into a vandermonde determinant @xmath86 therefore , @xmath6-even and @xmath6-odd polynomials @xmath59 have rather different properties .
all of them are symmetric in bosonic permutations @xmath87 , etc . ,
but they have opposite symmetry with respect to the permutations of the @xmath88 type . to find the explicit form of the polynomials @xmath59 , it is convenient to introduce new real coordinates @xmath89 where @xmath90 is the center - of - mass of the anyon sybsystem .
the corresponding complex coordinates are @xmath91 .
these coordinates are subject of the constraint : @xmath92 in these coordinates the bosonic symmetry of @xmath59 has the usual form , and we can apply the fundamental theorem of the theory of symmetric polynomials.@xcite according to it , @xmath93 can be expressed in the unique way as a polynomial in the three elementary symmetric polynomials : @xmath94 the first polynomial is equal to zero , @xmath95 , because of the constraint of eq .
( [ eq1.7 ] ) .
therefore , the polynomials @xmath59 are polynomials only in @xmath96 and @xmath97 .
the first @xmath6-even polynomials are : @xmath98 = const , @xmath99 , @xmath100 , i.e. , there exists only a single elementary polynomial of a given degree @xmath6 . however , two elementary polynomials , @xmath101 and @xmath102 , contribute to @xmath103 .
it is easy to check that the number of basis functions increases by one each time when @xmath6 takes values @xmath104 , where @xmath105 is an integer .
therefore , the number of @xmath6-even polynomials is equal to @xmath106 + 1 $ ] , where @xmath106 $ ] is the integral part of @xmath107 .
all @xmath6-odd polynomials can be obtained by multiplying @xmath6-even polynomials by @xmath108 .
the latter equals @xmath109 because of eq .
( [ eq1.6 ] ) .
we are now in position to choose a full basis of polynomials @xmath59 in symmetric coordinates @xmath60 .
these coordinates are most convenient for all the following calculations .
@xmath6-even polynomials can be chosen as @xmath110 where @xmath111 $ ] .
polynomials @xmath112 are linearly independent , and the total number of polynomials with a given @xmath6 is equal to @xmath106 + 1 $ ] .
all linearly independent @xmath6-odd polynomials can be obtained as : @xmath113 the total number of them equals @xmath114 + 1 $ ] .
this choice of polynomials determines the full set of quantum numbers in the wave function of eq .
( [ eq1.3 ] ) as @xmath6 , @xmath115 , and @xmath5 . to our best knowledge , in the previous studies
only the @xmath6-even polynomials have been taken into account .
@xcite when choosing polynomials @xmath112 , we have not imposed the hard core constraint and defer the discussion of the related properties to what follows .
it is an important feature of the aem that the wave functions ( and , therefore , electron densities , etc . ) of all eigenstates with @xmath116 and also @xmath117 are completely determined by the symmetry requirements .
they do not depend on the specific form of the hamiltonian and , in particular , on @xmath2 .
in this section we calculate the hamiltonian of the aem in the @xmath118 basis as the matrix of a point charge coulomb interaction .
it means , that we neglect form - factors of quasielectrons which have a scale of several magnetic lengths and are known only approximately.@xcite we postpone the discussion of inaccuracy originating from this approximation to sec .
[ sec : comparison ] .
unfortunately , functions @xmath118 are orthogonal only in quantum numbers @xmath6 and @xmath5 . as a result , the scalar products @xmath119 for @xmath120 , and
the matrix @xmath121 of these scalar products is block - diagonal .
the size of blocks is equal to 1 for @xmath122 and @xmath117 and increases by 1 each time when @xmath6 increases by 6 . with a non - diagonal matrix @xmath121 the schroedinger equation
has a form : @xmath123 and one has to find matrices @xmath124 and @xmath121 . to perform the calculations , it is convenient to employ variables @xmath51 , @xmath125 , and three @xmath126 and to take into account the constraint of eq .
( [ eq1.2 ] ) by the usual transformation : @xmath127 it adds the new variable @xmath128 , but all calculations become symmetric in anyon variables .
the jacobian of the transformation is equal to 1 .
the hamiltonian is diagonal in @xmath5 , therefore , we write out only the diagonal in @xmath5 matrix elements . for @xmath78 ,
the hamiltonian is also diagonal in @xmath6 , and eq .
( [ eq2.1 ] ) acquires a block - diagonal form . since all terms in the polynomials @xmath129 , eqs .
( [ eq1.9 ] ) and ( [ eq1.10 ] ) , have the same form , we concentrate in what follows on the matrix elements taken in the basis of the functions @xmath130 : @xmath131 here polynomials @xmath132 are substituted by monomials , and @xmath133 denotes a set of quantum numbers @xmath134 and @xmath135 . in the next parts of this section we describe in some detail the technique for performing different types of integrals .
the scalar products of functions @xmath130 , when written in the variables @xmath51 , @xmath125 , and @xmath53 , have the form : @xmath136 integration over @xmath51 and the gaussian integration over @xmath125 are straightforward .
since @xmath130 is multiplicative in the variables @xmath53 , one can rewrite eq .
( [ eq2.4 ] ) in the form : @xmath137 where @xmath138 since @xmath139 where @xmath140 is a bessel function , the angle integration in ( [ eq2.6 ] ) results in : @xmath141~ { \cal m}_{mm'}^{(\alpha)}(t)~ , \label{eq2.8}\ ] ] where @xmath142 , @xmath143 is the azimuth of @xmath128 , and @xmath144 this integral can be expressed in terms of the confluent hypergeometric function @xmath145 as@xcite @xmath146 here @xmath147 is the larger of the integers @xmath105 and @xmath148 . after the integration over @xmath143 in eq .
( [ eq2.5 ] ) , the coefficients @xmath149 take the form : @xmath150 where @xmath151 therefore , the scalar product @xmath149 of two functions @xmath130 is reduced to an one - fold integral from the product of three confluent hypergeometric functions .
matrix elements @xmath149 do not depend on @xmath5 .
scalar products @xmath152 of two functions @xmath118 , which include polynomials @xmath112 , are linear combinations of the coefficients @xmath149 .
they do not depend on @xmath5 , too .
the coefficients of these combinations can be found from the equations ( [ eq1.5 ] ) , ( [ eq1.9 ] ) , and ( [ eq1.10 ] ) , but the final expressions are rather cumbersome , especially for @xmath6-odd polynomials .
therefore , we do not write out here their explicit form . integrals ( [ eq2.11 ] ) can be simplified for @xmath153 .
indeed , in this case the kummer transformation@xcite @xmath154 results in a @xmath155 function with the first parameter @xmath156 equal to a negative integer .
this function reduces to a polynomial , and @xmath157 to a polynomial multiplied by @xmath158 .
therefore , the integral in eq .
( [ eq2.11 ] ) for @xmath149 can be performed exactly .
this transformation highly simplifies calculations .
indeed , in the large @xmath2 region , where the criterion of the applicability of the aem is satisfied , the statistical parameter @xmath36 can be neglected as it is shown in sec .
[ sec : energy ] below .
the hamiltonian of the anyon - anyon interaction is : @xmath159 matrix elements of @xmath160 in the basis of the functions @xmath130 can be calculated by analogy with the matrix elements @xmath149 .
the denominators @xmath161 lower the power of @xmath162 by 1 in one of the @xmath163 factors entering in eq .
( [ eq2.5 ] ) .
the final expression for the matrix element is : @xmath164 like the nonorthogonality matrix elements , matrix elements of @xmath160 also do not depend on the momentum @xmath5 .
for this reason , the matrix of the operator @xmath160 is diagonal in the angular momentum , @xmath6 , for arbitrary values of the momentum @xmath5 .
the hamiltonian of the anyon - hole interaction has a form : @xmath165 where the three - dimensional anyone - hole separation @xmath166 should be expressed in terms of the difference coordinates : @xmath167 similar equations hold for @xmath168 and @xmath169 .
it is convenient to introduce the fourier image @xmath170 where @xmath171 .
the integrations over @xmath51 and @xmath125 in the matrix elements of @xmath172 can be performed in the same way as in eq .
( [ eq2.4 ] ) , and the result can be expressed in terms of the coefficients @xmath173 , eq . ( [ eq2.6 ] ) , as : @xmath174 after the angular integration , the factors @xmath175 take the form of eq .
( [ eq2.8 ] ) , and eq . ( [ eq4.4 ] ) can be rewritten as : @xmath176 vectors @xmath177 are defined by the equation @xmath178 one can perform one angular integration in eq .
( [ eq4.5 ] ) if the phase @xmath143 is eliminated by change in the variables : @xmath179 here @xmath180 are phases of vectors @xmath177 . after the integration over @xmath181 , eq .
( [ eq4.4 ] ) takes its final form : @xmath182 here the direction of @xmath5 was chosen along the @xmath183 axis to eliminate the complex phase from the matrix element ( [ eq4.8 ] ) .
explicit expressions of @xmath184 and @xmath185 in terms of the integration variables are as follows : @xmath186 one can check by inspection that eq .
( [ eq4.8 ] ) is symmetric in the indices @xmath187 and @xmath188 , and that matrix elements are real for all sets of the quantum numbers @xmath189 compatible with our choice of the polynomials @xmath132 , eqs .
( [ eq1.9 ] ) and ( [ eq1.10 ] ) .
matrix elements of the operators @xmath190 and @xmath191 can be written by analogy with eq .
( [ eq4.8 ] ) .
equations ( [ eq3.2 ] ) and ( [ eq4.8 ] ) for the hamiltonian @xmath192 and eqs . ( [ eq2.4 ] ) and ( [ eq2.11 ] ) for the matrix @xmath193 determine completely the schroedinger equation ( [ eq2.1 ] ) .
for @xmath78 the operator of the anyon - hole interaction @xmath194 becomes diagonal in the angular momentum @xmath6 .
therefore , the schroedinger equation ( [ eq2.1 ] ) is also diagonal in @xmath6 since @xmath160 and @xmath193 are diagonal in @xmath6 for arbitrary @xmath195 .
it was shown in sec .
[ sec : functions ] that for each value of @xmath6 such that @xmath116 or @xmath196 there exists a single eigenfunction , and it does not depend on @xmath2 . for these values of @xmath6 the equations of sec .
[ sec : hamilt ] immediately give the energies of @xmath7 states .
when the number of @xmath132 polynomials with a given value of @xmath6 becomes two or more , equations of sec .
[ sec : hamilt ] give the coefficients of secular equations of the second , third , etc . ,
order which determine the @xmath2 dependent eigenfunctions and eigenvalues .
the order in which @xmath78 levels are arranged changes with increasing @xmath2 .
the main regularities can be understood using classical arguments . in the classical limit , which is achieved for a large exciton size
, the exciton ground state takes the shape of an equilateral triangle with anyons in the vertices and a hole in the center .
the anyon - anyon distance in this triangle , @xmath197 , found from the minimum of the electrostatic energy , is equal to @xmath198 it will be shown below that triangular configurations are described by the polynomials @xmath199 and @xmath200 . a straightforward calculation based on equations ( [ eq1.3 ] ) ,
( [ eq1.9 ] ) and ( [ eq1.10 ] ) shows that a mean - square value of the inter - anyon distance in these states is equal to @xmath201 comparing equations ( [ eq5.1 ] ) and ( [ eq5.2 ] ) one finds that @xmath202 for @xmath203 .
therefore , with increasing @xmath2 both the size , @xmath197 , and the angular momentum , @xmath6 , of the exciton ground state increase .
this means that the order of the energy levels changes , and the bottom state possesses the angular momentum @xmath204 .
since for @xmath78 the hamiltonian is diagonal in @xmath6 , the level interchange occurs usually as a level crossing . in fig .
[ fig1 ] the ground state energy is plotted as a function of @xmath2 for two values of the statistical charge , @xmath37 and @xmath153 , by curves a and b , respectively . it is seen that both curves show the same gross features , including increase of the ground - state energy and the angular momentum with @xmath2 . however , fine details are very different in the @xmath205 region .
we do not discuss these differences in more detail since they are expected to be sensitive to anyon form - factors@xcite which were not taken into account in our calculations .
however , since curves a and b practically coincide for @xmath206 , we believe that in this region the aem provides reliable results . in what follows
we restrict ourselves with this region and neglect the statistical charge , i.e. , consider the bosonic model , @xmath153 .
all data below are presented for this model .
it was shown in sec .
[ subsec : nonorth ] , eq .
( [ eq2.13 ] ) , that for @xmath153 matrix elements @xmath207 can be expressed in terms of elementary functions .
this fact permits one to reduce the three - fold integral of eq .
( [ eq4.8 ] ) to the one - fold integral : @xmath208 where functions @xmath209 , real and symmetric in indices , are polynomials in @xmath25 .
the lower polynomials are of a simple form : @xmath210 .
application of the equation ( [ eq5.4 ] ) highly simplifies all computations . for comparison , in fig .
[ fig1 ] is also shown the energy @xmath211 calculated in the classical limit , eq .
( [ eq5.1 ] ) .
it is seen that in the region @xmath212 it differs considerably from the exact quantum data . since for @xmath213
the term @xmath194 is non - diagonal in @xmath6 , the dispersion law @xmath214 can be found only numerically . in fig .
[ fig2 ] it is shown for two values of @xmath2 . the basis of polynomials used in computations included 22 @xmath6-even polynomials with @xmath215 , eq .
( [ eq1.8 ] ) , and 22 @xmath6-odd polynomials with @xmath216 , eq . ( [ eq1.9 ] ) .
the following regularities are distinctly seen . as it was argued above
, the levels with higher @xmath6 values draw closer to the spectrum bottom with increasing @xmath2 .
the level interchange manifests itself as avoided level crossings .
the level splitting near these crossings increases with @xmath217 and decreases with the difference @xmath218 .
these regularities can be understood if one takes into account that the bessel function in the integrand of eq .
( [ eq5.4 ] ) shows the power - law behavior , @xmath219 , in the small @xmath217 region . in particular , for @xmath220 the interaction of two branches can result in the negative exciton effective mass for small @xmath217 values .
negative exciton dispersion can appear even in the ground state as it can be seen in fig .
it is interesting to mention that exciton dispersion near @xmath78 is always positive in the two - semion problem.@xcite the above results show that the charge fractionalization determines both i ) the basic multiple - branch structure of the exciton energy spectrum and ii ) numerous specific features of the spectrum including the @xmath2 dependence of the arrangement of the branches , avoided branch intersections , etc .
the distribution of the electron density , @xmath221 , around a hole can be found from the equation : @xmath222 where the averaging is performed over the quantum state ( @xmath223 ) , where @xmath224 numerates exciton branches .
the density @xmath221 is exactly the quantity which i ) permits one to check reliability of the model and in which ii ) the specific pattern of the charge fractionalization manifest itself in a rather spectacular way .
it is one of the basic criteria of the aem that the excess charge density is small compared with the density of the iql , @xmath225 , for @xmath226 ( and compared with @xmath227 for @xmath228 ) .
therefore , for @xmath20 and @xmath21 iql s one can expect that the aem becomes applicable only when @xmath229 . this criterion will be applied in what follows .
the explicit expression for @xmath230 can be obtained in the same way as eq .
( [ eq5.4 ] ) . indeed , the operator @xmath194 and the operator of the electron density of eq .
( [ eq6.1 ] ) depend on the same arguments , @xmath231 .
therefore , the integrands differ only in the substitution @xmath232 by the fourier image of @xmath233 .
the final expression is : @xmath234 \nonumber \\ & & \times \int_{0}^{\infty } { dq~q~\exp(-3q^{2}/2)~j_{|l - l'|}(q|{\bf r}-{\bf d}|)}~ q_{lm , l'm'}(q)~ { \bar \chi}_{l'm'}^{\lambda}({\bf k } ) \chi_{lm}^{\lambda}({\bf k})~ , \label{eq6.2}\end{aligned}\ ] ] where @xmath235 is the angle between the vectors @xmath236 and @xmath237 .
@xmath238 shows the electron density on the hole . for @xmath7 ,
the density distribution @xmath239 is shown in fig .
[ fig3 ] for @xmath240 . in this case
@xmath224 can be completely identified by the index @xmath6 for @xmath116 , but there are two functions for @xmath241 .
for @xmath242 the energies , the eigenfunctions , and therefore also the densities @xmath243 , do not depend on @xmath2 .
the @xmath244 state has a high density ; @xmath245 .
therefore , the shape of the curve @xmath246 can not be reliable .
nevertheless , it is remarkable that numerical calculations performed in the spherical geometry for @xmath20 @xcite and @xmath21 iqls@xcite convincingly show that for @xmath247 the spectrum bottom is made by the @xmath244 exciton having the electron density which is very close to the fermi limit , @xmath248.@xcite functions @xmath249 become broader with increasing @xmath6 , and @xmath250 decreases .
the state @xmath251 from which @xmath6-odd polynomials originate is especially remarkable because it is the first to show a crater - like density distribution with a minimum at @xmath252 .
this minimum is a signature of the charge fractionalization , since for ordinary magnetoexcitons the density shows a maximum at @xmath253 .
we will discuss the properties of this state in more detail in sec .
[ sec : correlation ] . for @xmath254
there are two eigenfunctions ; they depend on @xmath2 . in fig .
[ fig3 ] they are shown for @xmath255 ; the lower energy component is drawn by a solid line .
[ fig4 ] shows the energy , @xmath256 , and the density @xmath257 for bottom states as a function of @xmath2 .
it is seen from the figure that for @xmath206 the density falls well below its critical value 0.05 , which supports our above conjecture , sec .
[ sec : energy ] , that the aem provides reliable results for @xmath206 . with increasing @xmath2 the angular momentum in the bottom state , @xmath6 , also increases .
it is a striking feature of the data that only states with @xmath258 , @xmath259 , reach the spectrum bottom ( we can not make definite conclusions about the @xmath251 state since it does nor reach the bottom for @xmath260 anyons , fig . [ fig1 ] ) .
the bottom states described by @xmath6-even and @xmath6-odd polynomials alternate .
we attribute the periodicity in @xmath6 to the superselection rule originating from the combination of the space and permutation symmetry .
indeed , we observe this periodicity in the semiclassical region where the quantization rule includes angular integration between two exchange points separated by the angle @xmath261 rather than the usual @xmath262 integration.@xcite polynomials @xmath132 with @xmath263 play a special role in the class of @xmath6-even polynomials .
all of them obey the hard core constraint . indeed
, the polynomial @xmath264 vanishes as @xmath265 each time when one of its arguments , @xmath60 , turns into zero . from the standpoint of the general theory,@xcite
hard - core functions are the only `` legitimate '' wave functions of an anyon system .
the exponent @xmath266 is the maximum order of the zero for a wave function with the angular momentum @xmath6 , and this maximum is achieved only for @xmath199 polynomials .
therefore , coulomb repulsion is strongly suppressed for these polynomials .
polynomials @xmath200 play a similar role in the class of @xmath6-odd polynomials .
it is a remarkable fact , that in the @xmath206 region all bottom states are either @xmath199 or @xmath200 polynomials . using the polynomials @xmath267 , @xmath268 and 1 , as an example
, we can follow the way in which hard core states move to the spectrum bottom when @xmath2 increases .
for small @xmath2 values the low energy component of the @xmath241 doublet has a pronounced maximum near @xmath253 as it is seen in figs .
[ fig3 ] and [ fig5 ] . for @xmath269
the low - energy component is close to @xmath270 ; for @xmath255 the overlap is 0.96 . for @xmath271 wave functions of both components
are strongly mixed , and they show similar distributions of the density , fig .
[ fig5 ] . for @xmath272
the function @xmath273 wins the competition .
for @xmath274 it dominates in the low - energy state ; the overlap is 0.97 .
the density @xmath275 has a single maximum for each value of @xmath115 .
one can obtain a simple analytic expression for the position of the maximum by averaging the density over @xmath276 .
this latter function , @xmath277 , reaches the maximum at @xmath278 , and the maximum of @xmath279 is very close to this value .
more detailed information on the nature of the @xmath78 bottom states comes from the correlation functions which are discussed in sec .
[ sec : correlation ] below . for comparison , in fig .
[ fig4 ] is also shown the @xmath2 dependence of the energy of a conventional magnetoexciton , @xmath280 , with the momentum @xmath78 . in the limit
@xmath281 this energy exactly coincides with the energy of an exciton in the many - electron system because of the hidden symmetry inherent in the problem ; see refs . and and references therein .
in the region of @xmath282 the accuracy of the aem is low .
however , it increases for @xmath206 when the charge fractionalization becomes important . in this region @xmath280 follows the usual coulomb law , @xmath283 , whereas for many - electron systems the dependence of the exciton binding energy on @xmath2 also is close to a coulomb law , but the numerator is considerably less than 1 since the electron density distribution has a width about @xmath2 .
the magnetoexciton and aem approaches are exact in the opposite limits .
the results should be matched in the intermediate region at @xmath284 . above in this section we discussed the electron density distribution only for @xmath78 anyon excitons .
the charge fractionalization manifests itself for these excitons in the crater shape of @xmath285 .
however , the most spectacular manifestation of the charge fractionalization can be expected in the large @xmath217 region , @xmath286 .
indeed , the exciton dipole moment @xmath237 differs from @xmath195 only by the rotation by @xmath287 , sec .
[ sec : functions ] .
therefore , one can expect that with increasing @xmath217 the electron density splits into bundles , their charges being multiples of @xmath288 .
the splitting of the electron shell into two well separated quasiparticles has been observed previously for a two - semion exciton.@xcite for a three - anyon exciton the pattern are much more impressive . for the bottom state ,
they are shown in fig .
[ fig6 ] for @xmath289 when the criterion of the large electron - hole separation is fulfilled .
the distribution which is cylindrically symmetric for @xmath290 transforms with increasing @xmath217 into a distribution with a single split - off anyon ( @xmath291 and 3 ) .
two anyons constituting the exciton core show a slight but distinct splitting in a perpendicular direction . this core can be considered as an anyon ion .
the core changes its shape with @xmath217 but remains stable in a wide range of @xmath217 . finally , for rather large @xmath217 values , it splits in the @xmath237 direction as it is seen in the last figure , @xmath292 .
the asymmetric density distribution for @xmath9 arises completely due to the admixture of @xmath6-odd polynomials to the @xmath254 state .
the well outlined profiles of the electron density seen in fig .
[ fig6 ] may be smeared by the oscillatory screening inherent in iqls.@xcite nevertheless , the basic pattern of the charge separation in an exciton should strongly influence the @xmath217 dependence of the magnetoroton - assisted recombination processes since charge density excitations are left in a crystal afterwards .
one can see in fig .
[ fig3 ] that @xmath85 state is the first state which shows a crater - like shape of the density @xmath293 .
this shape indicates the existence of the anyon substructure of an exciton as it was argued in sec . [
sec : density ] .
it is typical of all bottom states with @xmath294 . in this section
we compare properties of @xmath295 and @xmath273 states and show that despite the similarity in the shape of the density , they differ critically in the shape of the radial pair correlation function @xmath296 .
it is convenient to use the square of the wave function @xmath297 of eq .
( [ eq1.3 ] ) averaged over the hole coordinate . using eqs .
( [ eq1.3 ] ) and ( [ eq1.9 ] ) and performing gaussian integration over @xmath276 , one can write the following equation for the averaged @xmath298 function : @xmath299 choosing @xmath197 , @xmath300 , and @xmath301 , where @xmath301 is the angle between the vectors @xmath302 and @xmath303 , as independent variables , rewriting @xmath304 in terms of these variables , and looking for the maximum first over @xmath305 and then over @xmath197 and @xmath300 , one finds that the maximum of @xmath304 is reached for a configuration of an equilateral triangle with @xmath306 in the semiclassical limit , @xmath203 , this result coincides with eq .
( [ eq5.2 ] ) . to find the most probable configuration for the @xmath295 state it is convenient to work in @xmath307 variables , sec .
[ sec : functions ] , and perform averaging over @xmath276 .
simple calculation shows that @xmath308 the maximum of this expression under the constraint of eq .
( [ eq1.7 ] ) can be found in the same way as for eq .
( [ eq7.1 ] ) . finally , one recovers an equilateral triangle configuration with @xmath309 .
this result coincides with eq .
( [ eq7.2 ] ) for @xmath85 . because arbitrary @xmath6-odd polynomials have the form @xmath310 , eq .
( [ eq1.10 ] ) , and the equilateral triangle configuration is optimal for each of the multipliers , it is optimal also for their product , @xmath112 .
therefore , the most probable configuration has the same shape of an equilateral triangle both for @xmath6-even and @xmath6-odd states . to reveal a striking difference in the properties of @xmath273 and @xmath311 states
, one can calculate the radial pair correlation functions @xmath296 : @xmath312 substituting @xmath313 from eq .
( [ eq1.3 ] ) results in @xmath314 straightforward calculation shows that @xmath315 @xmath316 both functions are shown in fig .
the function @xmath317 has a hard core behavior , whereas @xmath318 does not vanish at @xmath252 . on the contrary
, it reaches its absolute maximum at this point .
the second maximum is by the factor 0.97 lower than the main one .
therefore , function @xmath295 violates the hard core constraint . of course , all bottom - state @xmath6-odd polynomials with @xmath319 show a hard core behavior .
the above conclusion on the triangular shape of the most probable configuration of the @xmath295 exciton implies existence of the 120@xmath8 correlations in the density correlation function . to investigate these correlations
it is convenient to work in the @xmath320 variables .
the two - particle correlation function @xmath321 depending on the angle @xmath322 between vectors @xmath323 and @xmath324 can be written as : @xmath325 the delta function takes into account the constraint of eq .
( [ eq1.7 ] ) , and @xmath326 is standing for the angle between @xmath323 and @xmath324 expressed in terms of the coordinates of these vectors .
integration over the variables @xmath320 in eq .
( [ eq7.8 ] ) results in the averaged correlation function , the main contribution comes from the area @xmath327 where @xmath328 reaches the maximum .
the last integral can be performed in polar coordinates , @xmath329 , @xmath330 , @xmath331 , and the normalized function @xmath332 takes the form : @xmath333 the integration over @xmath235 can be performed analytically , however , the final expression is rather cumbersome .
the result is plotted in fig .
two distinct maxima in the vicinity of @xmath261 and @xmath334 reveal 120@xmath8 correlations in the @xmath295 state .
the maxima are shifted from the angle @xmath335 since the function @xmath295 vanishes for the collinear configuration of anyons , @xmath336 .
apalkov and rashba@xcite have found the density - density correlation function , @xmath337 , for the @xmath21 iql with a single extra electron .
calculations were performed in the spherical geometry for the @xmath338 , @xmath339 quantum state .
the quantity @xmath340 , which is the difference between the three - quasiparticle angular momentum in the spherical geometry , @xmath341 , and the maximum value of this momentum , should be compared to the exciton angular momentum @xmath6.@xcite it is convenient to introduce mean values @xmath342 and @xmath343 here @xmath344 are unit vectors designating the positions of the electrons on the sphere .
electron density depends only on the polar angle @xmath345 and equals @xmath346 , where @xmath347 is the number of electrons .
if one introduces the deviation , @xmath348 , of the density from its mean value , the density - density correlation function can be written as @xmath349 in eq .
( [ eq7.12 ] ) the polar angles of the vectors @xmath350 and @xmath351 are chosen equal , @xmath352 . therefore , the correlation function , @xmath353 , is the function of the azimuth @xmath322 and depends on @xmath345 as a parameter . in eq .
( [ eq7.12 ] ) the singular term proportional to @xmath354 is omitted , as usual , since it makes no contribution to @xmath355 for @xmath356 .
the results of calculations are presented in fig .
[ fig9 ] for three values of @xmath345 .
the value of @xmath357 is close to the maximum of @xmath358 . since the anyon exciton wave function @xmath295 does not depend on @xmath2 , the data of figs .
[ fig8 ] and [ fig9 ] can be compared .
there is a striking similarity between them .
all three curves in fig .
[ fig9 ] show flat but distinct maxima near the same values of the argument , @xmath359 and @xmath334 .
these maxima are much flatter than the maxima in fig .
this difference can be attributed to the smearing originating from the quasielectron form - factors .
nevertheless , there is no doubt that it is the three quasielectron substructure of the many electron state @xmath85 which manifests itself in 120@xmath8 correlations .
therefore , the state @xmath295 shows properties intermediate between an anyon exciton and tight exciton .
unfortunately , reliable calculation of @xmath360 can not be performed for the @xmath361 states since their size exceeds the accessible sphere size .
in this section we make a comparison of the aem with computational results for excitons in the spherical geometry .
three - anyon excitons are expected to appear in two charge conjugate iql s , @xmath20 and @xmath21 .
their properties in the semiclassical region , @xmath362 and @xmath203 , should be identical . for intermediate @xmath2 and @xmath6 values , excitons of these two liquids
are expected to show rather different properties .
the data obtained by finite - size computations@xcite substantiate these expectations .
there are different reasons why @xmath20 and @xmath21 excitons are expected to have different properties in the intermediate region .
a simple electrostatic reason is a different nonlinear screening of the unit positive charge by these two iql s .
the effect of the screening on the exciton energy spectrum was discussed in ref . . however , there are also different mechanisms resulting in the difference in the properties of these excitons .
they are related to rather different form - factors and different statistical properties of the quasielectrons of the @xmath20 and @xmath21 iql s . neglecting spin effects
, we can consider electrons as spin - less fermions .
then , because of the charge symmetry , the properties of the quasielectrons of the @xmath21 iql are identical to the properties of the quasiholes of the @xmath20 iql .
therefore , we will start with a comparative study of quasielectrons and quasiholes of the @xmath20 iql .
the form - factors of charged particles of the @xmath20 iql were investigated in a number of papers .
the data are summarized in refs . - .
it is known that a quasihole has a narrow profile with a radius up to two magnetic lengths .
the density decreases away from the center of a quasihole nearly monotonically . on the contrary
, a quasielectron has a pronounced density dip at the center , the density maximum at about two magnetic lengths , and the radius of about four magnetic lengths .
therefore , the model of point anyons developed above matches much better the excitons of the @xmath21 iql than the excitons of the @xmath20 iql .
another aspect of the problem is related to statistical properties of anyons . in the aem
the effect of the fractional permutational statistics of anyons was taken into account by including the factor @xmath363 into the halperin pseudo - wave - function @xmath61 , eq .
( [ eq1.3 ] ) .
it was shown in sec .
[ sec : energy ] that the effect of this factor can be neglected for @xmath206 , i.e. , in the region where the aem is expected to be valid .
however , one should also take into account the nontrivial populational statistics of anyons .
this can be done using the theory of composite fermions@xcite and the approach to the dimensionality of the quasiparticle space based on exclusion statistics.@xcite bosonic haldane dimension , @xmath364 , is an effective number of the single - quasiparticle states defined in such a way that the usual bose distribution @xmath365 results in the correct number of states in the hilbert space of @xmath366 quasiparticles , @xmath367 .
the number of quasiparticle states , @xmath368 , can be found by counting the number of states in the hilbert space of composite fermions.@xcite if quasiparticles obey bose statistics , @xmath369 does not depend on @xmath366 . by counting the composite fermion number - of - states in the spherical geometry ,
it was shown in ref .
that for @xmath370 , @xmath105 is an integer , the bosonic dimension of quasiholes equals @xmath371 and the bosonic dimension of quasielectrons equals @xmath372 here @xmath347 is the number of electrons .
these equations are consistent with the diagonal coefficients of the exclusion statistics of quasiholes and quasielectrons ( @xmath373 and @xmath374 , respectively ) , found by different authors.@xcite equations ( [ eq8.1 ] ) and ( [ eq8.2 ] ) indicate that quasiholes show a bosonic behavior , while quasielectrons are subject of the constraint which is even more restrictive than the constraint imposed by fermi statistics .
these conclusions were supported@xcite by counting the maximum value of the angular momentum , @xmath375 , for the system of @xmath366 quasiparticles .
it is equal to @xmath376 for @xmath377 quasiholes , and to @xmath378 for @xmath379 quasielectrons . since @xmath380 is the angular momentum of a single quasiparticle,@xcite eq .
( [ eq8.3 ] ) confirms the bosonic behavior of quasiholes , whereas eq .
( [ eq8.4 ] ) confirms the existence of the restriction on the population of single - quasiparticle states by quasielectrons .
et al._@xcite were the first to discover this restriction by means of numerical calculations and to propose eqs .
( [ eq8.1 ] ) and ( [ eq8.2 ] ) .
they attributed the restriction to the hard core constraint for quasielectrons having a dynamical ( short - range repulsion ) rather than statistical origin . for a macroscopic system , @xmath381 , equations ( [ eq8.1 ] ) and ( [ eq8.2 ] ) give coinciding results , @xmath382 , in the dilute limit , @xmath383 .
however , in an exciton the quasiparticles are confined inside the volume about @xmath384 , where @xmath162 is the exciton radius . therefore , the second term of eq .
( [ eq8.2 ] ) which signifies the deviation from the bosonic behavior of quasielectrons , can be of importance .
quasielectrons of the @xmath21 iql are described by eqs .
( [ eq8.1 ] ) and ( [ eq8.3 ] ) .
therefore , the population of single - quasiparticle states obeys the bose statistics , and eq .
( [ eq1.3 ] ) for wave functions is absolutely adequate since it includes polynomials @xmath112 symmetrical in variables @xmath385 .
for this reason , and also taking into account narrow form - factors of quasielectrons of the @xmath21 iql , one can conclude that the @xmath21 iql is the best candidate for comparison with the aem .
we believe , that the criterion @xmath206 , established in sec . [
sec : density ] by evaluating the density @xmath285 , is applicable to the @xmath21 iql . the situation is more involved for the @xmath20 iql .
quasielectrons in this case are described by eqs .
( [ eq8.2 ] ) and ( [ eq8.4 ] ) .
one can apply eq .
( [ eq8.2 ] ) to the area inside an exciton and order that @xmath386 , which results in @xmath387 .
the number of electrons inside an exciton can be evaluated as @xmath388 .
evaluating the exciton radius @xmath162 as @xmath389 and using the classical equation ( [ eq5.1 ] ) , one comes to the criterion @xmath390 .
a wide quasielectron form - factor imposes a similar restriction on @xmath2 .
since the reliable finite - size computations can be performed only for @xmath391 , the prospects for a quantitative comparison of the results obtained by both approaches seem less favorable for the @xmath20 iql than for the @xmath21 iql . however , we feel that the above rigid criterion relaxes considerably when a qualitative description of the spectrum - bottom states is concerned .
we are now in position to compare the basic results of the aem with the computational results for finite - size systems .
the basic statement of the theory of anyon excitons , that the charge fractionalization results in a multiple - branch energy spectra of excitons,@xcite was confirmed by finite - size computations performed for both @xmath20 and @xmath21 iql s . actually , it provides the basic idea for representing the energy spectra obtained in the spherical geometry for a discrete set of @xmath6 values in the form of exciton branches @xmath392 .
we begin with summarizing some results obtained in ref . by finite - size calculations for excitons of the @xmath21 iql and compare them with the results obtained in this paper in the framework of the aem .
classification of excitons in terms of tight and anyon excitons is used , cf .
[ sec : introduction ] .
_ number of exciton species_. anyon exciton is a bound state of a three - quasielectron complex from the low - energy ( anyon ) sector@xcite and a hole .
the angular momentum of the exciton , @xmath6 , is equal to @xmath393 , where @xmath394 is the angular momentum of the complex , and @xmath395 is the maximum value of this momentum which can be found from eq .
( [ eq8.3 ] ) .
therefore , the number of exciton species is equal to the number of the three - quasielectron complexes in the low - energy sector , and @xmath396 should be compared to the exciton angular momentum @xmath6 of the aem .
only a single three - quasielectron complex exists if the angular momentum @xmath396 equals @xmath396 = 0 , 2 , 3 , 4 , 5 , or 7 .
the complex @xmath397 is absent because composite fermions obey fermi statistics . for @xmath398
the number of @xmath6-even states increases by one ( see fig . 1 in ref . ) .
these properties are in a complete agreement with the classification of the polynomials @xmath112 of sec .
[ sec : functions ] , eqs .
( [ eq1.9 ] ) and ( [ eq1.10 ] ) .
the @xmath399 exciton is absent because of the constraint of eq .
( [ eq1.2 ] ) .
_ l _ = 0 _ branch_. despite the fact that the @xmath30 exciton appears , according to its quantum number , as the first state in the series of anyon excitons , it possesses rather special properties .
this exciton originates from the @xmath400 three - quasielectron complex which is quite dense , and the energy of this complex is high .
it is nearly the same as the energy of some states from the next sector .
these data imply that the @xmath30 anyon exciton actually merges with the @xmath30 tight exciton .
because of these arguments , the @xmath244 exciton was assigned in ref . as a tight exciton rather than an anyon one .
this assignment is supported by an independent argument .
@xmath30 excitons of the @xmath401 and @xmath402 iql s show nearly identical properties , whereas the anyon exciton assignment of the latter entity is excluded by the symmetry arguments based on the composite fermion theory.@xcite the shape of the @xmath244 exciton density distribution of fig .
[ fig3 ] is in agreement with this assignment .
the magnitude of @xmath403 exceeds the maximum density compatible with the aem ; see the discussion at the beginning of sec .
[ sec : density ] .
_ electron density_. for a system with a single extra electron against a background of the @xmath21 iql , the electron density has a pronounced maximum at @xmath253 if @xmath404 or 4 .
the @xmath405 state is the first state with a crater - like electron density distribution having a density dip at @xmath253 ( see fig . 2 in ref . ) .
the crater shape of the density indicates the charge fractionalization .
these data are in a qualitative agreement with the electron density distributions of fig .
[ fig3 ] for excitons having angular momenta @xmath6 = 0 , 2 , 3 , and 4 .
the general shape of the curves is the same , but there are differences in the magnitudes of the densities at @xmath253 . since wave functions of these excitons do not depend on @xmath2 , electron densities for the exciton and three - quasielectron states with the angular momenta @xmath406 , respectively , can be compared .
_ bottom states_. @xmath30 exciton remains the bottom state in the whole area @xmath391 .
@xmath407 , @xmath85 , and @xmath408 excitons move down to the spectrum bottom , but they start competing with the @xmath244 exciton only for @xmath409 when the accuracy of finite - size calculations becomes ambiguous .
@xmath410 exciton is also seen in the low - energy part of the spectrum but never reaches the spectrum bottom.@xcite these data are in agreement with fig .
[ fig4 ] where the sequence of the first bottom states includes @xmath411 and 6 .
_ charge fractionalization : density correlation function_. it was shown in sec . [
sec : correlation ] that @xmath321 reveals 120@xmath8 correlations both for the many - electron and aem wave functions .
these correlations signal the charge fractionalization .
_ intrinsic angular momenta of anyons_. there is a convincing one - to - one correspondence between the excitons of many electron systems having angular momenta @xmath412 in the spherical geometry and the excitons of the aem with the same values of the @xmath72 projections of the angular momentum .
therefore , our data provide no indication of the existence of the intrinsic angular momenta of anyons.@xcite on the contrary , our data are in agreement with the recent conjecture on the absence of anyon spins in the plane limit.@xcite the aem predicts identical exciton spectra for the @xmath21 and @xmath20 iql s .
however , finite - size calculations@xcite result in a rather different symmetry of the low - energy exciton states for @xmath391 . the exact classification of the exciton states based on the composite fermion theory shows that only @xmath251 and @xmath413 anyon excitons can exist in the @xmath0 iql.@xcite this conclusion is supported by numerical data .
therefore , a challenging question arises : why @xmath414 and @xmath415 excitons of the aem are missing from the many - electron spectra of the @xmath0 iql ?
we argue , that these excitons are excluded because of their small bosonic dimension which can not accommodate three quasielectrons . in what follows
we compare the results of the finite - size computations of ref . for the @xmath20 iql with the aem data .
_ excluded states_. finite - size calculations and composite fermion theory show that anyon excitons of the @xmath20 iql can only posses angular momenta @xmath251 and @xmath413 . all excitons with @xmath6 = 0 , 1 , 2 , and 4 can only appear as tight excitons . @xmath30
tight exciton forms the spectrum bottom for @xmath391 , but there is no low - energy @xmath414 and @xmath415 excitons .
it seems probable that there can exist only one , @xmath30 , tight exciton near the spectrum bottom .
therefore , it is necessary to understand why @xmath414 and @xmath415 excitons of the aem , which possess low energies , do not appear as anyon excitons in many - electron systems .
the criterion @xmath416 results in the minimum anyon - exciton radius of @xmath417 .
it is seen in fig .
[ fig3 ] that the criterion @xmath418 is violated for @xmath414 and @xmath415 excitons .
therefore , we attribute the exclusion of the @xmath414 and @xmath415 exciton states from the many - electron spectrum to the reduction of the bosonic dimension of the quasielectron space because of the second term of eq .
( [ eq8.2 ] ) .
_ bottom states .
_ tight @xmath244 exciton remains the bottom state up to @xmath284 when the @xmath85 anyon exciton reaches the bottom .
@xmath419 and @xmath254 excitons start to compete with the @xmath30 exciton only for @xmath420 when the accuracy of finite - size calculations is low .
the data are in agreement with fig .
[ fig4 ] .
_ electron density distribution for _ @xmath78 excitons .
tight excitons have narrow electron density distributions .
the density of the @xmath244 exciton is sharply peaked at @xmath253 .
@xmath85 and @xmath421 anyon excitons show pronounced density dips at @xmath253 and two maxima of the density .
it was hypothesized@xcite that the @xmath253 dip originates from the dip in the quasielectron form - factor , whereas the two - maxima shape indicates the existence of a two - anyon core and one split - off anyon .
it is remarkable that the density distribution in the @xmath85 and @xmath421 excitons only weakly depends on @xmath2 .
weak @xmath2 dependence of the density distribution for @xmath85 and @xmath421 excitons is in agreement with the existence of a single polynomial @xmath112 for @xmath412 , sec .
[ sec : functions ] .
splitting of the exciton shell of an @xmath78 exciton can not be described within the framework of the point - anyon aem .
_ dependence of the density distribution on _ @xmath195 .
when @xmath217 increases , the @xmath253 dip in the electron density of the @xmath85 exciton transforms into a narrow maximum , and the density distribution acquires a three maxima shape .
it was proposed@xcite that in this region of @xmath217 values an exciton consists of a single - anyon core and a two - anyon shell .
splitting off of anyons from the core with increasing @xmath217 is in a qualitative agreement with fig .
the model of anyon exciton developed and solved in the paper includes three quasielectrons ( anyons ) and one hole .
it is applicable to exciton spectra of two charge conjugate iql s , @xmath0 and @xmath1 , and is exact in the limit of a large separation between electron and hole confinement planes , @xmath16 .
anyon excitons possess multiple branch energy spectra , @xmath392 .
an exciton is described by a 2d momentum @xmath5 and two internal quantum numbers @xmath224 which numerate exciton branches .
one of these quantum numbers can be chosen as the exciton angular momentum , @xmath6 , in the @xmath7 state .
a full set of basis functions was chosen with a proper account of the magnetic translational symmetry and permutational symmetry .
the functions include two types of polynomials symmetric in anyon permutations , one of them , apparently , was considered for the first time .
analytic expressions for all matrix elements were derived . as a result , exact expressions for the energy spectrum of a four - particle system were found for @xmath7 . all exciton states with even angular momenta @xmath122 and odd momenta @xmath422
are non - degenerate , and their wave functions are completely determined by symmetry requirements .
this property manifests itself in finite - size data in a weak @xmath2 dependence of the electron density distribution .
properties of @xmath9 excitons were investigated by numerical solution of the schroedinger equation .
the aem is not applicable for small @xmath2 values , @xmath22 .
however , the analysis of the internal criteria of the aem and comparison with the finite - size data show that it gives satisfactory results for @xmath12 , especially for excitons of the @xmath1 iql .
finite - size computations which are reliable for @xmath22 , and the aem , whose accuracy increases with @xmath2 , constitute two supplementary approaches which match at @xmath12 . according to the aem , exciton states with the angular momenta obeying the superselection rule @xmath10 , where @xmath11 is an integer , form the sequence of the bottom states for @xmath3 .
all these states are hard core states , i.e. , their wave functions turn into zero if any two of the anyon coordinates coincide .
the equilateral anyon configurations are the most probable ones .
angular momenta @xmath6 of the bottom states increase as @xmath423 , and the size of these states as @xmath2 .
the tight @xmath30 state is the bottom state for @xmath22 , whereas @xmath251 state , which possesses the properties of both anyon and tight excitons , can reach the bottom for @xmath12 .
anyon substructure of excitons manifests itself in their different properties .
first , the multiple branch energy spectrum originates because of the existence of the internal degrees of freedom .
second , for @xmath78 excitons a crater - like shape of the electron density distribution with a pronounced dip near the hole signals the existence of the charge fractionalization .
third , 120@xmath8 correlations in the electron density were found both in the framework of the aem and in finite - size data .
fourth , for @xmath9 excitons a spectacular splitting of the electron density into bunches provides a direct manifestation of the fractional charge substructure .
aem results in identical low - energy spectra for the excitons of the @xmath0 and @xmath1 iql s .
this result is definitely correct in the @xmath16 limit .
finite - size data suggest that in the intermediate region , @xmath424 , the @xmath0 iql possesses less anyon exciton branches than the @xmath1 iql . following ref .
, we attribute this phenomenon to the difference in the populational statistics of the quasielectrons of the charge conjugate iql s .
anyon statistics enters into the theory in two ways .
first , through the fractional - power factors in the wave function ; they ensure correct interchanging statistics .
second , through the populational statistics .
we have shown that the first mechanism can be neglected for @xmath3 , whereas the second one is of a critical importance in the intermediate region .
it follows from the composite fermion theory that quasielectrons of the @xmath1 iql obey bosonic populational statistics . as a result
, quantum numbers of anyon excitons found from the aem and from the finite - size data exactly coincide . on the contrary ,
the bosonic dimension , @xmath425 , of the quasielectron space of the @xmath0 iql rapidly decreases with the number of quasielectrons , @xmath379 .
the condition @xmath426 , written for the area about the exciton size , eliminates the exciton states of the aem with the angular momenta @xmath414 and @xmath415 from the many - electron spectrum and brings in agreement the aem and many - electron data .
therefore , for @xmath12 the aem supplemented with anyon statistics arguments matches the many - electron data for both iql s and sheds light on the origin of the difference in their exciton spectra .
the above arguments are rather general and can be applied to different problems of the theory of the bound states of several anyons . in conclusion
, the anyon exciton model unveils the general pattern of the exciton spectra of iql s .
it predicts the properties of excitons in the large @xmath2 limit , @xmath16 , and is in agreement with finite - size data in the intermediate region , @xmath12 .
the theory suggests that the region @xmath3 is most favorable for investigating the anyon substructure of excitons in optical experiments .
we are grateful to a. l. efros and y .- s .
wu for discussions and comments , and to e. v. tsiper for suggestions related to @xmath6-odd polynomials .
e.i.r . acknowledges the support of the san diego supercomputer center where finite - size computations were performed .
we also acknowledge the support of the quest of the ucsb by subagreement kk3017 .
present address : department of physics , heriot - watt university , edinburgh eh14 4as , united kingdom .
r. b. laughlin , phys .
lett . * 50 * , 13 ( 1983 ) .
d. c. tsui , h. stormer , and a. c. gossard , phys .
48 * , 1559 ( 1982 ) .
b. i. halperin , phys .
* 52 * , 1583 ( 1984 ) .
d. arovas , j. r. schrieffer , and f. wilczek , phys .
lett . * 53 * , 722 ( 1984 ) .
i. kukushkin , v. timofeev , k. von klitzing , and k. ploog , festkrperprobleme * 28 * , 21 ( 1988 ) ; d. heiman , b. b. goldberg , a. pinczuk , c. w. tu , a. c. gossard , and j. h. english , phys . rev
. lett . * 61 * , 605 ( 1988 ) ; a. j. turberfield , s. h. heines , p. a. wright , r. a. ford , r. g. clark , j. f. ryan , j. j. harris , and c. t. foxon , phys .
* 65 * , 637 ( 1990 ) ; e. m. goldis , s. a. brown , r. b. dunford , a. g. davies , r. newbury , r. g. clark , p. e. simmonds , j. j. harris , and c. t. foxon , phys .
b * 46 * , 7957 ( 1992 ) .
a. pinczuk , b. s. dennis , l. n. pfeiffer , and k. west , phys .
lett . * 70 * , 3983 ( 1993 ) .
s. m. girvin , a. h. macdonald , and p. m. platzman , phys . rev .
* 33 * , 2481 ( 1986 ) .
a. h. macdonald , e. h. rezayi , and d. keller , phys .
. lett . * 68 * , 1939 ( 1992 ) .
wang , j. l. birman , and z .- b .
su , phys .
. lett . * 68 * , 1605 ( 1992 ) .
v. m. apalkov and e. i. rashba , pisma zh .
. fiz . * 55 * , 38 ( 1992 ) [ jetp lett .
* 55 * , 37 ( 1992 ) ] . e. i. rashba and m. e. portnoi , phys .
70 * , 3315 ( 1993 ) . x. m. chen and j. j. quinn , phys .
b * 50 * , 2354 ( 1994 ) .
i. v. kukushkin , r. j. haug , k. von klitzing , k. eberl , and k. ttemeyer , phys .
b * 50 * , 11 259 ( 1994 ) . e. i. rashba and v. m. apalkov , in _ optical phenomena in semiconductor structures of reduced dimensions _ , edited by d. j. lookwood and a. pinczuk , nato advanced study institute , series b : physics , vol .
248 ( kluver , dordrecht , 1993 ) , p. 63 .
d. heiman , f. plentz , a. pinczuk , l. n. pfeiffer , and k. w. west , bull .
. soc . * 41 * , no . 1 , 76 ( 1996 ) . v. m. apalkov , f. g. pikus , and e. i. rashba , phys .
b , * 52 * , 6111 ( 1995 ) . v. m. apalkov and e. i. rashba , solid state commun . *
95 * , 421 ( 1995 ) .
x. m. chen and j. j. quinn , phys .
lett . * 70 * , 2130 ( 1993 ) .
e. h. rezayi , in _ optical phenomena in semiconductor structures of reduced dimensions _
( ref . ) , p. 79 . v.
m. apalkov and e. i. rashba , phys .
b * 46 * , 1628 ( 1992 ) .
f. d. m. haldane , phys .
lett . * 51 * , 605 ( 1983 ) .
m. e. portnoi and e. i. rashba , modern phys .
b * 9 * , 123 ( 1995 ) .
l. p. gorkov and i. e. dzyaloshinskii , zh .
. fiz . * 53 * , 717 ( 1967 ) [ sov .
jetp * 26 * , 449 ( 1968 ) ] .
e. i. rashba and v. m. apalkov , _ conference book of the 11th international conference on high magnetic fields in semiconductor physics _
( mit , cambridge , 1994 ) , p. 648 . j. zang and j. l. birman , phys .
b * 51 * , 5574 ( 1995 ) . x. m. chen and j. j. quinn , phys . rev .
b * 51 * , 5578 ( 1995 ) .
s. he , x .- c .
xie , and f .- c .
zhang , phys .
lett . * 68 * , 3460 ( 1992 ) .
g. dav and j. k. jain , phys .
lett . * 69 * , 2843 ( 1992 ) .
n. r. cooper and d. b. chklovskii , preprint cond - mat/9607002 .
r. k. kamilla , x. g. wu , and j. k. jain , solid state commun .
* 99 * , 289 ( 1996 ) .
wu , phys .
lett . * 52 * , 2103 ( 1984 ) ; * 53 * , 111 ( 1984 ) .
g. birkhoff and s. mac lane , _ a survey of modern algebra _ ( macmillan , new york , 1977 ) .
d. endesfelder and o. terzidis , z. phys .
b - condensed matter * 87 * , 51 ( 1992 ) .
t. chakraborty and p. pietilinen , the quantum hall effect ( springer - verlag , berlin , 1995 ) .
p. bran and r. morf , phys .
b * 43 * , 12 654 ( 1991 ) and references therein .
u. girlich and m. hellmund , phys .
b * 49 * , 17 488 ( 1994 ) and references therein . i. s. gradshteyn and i. m. ryzhik , _ table of integrals , series and products _ ( academic , new york , 1980 ) . a. h. macdonald and e. h. rezayi , phys . rev . *
42 * , 3224 ( 1990 ) ; v. m. apalkov and e. i. rashba , pisma zh .
. fiz . * 54 * , 155 ( 1991 ) [ jetf lett . *
54 * , 155 ( 1991 ) ] ; n. sivan and s. levit , phys .
b * 46 * , 3953 ( 1992 ) .
f. c. zhang , v. z. vulovic , y. guo , and s. das sarma , phys .
b * 32 * , 6920 ( 1985 ) ; e. h. rezayi and f. d. m. haldane , _ ibid . _ , p. 6924 . v.
m. apalkov and e. i. rashba , unpublished .
t. v. tatarinova , e. i. rashba , and a. l. efros , phys .
b * 50 * , 17 349 ( 1994 ) . j. k. jain , adv . in phys . *
41 * , 105 ( 1992 ) .
f. d. m. haldane , phys .
lett . * 67 * , 937 ( 1991 ) .
wu , phys .
lett . * 73 * , 922 ( 1994 ) .
m. d. johnson and g. s. canright , phys .
b * 49 * , 2947 ( 1994 ) .
j. yang , phys .
b * 50 * , 11 196 ( 1994 ) . w .-
su , y .- s .
wu , and j. yang , preprint cond - mat/9608075 .
lee and x .- g .
wen , phys . rev .
b * 49 * , 11 066 ( 1994 ) .
t. einarsson , s. l. sondhi , s. m. girvin , and d. p. arovas , nucl .
b * 441 * , 515 ( 1995 ) . | elementary excitations of incompressible quantum liquids ( iql s ) are anyons , i.e. , quasiparticles carrying fractional charges and obeying fractional statistics . to find out how the properties of these exotic quasiparticles manifest themselves in the optical spectra ,
we have developed the anyon exciton model ( aem ) and compared the results with the finite - size data for excitons of @xmath0 and @xmath1 iql s .
the model considers an exciton as a neutral composite consisting of three quasielectrons and a single hole .
the aem works well when the separation between electron and hole confinement planes , @xmath2 , obeys the condition @xmath3 , where @xmath4 is the magnetic length . in the framework of the aem an exciton possesses momentum @xmath5 and two internal quantum numbers ,
one of which can be chosen as the angular momentum , @xmath6 , of the @xmath7 state .
charge fractionalization manifests itself in striking differences between the properties of anyon excitons and ordinary magnetoexcitons .
existence of the internal degrees of freedom results in the multiple branch energy spectrum , crater - like electron density shape and 120@xmath8 density correlations for @xmath7 excitons , and the splitting of the electron shell into bunches for @xmath9 excitons . for @xmath3
the bottom states obey the superselection rule @xmath10 , @xmath11 are integers , and all of them are hard core states . for @xmath12
there is one - to - one correspondence between the low - energy spectra found for the aem and the many - electron exciton spectra of the @xmath1 iql , whereas some states are absent from the many - electron spectra of the @xmath0 iql .
we argue that this striking difference in the spectra originates from the different populational statistics of the quasielectrons of charge conjugate iql s and show that the proper account of the statistical requirements eliminates excessive states from the spectrum . apparently , this phenomenon is the first manifestation of the exclusion statistics in the anyon bound states . |
the discovery of broad emission lines in the polarized spectrum of a number of seyfert ( sey ) 2 galaxies has provided the basis for a unified model of sey galaxies in which the main discriminating parameter is the inclination of our line of sight with respect to an obscuring torus surrounding the source . in the simple version of this model
, one then expects the direct continuum of type 1 and 2 objects to be identical except for the effects of obscuration by the torus material .
the x - ray spectra of type 1 objects are relatively well known and consist of a power law of index @xmath12 , a cut - off at around 200 kev plus a compton reflection component with r ( the solid angle in units of 2@xmath8 , subtended by the reflecting material ) @xmath1 0.6 - 1 and a fe k@xmath9 feature ( perola et al .
2002 , matt 2001 , pounds and reeves 2002 ) .
conversely , the x - ray spectra of type 2 are much less well defined due to the extra complexity of absorption : constraints on individual spectra are rather poor due to the limited photon statistics and provide a greater variety of spectral components ( matt et al .
2000 , malaguti et al . 1999 ,
turner et al . 2000 ) .
one way to better constrain the x - ray spectral properties is to consider average spectra .
when this is done , the spectra of type 2 objects are found to be substantially harder than that of type 1 , which implies a lower energy cut - off ( zdziarski et al .
1995 , 2000 ) .
however , this intrinsic difference in the spectrum of sey 1 and 2 s can not be explained by the simplest version of the unified agn model , providing its strongest challenge yet .
here , we study the average spectra of sey 1 and 2 observed by bepposax over the energy band 3 - 200 kev , i.e. the first simultaneous broad band spectra with coverage above 10 kev . the purpose of this work is to impose the tightest possible constraints on both spectral shapes in order to check if their intrinsic continuum is indeed different and to discuss our findings in the framework of the unified theory .
the sample used for this work consists of 9 sey 1 ( fairall9 , 3c120 , ngc3783 , ngc4593 , mcg-6 - 30 - 15 , ngc5548 , mkn509 , eso141-g55 , ngc7469 ) and 13 sey 2 ( ngc526a , ngc2992 , ngc5506 , eso103-g35 , ngc7172 , ngc7314 , ngc7582 , mcg-5 - 23 - 16 , ngc1365 , ngc2110 , ngc4258 , ngc5252 , ngc5674 ) galaxies .
the strongest bepposax detections of sey ( such as ngc4151 and ic4329a ) were excluded so that single sources would not significantly bias our results .
we also excluded from our sey 2 samples sources with column density in excess of 10@xmath10 atoms @xmath5 in order to maintain a certain homogeneity in the absorption properties .
the count spectra of the individual observations were added with weights corresponding to the length of time of each observation and the sum was simply normalized by the total number of sources in each subset . only mecs ( 3 - 10 kev ) and
pds data ( 15 - 200 kev ) were considered in order to avoid the complexity due to the soft excess particularly in sey 2 .
standard data reduction was performed using the software package `` saxdas '' ( see http://www.sdc.asi.it/software and the cookbook for bepposax nfi spectral analysis , fiore , guainazzi & grandi 1998 ) .
data were linearized and cleaned for earth occultation periods and periods of high particle background ( satellite passages through the south atlantic anomaly ) .
data have been accumulated for earth elevation angles @xmath11 degrees and magnetic cut - off rigidity @xmath12 . for the pds data we adopted a fine energy and temperature dependent rise time selection , which decreases the pds background by @xmath13 .
this improves the signal to noise ratio of faint sources by about 1.5 ( frontera et al . 1997 , fiore , guainazzi & grandi 1998 ) .
data from the four pds units and the two mecs units were merged after equalization and single mecs and pds spectra were then accumulated .
the normalization constant , introduced to allow for a well known difference in the absolute cross - calibration between the two detectors , was left free to vary between 0.7 -0.95 ( fiore , guainazzi and grandi , 1998 ) .
all quoted errors refer to 90% confidence level for 1 interesting parameter ( @xmath14=2.7 ) .
first we fitted the 3 - 200 kev spectra with an absorbed power law plus a gaussian line ( model 1 in table 1 ) . as expected
, this simple model provides an unsatisfactory fit to the data mainly due to the presence of a strong reflection component and the possible presence of a high energy cut - off .
the iron line energy is compatible with reflection in cold material and the equivalent width ( ew ) is @xmath1 100 - 200 ev . in both classes of objects
the line width is slightly broad but this is likely due to a redshift effect as sources with different z are summed together : in particular 5 seyfert 1s and 3 seyfert 2s galaxies have redshift greater than 0.01 which is enough to produce a broadening of the line ; an important contribution to the line width could also be given by fairall 9 ( z=0.047 ) to the sey 1 average spectrum and by ngc5506 ( z=0.062 ) to the seyfert 2s average spectrum .
a detailed analysis of the iron line is however beyond the scope of the present paper and it is therefore postponed to a later work .
+ although the absorbed power law model points to an overall shape similar for both types of galaxies , the best fit photon index for sey 1 and 2 turns out to be significantly different : the sey 1s retain the canonical @xmath6 of @xmath11.9 while the sey 2s are flatter with @xmath15 1.75 . as expected sey 2 galaxies
are more absorbed than sey 1 with an average column density of @xmath1 4 @xmath3 10@xmath4 atoms @xmath5 , however some amount of absorption is also present in sey 1 galaxies with n@xmath2 @xmath16 10@xmath4 atoms @xmath5 .
a recent analysis of a sample of sey 1 with bepposax indicates that cold absorption of this amount is generally not present in this type of object ( n@xmath2 @xmath17 10@xmath18 atoms @xmath5 ) ; however a warm absorber seems to be a common feature of type 1 objects and is consistent with the presence of some absorption in our average spectrum ( perola et al . 2002 ) .
next we introduced a compton reflection component and a high energy cut - off via the pexrav model in xspec : this model provides an improvement in the fit which is significant at the 99.99% confidence level for 2 extra degrees of freedom for both classes ( table 1 model 2 ) .
the inclination assumed for the seyfert 1s is cos i=0.45 in model 2@xmath19 and 0.87 in model 2@xmath20 ; on the other hand since the sey 2s are most likely to be seen edge - on , we assumed in this case cos i=0.45 only
. furthermore , we assume the reflection medium to be close to neutral and the abundances used are those of anders and ebihara ( 1982 ) .
it is evident from table 1 that while the amount of reflection is consistent between the two types , the primary continuum is still significantly different : the average sey2 spectrum is substantially harder than that of sey 1 and the cut - off is at lower energies ( see table 1 ) . on the other hand ,
the average spectrum of sey 1s is perfectly canonical having @xmath6=1.9 , a reflection strength r @xmath1 0.6 ( or 1 depending on the inclination angle assumed ) and a cut - off energy at around 200 kev ; the iron line is confirmed at 6.4 kev and the equivalent width is 100 - 120 kev .
our results confirm the findings of perola et al .
( 2002 ) on a sample of seyfert 1 galaxies observed by bepposax and analyzed individually with the same model .
when their data are combined to get weighted mean values ( @[email protected] , [email protected] for cos@xmath22=0.9 and e@xmath23 130@xmath2120 kev ) , these are very similar to our best fit parameters of model 2@xmath20 in table 1 ; their average line energy is at [email protected] kev while the equivalent width is 122@xmath2116 again in agreement with our results .
also , if the iron line comes from the same medium as the compton reflecting continuum , one would expect a simple relationship between r , @xmath6 and ew , which for the average photon index of the sey 1 and the reflection value obtained for cos@xmath22=0.87 provides an ew value very similar to that observed ( see george and fabian 1991 results rescaled to the anders and grevesse ( 1993 ) abundances ) : therefore the observed ew is of the magnitude expected from the simultaneously measured strength of the reflection component .
the difference seen between sey 1 and 2 is however troublesome as it is not easily reconciled with the unified theory , which is however independently supported by other observational evidences . + it is well known that there is an inter - dependence of the parameters in the pexrav model , whereby increasing the reflection and/or the high energy cut - off leads to a steeper power law . to test this possibility ,
we have imposed a photon index of 1.9 ( equal to that obtained for the sey 1 in the cos i=0.45 configuration ) on the sey 2 and measured the contours of reflection strength against cut - off energy for this case : figure 1 shows the results compared to the sey 1 case while model 2@xmath24 in table 1 provides the best fit parameters .
the cut - off energy turns out to be similar to that measured in type 1 objects but r is now much higher ( @xmath13 ) .
a compton reflection component as strong as this is unusual for compton thin sey 2 galaxies ( bassani et al .
1999 ) which characterize our sample . also , following the same reasoning as for type 1 objects , the measured iron line ew is too low for the observed reflection strength especially if a contribution of the order of 30 - 40 ev has to be taken into account for the line component originating by transmission through a torus of column density of @xmath1 4 @xmath3 10@xmath4 @xmath5 .
a reflection component as strong as that measured is , however , still possible if the hard x - ray emission is isotropic and/or there is a time lag in the response of the reflection flux to variations of the source intensity ( cappi et al .
1996 , weaver et al .
+ a model independent way to verify the presence of a stronger reflection in type 2 with respect to type 1 objects is to measure the sey 2 to sey 1 pds ratio and search for the presence of a bump in the pds data .
the sey2 to sey1 ratio is shown in figure 2 , which clearly excludes the presence of stronger reflection in type 2 objects .
indeed the average spectra of sey 1s and sey 2s appear to be very similar in the high energy ( pds ) band , but they are different in the mecs range where effect of absorption takes place .
+ the ratio shown in figure 2 was produced by first summing the background - subtracted data for the two types of sources individually , normalizing each for the number of sources in each sample and then dividing in order to obtain the ratio .
the data were then rebinned in order to have a minimum ( 3 sigma ) detection in each energy bin throughout the ratio .
a visual inspection of the distribution of spectra for each sample ( figures 3 ) confirms that neither the sey 1 or sey 2 groups are dominated by one source .
the origin of the negative counts which can be seen in the individual spectra is due to the fact that we are working in a high energy regime where we are background dominated and statistical variations in background can be higher than the source flux itself .
this is confirmed by the observation that the maximum negative flux corresponds to a significance of only 1.5@xmath25 , thus ruling out the possibility that systematics in the background subtraction are responsible .
the sey 1 sample is much more homogeneous than the sey 2 group , which would be expected due to the wider range of geometric orientations possible for the latter type .
it can be seen that there is no discernible excess in the 20 - 100 kev band , thus excluding the presence of stronger reflection in sey 2 s than in sey 1 s ; if we restrict the analysis to the 40 - 70 kev band where excess counts are visible by eye , the significance is still too low ( 2.3@xmath25 ) to claim the presence of an excess . + having excluded reflection as the possible cause of a flatter spectrum , the only other remaining way to steepen the sey 2 spectrum is via a more complex absorption than assumed . this is not an unreasonable assumption since within our sample there are objects with different column densities and summing their spectra together can artificially produce a complex absorption where the covering fraction depends on the distribution of fluxes .
we have therefore introduced in the sey 2 fit an extra absorber that partially covers the source ( model 3@xmath19 in table 1 ) .
the fit improves significantly at more than 99% confidence level for two extra degrees of freedom and provides average spectral parameters for the intrinsic continuum very similar to those measured in sey 1s : @xmath15 1.9 , r @xmath10.7 - 1.2 and e@xmath7 200 kev .
the two best fits give rise to two different mecs / pds normalization factors ( @xmath10.75 and @xmath10.83 ) both of which are within the range indicated in the cookbook of bepposax nfi spectral analysis ( fiore et al .
however , in order to further check the robustness of these results , we have tried to re - analyse the sey 2 data using the normalization found for the sey 1 s and vice versa . in both cases
the values of @xmath6 , r and cut - off are consistent with those found in the original analysis within 2@xmath25 .
the composite spectra of both types of seyferts are compared in figure 4 ; again from this figure is evident that the two spectra are very similar in the pds energy range , while differing considerable at lower energies .
we have measured with bepposax the average spectra of sey 1 galaxies confirming previous indications that typically these galaxies have a power law shape with photon index @xmath15 1.9 , a reflection component of strength [email protected] - 1 and a high energy cut - off which is now more clearly defined and located at around 200 kev .
sey 2s on the other hand , have an average spectrum which is substantially harder and with a lower energy cut - off than sey 1s .
our analysis excludes the possibility that stronger reflection in sey 2s is responsible for their harder spectrum ; instead the introduction of a complex absorber provides an average sey 2 s spectrum very similar to that of sey 1 s .
this complex absorption is not necessarily intrinsic to individual sources but is likely to be an artifact of summing spectra with different amounts of absorption .
however since complex absorption has been observed in some individual source spectra ( e.g. malaguti et al .
1999 , turner et al .
2000 ) we can not exclude a - priori its possible contribution to the average seyfert 2 s spectrum . in any case our result on the primary power law continuum argues strongly for a very similar production mechanism not only in both types of seyferts but in each individual object and is therefore also relevant for the synthesis of the x - ray background : sey 2 spectra can now be taken to be similar to their type 1 counterparts in the agn contribution estimates .
anders , e. , ebihara , m. 1982 , geochim .
acta 46 , 2363 anders , e. , grevesse , n. 1993 , geochim .
acta 53 , 197 bassani l. , cappi m. , malaguti g. , palumbo g.g.c .
, dadina m. et al . 1999 ,
mem s.a.it vol.70 - 1 , 65 cappi m. , mihari t. , matsuoka m. , hayashida k. , weaver k.a . , and otani c. 1996 , apj 458 , 149 fiore f. , guainazzi m. , grandi p. 1998 , _
beppo_sax cookbook frontera f. , costa e. , dal fiume d. , et al .
1997 a&as , 122 , 357 george , i. m. and fabian , a. c. 1991 , mnrass 249 , 352 malaguti , g. , bassani , l. , cappi , m. , comastri , a. , di cocco , g. , fabian , a. c. , et al . 1999 , a&a342l , 41 matt , g. , fabian , a. c. , guainazzi , m. , iwasawa , k. , bassani , l. , malaguti , g. 2000 , mnras , 318 , 173 matt , g. 2001 , x - ray astronomy , aip conference proceedings edited by n.e .
white et al . , vol .
599 , p. 209 perola , g. c. , matt , g. , cappi , m. , fiore , f. , et al .
2002 , 389 , 802 pounds , k.a .
and reeves , j.n .
2002 , astro - ph/0201436 , proceedings of the symposium on ` new visions of the x - ray universe in the xmm - newton and chandra era ' , estec , the netherlands turner , t. j. , perola , g. c. , fiore , f. , matt , g. , george , i. m. , piro , l. , bassani , l. 2000 , apj , 531 , 245 weaver k.a .
, nousek j. , yaqoob t. , mushotzky r.f . , makino f. , and otani c. 1996 , apj 458 ,
160 zdziarski , a. , johnson , w. n. , done , c. , smith d. , mcnaron - brown k. 1995 , apj , 438l , 63 zdziarski , a. , poutanen , j. , johnson , w. n. 2000 , apj , 542 , 703 llllcllll n@xmath2(1 ) @xmath3 10@xmath4 & 1.01@xmath26 & 0.93@xmath27 & 0.87@xmath28 & & 4.12@xmath29 & 3.67@xmath30 & 3.83@xmath31 & 3.16@xmath32 + n@xmath2(2 ) @xmath3 10@xmath4 & & & & & & & & 27@xmath33 + c@xmath34 & & & & & & & & 0.26@xmath35 + photon index & 1.85@xmath36 & 1.92@xmath35 & 1.88@xmath35 & & 1.74@xmath37 & 1.75@xmath38 & * 1.90 * & 1.89@xmath39 + e@xmath40 & & 216@xmath41 & 197@xmath42 & & & 128@xmath43 & 221@xmath44 & 226@xmath45 + r & & 1.25@xmath46 & 0.64@xmath47 & & & 1.19@xmath48 & 3.02@xmath49 & 0.96@xmath50 + e@xmath51 ( kev ) & 6.39@xmath52 & 6.37@xmath53 & 6.38@xmath52 & & 6.48@xmath54 & 6.47@xmath55 & 6.44@xmath35 & 6.46@xmath53 + @xmath56 ( kev ) & 0.26@xmath57 & 0.18@xmath58 & 0.22@xmath59 & & 0.25@xmath60 & 0.21@xmath61 & 0.10@xmath54 & 0.12@xmath62 + ew ( ev ) & 148@xmath63 & 106@xmath64 & 120@xmath65 & & 189@xmath66 & 145@xmath67 & 97@xmath68 & 103@xmath69 + @xmath70/dof & 257.41 ( 70 ) & 87.12 ( 68 ) & 90.48 ( 68 ) & & 242.62 ( 70 ) & 101.92 ( 68 ) & 108.54 ( 69 ) & 90.4 ( 66 ) + @xmath71 & 3.68 & 1.25 & 1.33 & & 3.47 & 1.50 & 1.57 & 1.37 + | we have studied the average 3 - 200 kev spectra of seyfert galaxies of type 1 and 2 , using data obtained with bepposax .
the average seyfert 1 spectrum is well - fitted by a power law continuum with photon spectral index @xmath0 1.9 , a compton reflection component [email protected] - 1 ( depending on the inclination angle between the line of sight and the reflecting material ) and a high - energy cutoff at around 200 kev ; there is also an iron line at 6.4 kev characterized by an equivalent width of 120 ev .
seyfert 2 s on the other hand show stronger neutral absorption ( n@xmath2=3 - 4 @xmath3 10@xmath4 atoms @xmath5 ) as expected but are also characterized by an x - ray power law which is substantially harder ( @xmath6 @xmath1 1.75 ) and with a cut - off at lower energies ( e@xmath7130 kev ) ; the iron line parameters are instead substantially similar to those measured in type 1 objects .
there are only two possible solutions to this problem : to assume more reflection in seyfert 2 galaxies than observed in seyfert 1 or more complex absorption than estimated in the first instance . the first possibility is ruled out by the seyfert 2 to seyfert 1 ratio while the second provides an average seyfert 2 intrinsic spectrum very similar to that of the seyfert 1 .
the extra absorber is likely an artifact due to summing spectra with different amounts of absorption , althought we can not exclude its presence in at least some individual sources . our result argues strongly for a very similar central engine in both type of galaxies as expected under the unified theory . |
random fluctuations in nonlinear systems in engineering and science are often non - gaussian @xcite . for instance , it has been argued that diffusion by geophysical turbulence @xcite corresponds to a series of pauses " , when the particle is trapped by a coherent structure , and flights " or jumps " or other extreme events , when the particle moves in the jet flow .
paleoclimatic data @xcite also indicate such irregular processes .
there are also experimental demonstrations of lvy flights in foraging theory and rapid geographical spread of emergent infectious disease .
humphries et . al .
@xcite used gps to track the wandering black bowed albatrosses around an island in southern indian ocean to study the movement patterns of searching food .
they found that by fitting the data of the movement steps , the movement patterns obeys the power - law property with power parameter @xmath1 . to get the data set of human mobility that covers all length scales , brockmann @xcite collected data by online bill trackers , which give successive spatial - temporal trajectories with a very high resolution .
when fitting the data of probability of bill traveling at certain distances within a short period of time ( less than one week ) , he found power - law distribution property with power parameter @xmath2 , and observed that @xmath0stable lvy motions are strikingly similar to practical data of human influenza .
lvy motions are thought to be appropriate models for a class of important non - gaussian processes with jumps @xcite .
recall that a lvy motion @xmath3 , or @xmath4 , is a stochastic process with stationary and independent increments .
that is , for any @xmath5 with @xmath6 , the distribution of @xmath7 only depends on @xmath8 , and for any @xmath9 , the random variables @xmath10 , @xmath11 , are independent .
a lvy motion has a version whose sample paths are almost surely right continuous with left limits .
stochastic differential equations ( sdes ) with non - gaussian lvy noises have attracted much attention recently @xcite . to be specific ,
let us consider the following n - dimensional _ state system _ : @xmath12 where @xmath13 is a vector field ( also called a drift ) , and @xmath14 is a symmetric @xmath0stable lvy motion ( @xmath15 ) , defined in a probability space @xmath16 .
assume that we have either + ( i ) a discrete time m - dimensional _ observation system _ : @xmath17 where @xmath18 is a white sequence of gaussian random variables , i.e. @xmath18 s are mutually independent standard normal random variables , and @xmath19 is a sequence of nonnegative numbers ; + or + ( ii ) a continuous time m - dimensional _ observation system _
: @xmath20 where @xmath21 is a given vector field and @xmath22 is a brownian motion . in the present paper ,
we estimate system states , with help of observations , and in particular , we try to capture transitions between metastable states by examining most probable paths for system states .
this paper is organized as follows .
we consider state estimates with discrete time and continuous time observations in sections 2 and 3 , respectively .
to demonstrate our ideas , we consider the following scalar sde driven by a symmetric @xmath23-stable lvy motion @xmath24 together with observations @xmath25 are taken at discrete time instants @xmath26 as follows @xmath27 where @xmath18 is a white sequence of gaussian random variables . for @xmath28 ,
a symmetric @xmath23-stable lvy motion @xmath29 has the generating triplet @xmath30 , where the jump measure @xmath31 with @xmath32 given by the formula @xmath33 . for more information see @xcite
thus , the generator for the solution process @xmath34 in is @xmath35 \ ; \nu_\alpha({\rm d}y).\end{aligned}\ ] ] in fact , this linear operator @xmath36 is a nonlocal laplace operator ( ( * ? ? ?
7 ) ) , denoted also by @xmath37 , for @xmath38 .
the generator @xmath36 carries crucial information about the system state @xmath39 , and hence will be useful in our investigation of state estimation . also note that the non - gaussianity of the lvy noise manifests as nonlocality ( an integral term ) in the generator .
the adjoint operator for the generator @xmath36 is @xmath40 \ ; \nu_\alpha(dy).\end{aligned}\ ] ] the fokker - planck equation for the sde ( [ state1 ] ) is ( @xcite ) : @xmath41 \ ; \nu_\alpha(dy).\end{aligned}\ ] ] denote @xmath42 .
similarly as in @xcite , we have the following theorem which determines the time evolution of the conditional probability density function @xmath43 . for convenience , we often write @xmath44 for @xmath43 .
( conditional density for continuous - discrete problems ) .
let system ( [ state1 ] ) satisfy the hypotheses that @xmath13 is lipschitz in space and the initial state @xmath45 , with the property @xmath46 , is independent of @xmath47\}$ ] .
suppose that the prior density @xmath48 for([state1 ] ) exists and is once continuously differentiable with respect to @xmath49 and twice with respect to @xmath50 .
let @xmath21 be continuous in both arguments and bounded for each @xmath26 with probability @xmath51 .
+ then , between observations , the conditional density @xmath52 satisfies the fokker - planck equation @xmath53 where @xmath54 is the operator in ( [ ccdd ] ) . at an observation @xmath55 ,
the conditional density satisfies the following difference equation @xmath56 where @xmath57 is @xmath58^t r_k^{-1}[y_k - h(x , t_k)]\}.\ ] ] the conditional density in the absence of observation , satisfies the fokker - planck equation .
therefore , between observations , conditional density @xmath52 satisfies the fokker - planck equation ( [ kolmog ] ) .
thus , it remains to determine the relationship between @xmath59 and @xmath60 since @xmath61 , we have by bayes rule @xmath62 now , since the noise @xmath63 is white , @xmath64 similarly , we compute @xmath65 therefore , @xmath66 this completes the proof .
this theorem provides the foundation for computing conditional density for system state of sde ( [ state1 ] ) , under discrete time observations .
define @xmath67 .
+ this provides the most probable orbit ( @xcite ) starting at @xmath68 .
these most probable orbits are the maximal likely orbits for a dynamical system under noisy fluctuations .
let us consider an example .
[ discrete ] let us consider a scalar system with state equation @xmath69 the discrete - time scalar observation is : @xmath70 with @xmath71 , and @xmath72 . in the absence of lvy noise , this system has two stable states : @xmath73 and @xmath74 . when the noise kicks in , these two states are no longer fixed .
the random system evolution near these two states , together with possible transitions between them , is sometimes called a metastable phenomenon @xcite . for convenience ,
we call @xmath73 and @xmath74 ( and random motions nearby ) metastable states .
the corresponding nonlocal fokker - planck equation is computed on @xmath75 with a finite difference scheme ( @xcite ) under the natural boundary condition .
space stepsize @xmath76 and time stepsize @xmath77 .
the initial probability density is taken either as a gaussian distribution or a uniform distribution . in figures
[ discrete1 ] and [ discrete2 ] , we show the conditional density @xmath44 , together with the corresponding most probable orbit ( taken as the state estimation for @xmath39 ) , together with observations @xmath25 and a true state path @xmath39 .
the initial density is either gaussian or uniform , but centered at the metastable state @xmath73 .
notice that the estimated state captures the transitions from @xmath73 to @xmath74 and then back to @xmath73 , during the time period @xmath78 .
conditional probability and state estimation when @xmath79 .
the initial probability density is gaussian centered at @xmath73 .
online version : the red curve in the right panel is the estimate state orbit ( i.e. , most probable orbit ) . , title="fig:",width=234,height=188 ] conditional probability and state estimation when @xmath79 .
the initial probability density is gaussian centered at @xmath73 .
online version : the red curve in the right panel is the estimate state orbit ( i.e. , most probable orbit ) . , title="fig:",width=234,height=188 ] conditional probability and state estimation when @xmath80 .
the initial probability density is uniform on @xmath81 .
online version : the red curve in the right panel is the estimate state orbit ( i.e. , most probable orbit ) .
, title="fig:",width=234,height=188 ] conditional probability and state estimation when @xmath80 .
the initial probability density is uniform on @xmath81 .
online version : the red curve in the right panel is the estimate state orbit ( i.e. , most probable orbit ) . , title="fig:",width=234,height=188 ]
we consider the following scalar _ state system _ with a symmetric @xmath0stable lvy motion @xmath82 together with a continuous time scalar _ observation system _ : @xmath83 where @xmath21 is a given vector field and @xmath22 is a brownian motion
. where @xmath54 is the adjoint operator of the generator @xmath36 : @xmath86 \ ; \nu_\alpha({\rm d}y),\end{aligned}\ ] ] and @xmath87 is the initial density of @xmath39 ( say a uniform distribution near the metastable state @xmath73 ) . the zakai equation may be numerically solved with a finite difference method based on @xcite together with a discretization of the noisy term at the current space - time point and @xmath88 .
the initial probability density is taken either as a gaussian distribution or a uniform distribution .
for other numerical methods , see , for example , @xcite . the conditional density @xmath84 provides information for the system evolution . with the observation , we can infer possible transitions from the metastable state @xmath73 to the metastable state @xmath74 , within a time range @xmath93 .
if the system starts with a probability distribution @xmath94 near the metastable state @xmath95 , then the conditional density @xmath84 helps us to infer whether the system will get near the other metastable state @xmath74 , and vice versa .
this may be achieved by examining the most probable orbits for the system , under the observation .
define @xmath67 .
+ this provides the most probable orbit ( @xcite ) starting at @xmath68 .
we take this as our state estimation for @xmath39 , as in @xcite .
[ continuous ] let us consider the following scalar sde state equation with a symmetric @xmath0stable lvy motion : @xmath96 the scalar observation equation is given by @xmath97 when noise is absent , the state system has two stable states : @xmath73 and @xmath74 . in figures ( [ continuous1 ] ) - ( [ continuous2 ] ) , we show the conditional density @xmath44 , together with the corresponding most probable orbit ( taken as the state estimation for @xmath39 ) , together with a true state path @xmath39 .
the initial density is either gaussian or uniform , but centered at the metastable state @xmath73 .
notice that the estimated state captures the transitions from @xmath73 to @xmath74 , then back to @xmath73 , and finally to @xmath74 , during the time period @xmath98 .
conditional probability ( left ) and state estimation ( right ) when @xmath79 .
the initial probability density is gaussian centered at @xmath73.,title="fig:",width=234,height=188 ] conditional probability ( left ) and state estimation ( right ) when @xmath79 .
the initial probability density is gaussian centered at @xmath73.,title="fig:",width=234,height=188 ] conditional probability ( left ) and state estimation ( right ) when @xmath79 .
the initial probability density is uniform on @xmath99.,title="fig:",width=234,height=188 ] conditional probability ( left ) and state estimation ( right ) when @xmath79 .
the initial probability density is uniform on @xmath99.,title="fig:",width=234,height=188 ] b. grigelionis and r. mikulevicius , nonlinear filtering equations for stochastic processes with jumps . in _ the oxford handbook of nonlinear filtering _
, d. crisan and b. l. rozovskii ( eds . ) , oxford university press , p. 95 - 128 , 2011 .
s. popa and s. s. sritharan , nonlinear filtering of ito - levy stochastic differential equations with continuous observations .
_ communications on stochastic analysis _ , vol .
3 , ( 2009 ) , pp . 313 - 330 .
d. schertzer , m. larcheveque , j. duan , v. yanovsky and s. lovejoy , fractional fokker
planck equation for nonlinear stochastic differential equations driven by non - gaussian lvy stable noises .
_ j. math .
_ , * 42 * ( 2001 ) , 200 - 212 .
w. a. woyczynski , lvy processes in the physical sciences . in _
lvy processes : theory and applications _ , o. e. barndorff - nielsen , t. mikosch and s. i. resnick ( eds . ) , 241 - 266 , birkhuser , boston , 2001 . | a goal of data assimilation is to infer stochastic dynamical behaviors with available observations .
we consider transition phenomena between metastable states for a stochastic system with ( non - gaussian ) @xmath0stable lvy noise .
with either discrete time or continuous time observations , we infer such transitions by computing the corresponding nonlocal zakai equation ( and its discrete time counterpart ) and examining the most probable orbits for the state system .
examples are presented to demonstrate this approach .
* short title : * transitions in non - gaussian stochastic systems + * key words : * nonlocal zakai equation ; nonlocal laplace operator ; non - gaussian noise ; transitions between metastable states ; most probable orbits * pacs * ( 2010 ) : 05.40.ca , 02.50.fz , 05.40.fb , 05.40.jc |
recently it was known that the large difference between the planck scale @xmath7 gev and the electroweak scale @xmath8 gev ( so - called the hierarchy problem ) can be produced by introducing the compact extra dimension(s ) .
there are two models with extra dimension(s ) .
one is the add model ( large extra - dimension model ) @xcite , where the extra dimensions are below @xmath9 ( @xmath10 ) and the extra submanifold is at least two dimensional .
although the add model resolves the hierarchy problem , it causes another hierarchy problem between the electroweak scale and the radius of the extra dimensions .
this comes from the power law behavior of the hierarchy , that is , @xmath11 , where @xmath12 and @xmath13 are the typical radius and the dimension of the extra submanifold , respectively .
the other is the randall - sundrum ( rs ) model ( small extra - dimension model ) @xcite , where the classical rs metric is @xmath14 and the extra manifold is a one - dimensional orbifold @xmath15 .
@xmath16 is smaller than the 5-dimensional planck mass @xmath17 gev , @xmath18 is the interbrane distance , and the hidden and visible branes are at @xmath19 and @xmath20 , respectively .
the electroweak scale can be derived from the planck scale via @xmath21 , where @xmath22 is the geometrical warp factor appearing in the metric of our brane .
the unwanted extra hierarchy problem does not exist due to the exponential dependence of the hierarchy on the radius @xmath23 unlike the add theory .
when a kind of modulus stabilization with a bulk scalar is added to the rs theory ( since it is impossible to determine the value of the interbrane distance @xmath24 from the dynamics of the original " rs model ) and a kind of dimensional reduction is done , the modulus field ( radion ) couples to the standard model ( sm ) fields like @xmath25 , where @xmath26 is the physical field of the radion , @xmath6 is a dimensionful coupling constant of order @xmath27 , and @xmath28 is the trace of the energy - momentum tensor of the sm fields .
the magnitudes of @xmath6 and the radion mass @xmath0 are phenomenologically important , since they are related to decay rates and cross sections .
there are two kinds of considerations on these parameters until now .
one is experimental , and the other is theoretical .
the experimental consideration is related to the experimentally measured quantities , for example , the decay rates and production cross sections of radion , and the radion loop corrections to the weak mixing angle .
it is noteworthy that the branching ratio of the radion into two gluons is the largest for @xmath29 gev , although it comes from the trace anomaly at the one - loop level .
the recent l3 results on the sm higgs search excluded @xmath30 gev for @xmath31 gev @xcite .
the constraint from the weak mixing angle with errors of order @xmath32 can exclude @xmath33 gev for @xmath34 gev @xcite .
and the lower bound of the radion mass is about @xmath35 gev from limits of neutrino phenomenology and tasso and cleo on inclusive b decays @xcite .
the theoretical one is on the perturbative unitarity .
the perturbative unitarity for @xmath36 ( @xmath37 : sm higgs boson ) is broken for relatively small @xmath38 gev for @xmath39 gev @xcite , and the unitarity bound in the @xmath40 process is @xmath41 gev at @xmath42 tev @xcite . in this paper , we have constrained the two parameters up to one - loop level in terms of two conditions , warp factor @xmath43 and @xmath4 gev ( @xmath44 ) .
the constraint , @xmath45 , was imposed to solve the hierarchy problem @xcite . from the above conditions , the one - loop allowed bounds of @xmath5 and @xmath46 are @xmath47 and @xmath2 , which can be stronger than the previous constraints except the lower bound of radion mass .
it is noteworthy that the one - loop upper bound of the radion mass @xmath5 is about five times larger than the tree - level one of about 49 gev , and still smaller than the masses of the low - lying kaluza - klein ( kk ) modes .
this paper is organized as follows : in section ii , we discuss what fields give important contributions to the effective potential , and analyze the contributions of the scalar and kk mode sectors and those of the fermion and gauge sectors . in section iii , we find the bounds on @xmath5 and @xmath46 from @xmath45 and @xmath4 gev .
the conclusions are in section iv .
when the hierarchy problem is solved with only the one - loop corrections of bulk fields , they generate a very light radion @xcite , which is excluded by experiments of neutrino phenomenology and @xmath48 decays @xcite .
therefore , the stabilization mechanism at the tree level is required .
the mechanism proposed by goldberger and wise @xcite can be a promising one , because they stabilized the modulus without any severe fine - tuning of the parameters in the full theory . in the goldberger - wise stabilization mechanism
, there is a bulk scalar field @xmath49 which has large quartic self - interactions on the hidden and visible branes alone , and a @xmath50-dependent @xmath51 @xmath52 . after replacing the field @xmath49 in the original lagrangian with its @xmath51 @xmath52 and integrating the lagrangian over @xmath50
, we have the modulus stabilizing potential .
the typical energy of this mode is the expectation value of an operator @xmath53 to @xmath52 of order @xmath54 , which makes the effective gravitational coupling constant @xmath55 of the bulk scalar about @xmath56 by the 5-dimensional planck suppression .
thus the quantum correction of the 5-dimensional graviton to the bulk scalar @xmath51 @xmath52 is unimportant in the full theory . and
the quantum corrections from the brane self - interactions of @xmath57 can shift the @xmath51s of the bulk scalar at the branes , @xmath58 and @xmath59 , where @xmath60 and @xmath61 .
these corrections can change only the coefficients of the radion potential and be equivalent to rescaling of the coefficients .
the 4-dimensional graviton coupling to the radion is @xmath62 , where @xmath63 is the 4-dimensional ricci scalar .
when the radion @xmath64 is replaced with @xmath65 , where @xmath66 is the @xmath51 of @xmath64 , the one - graviton coupling constant to the physical field of the radion @xmath26 is planck - suppressed , @xmath67 , since @xmath63 has linear terms of the graviton field @xmath68 in the weak field approximation .
so we can neglect the quantum effects of the 4-dimensional graviton .
there are low - lying kk modes of the bulk scalar and the bulk graviton with masses of @xmath35 tev and non - renormalizable couplings suppressed by tev scale @xcite .
the effects of these low - lying kk modes on the electroweak scale physics are absorbed into the renormalizations of the planck and tev brane tension terms @xmath69 @xcite . since the radion has highly non - renormalizable self - interactions such as @xmath70 in eq .
( [ vphi ] ) , we treated the radion self - interactions only at the tree level .
for example , the non - renormalizable quantum gravity has a serious problem that the one - loop effective action on the mass - shell is dependent on the gauge fixing parameters @xcite , breaking the dewitt - kallosh theorem @xcite .
therefore , we consider the loop corrections of the kk modes of bulk fields and the sm particles to the radion effective potential in our brane .
but the one - loop corrections from the kk modes of the bulk scalar and graviton can contribute considerably to the radion potential as @xmath96 , where @xmath97 and @xmath98 are shifts from the classical planck and tev brane tensions @xmath99 @xcite respectively , belonging to the tension shifts @xmath100 and @xmath101 of @xmath102 in ref . @xcite . for small @xmath101 , the addition of @xmath103 to the potential @xmath70 in eq .
( [ vphi ] ) gives a minimum for large @xmath104 , and the adjustment of the tension shift @xmath100 on the planck brane can make the effective four - dimensional cosmological constant vanish . since the shift of the tev brane tension @xmath98 is absorbed into the renormalization of the tev brane tension term @xcite @xmath105 , it can not be obtained from calculation and can only be determined by a renormalization condition relating it to observable quantities @xcite . because the 4-dimensional cosmological constant of order @xmath106 @xcite is regarded as the minimum value of the effective potential @xmath107 of the rs theory , the value of the potential can be approximated as zero at the observed values of @xmath37 and @xmath64 , which should be a stable minimum point of the potential @xcite .
the two renormalization conditions are expressed as @xmath108 at the observed values of @xmath37 and @xmath64 . from these first and second conditions
, we can determine the shifts @xmath100 and @xmath98 respectively , if all the other parameters are known .
the size of the shift @xmath98 can be of order @xmath109 much smaller than @xmath110 .
it is noteworthy that the sign of the kk mode contribution is important in the vacuum stability since it can change the shape of potential .
if @xmath98 is fixed to be negative and @xmath111 increases , the potential becomes deeper and has more parameter points which give stable vacua .
but when the sign is reversed , the number of the allowed points gets smaller , and thus the negative @xmath98 can be favored in terms of stability .
in the randall - sundrum theory , the extra dimension must have a finite size about @xmath133 to solve the hierarchy problem , which gives the constraint , warp factor @xmath134 . and in the standard model
, the @xmath51 of the higgs field is about 246 gev .
this can give a constraint on the higgs @xmath51 of the effective potential in the rs model .
but when the quantum corrections of radion are added to the sm particles , the experimentally determined @xmath51 of the higgs field may be much different from the sm value , that is , the corrections may destroy the above constraint .
when we consider the quantum corrections of the radion to the masses of the sm particles using the lagrangian @xmath135 , the pole masses of leptons and heavy quarks , for example , are @xmath136 \right\ } , \end{aligned}\ ] ] where @xmath137 are the renormalized masses with the sm loop corrections alone , @xmath138 are the mass parameters in the renormalized lagrangian , @xmath88 is a renormalization scale , and @xmath139 . if @xmath140 , then the radion correction can be large so that the shift of the higgs @xmath51 from the sm one can be large .
( this argument is similarly applied to the gauge bosons and the higgs boson . )
so the condition @xmath141 gev can be spoilt for @xmath140 . but this small range is excluded by the constraint from the weak mixing angle that @xmath142 gev are excluded for @xmath143 gev @xcite .
therefore , it can be natural that the physical pole mass ( @xmath144 @xmath145 ) is almost equal to the sm mass ( @xmath144 @xmath137 ) .
this means that the higgs @xmath51 @xmath146 gev can be a physical constraint on our effective potential for the natural range of @xmath6 .
the radion corrections are typically smaller than @xmath147 of @xmath137 , so we can expect that the errors of the higgs @xmath51 from the radion contributions are smaller than @xmath147 .
the procedure to determine @xmath5 and @xmath46 is as follows .
first , we find a parameter point of ( @xmath148 ) which gives a stable vacuum satisfying the two constraints , warp factor @xmath149 and higgs @xmath51 @xmath150 gev .
note that @xmath98 is not considered for the tree - level potential .
of course , the point has to be in the region where perturbation is valid and @xmath132 is real .
next , the values of @xmath5 and @xmath46 are determined from that point by means of equations @xmath151 and @xmath152 . since when a minimum point is found , @xmath97 can be determined trivially via the first equation in eq .
( [ rencond ] ) , we do not include the parameter @xmath97 . because the rs metric @xmath153 is a solution of the classical " 5-dimensional einstein s equation , in order to maintain the classical metric we choose parameter ranges where
the loop effects of the quantum gravity can be neglected . for the tree level potential , the allowed region for the 5-dimensional planck mass of a typical size @xmath154 and
the higgs mass @xmath155 gev is @xmath156 and @xmath157 from fig .
[ tree1 ] , where @xmath152 and @xmath158 .
because of the exponential dependence of @xmath159 , the large change of @xmath160 corresponds to the very small change of @xmath161 , whose central value is about 36 ( @xmath162 ) .
this means that the range of the radion @xmath51 @xmath66 is very narrow . from fig .
[ lam - num ] , most of the allowed parameter points ( about 93 percent ) are concentrated around the line @xmath163 gev , and thus the point with a smaller or larger @xmath46 is rare .
therefore , the _ naturally _ allowed region can be much narrower in @xmath46 than the above allowed one .
for the one - loop effective potential , the allowed region for the 5-dimensional planck mass @xmath154 and the higgs mass @xmath155 gev is @xmath1 and @xmath2 from fig .
[ loop1 ] ( the central value of @xmath164 is still about 36 ) . from fig .
[ lam - num ] , almost all the data are focused at @xmath165 gev like the tree level case .
therefore , we can conclude that the _ naturally _ allowed regions are similar for the tree and one - loop cases , and there is a small shift of the central value of @xmath46 ( or equivalently @xmath66 ) due to the one - loop corrections .
when @xmath98 and other parameters are changed continuously , the allowed region in fig . [ loop1 ] can be broader in @xmath46 . from numerical analysis
, we have found that only the negative values of @xmath98 can be allowed . as discussed in section ii
, negative values of the tension shift @xmath98 produce many parameter points which were not allowed at the tree level , and considerable parts of these new points make the radion mass sufficiently larger than the tree - level upper bound of the mass .
it is phenomenologically noteworthy that the one - loop upper bound of the radion mass @xmath5 is rather larger than the tree - level one by about five times .
but the radion is still the first signal of the rs theory lighter than the lowest - lying kk mode with a mass of order @xmath166 , because the radion mass is smaller than about @xmath167 gev . and the branching ratios of the radion into gluon or @xmath168 boson pairs are dominant according to the mass @xmath5 .
the tree level perturbative unitarity can not be stronger than the two constraints , warp factor @xmath149 and higgs @xmath51 @xmath150 gev , in constraining @xmath0 and @xmath6 , since the unitarity bound in the @xmath36 process is @xmath169 at @xmath170 @xcite and the bound in the @xmath40 process is @xmath171 gev at @xmath42 tev @xcite . for the radion mass @xmath31 gev ,
the bounds from the sm higgs search @xcite are helpful , since @xmath172 gev can be excluded .
the limits from neutrino phenomenology and tasso and cleo on inclusive @xmath48 decays give lower bounds of the radion mass about 1 gev @xcite , which agree to our lower bound .
and the region @xmath173 gev for cutoff @xmath174 tev is excluded by the constraint from the weak mixing angle @xcite , since @xmath175 and @xmath150 gev make @xmath5 less than about 260 gev .
most of the previous constraints gives only lower bounds . only the unitary bound in the @xmath40 process gives the weak upper bound of @xmath5 .
compared with the above constraints , the @xmath175 and @xmath150 gev constraints can produce the bounds of @xmath5 and @xmath46 strongly , except the mass lower bound .
assuming the goldberger - wise mechanism in the randall - sundrum theory , we considered the allowed regions of @xmath0 and @xmath46 from two constraints , warp factor @xmath149 and higgs @xmath51 @xmath150 gev , by means of the tree level and one - loop level potentials .
the allowed regions are @xmath176 and @xmath177 for the tree level potential , and @xmath178 and @xmath179 for the one - loop level potential .
it is phenomenologically important that the one - loop allowed upper bound of radion mass @xmath5 is considerably larger than the tree - level one , but the radion is still the first experimental signature of the rs model , since it is lighter than the kk modes .
the mass of the radion is less than about 260 gev , and thus it decays into gluon or @xmath168 boson pairs dominantly . and its loop contributions are small due to the average value of the effective coupling constant @xmath180 ( the average effective fine structure constant is about @xmath181 ) .
the conditions of @xmath182 and @xmath150 gev can give the strong bounds of @xmath0 and @xmath6 .
+ * note added : * after completing this paper , we received an interesting paper by u. mahanta @xcite , who considered the one - loop contributions of the radion and the kk modes of graviton to the higgs potential .
we did not include the radion contribution due to the non - renormalizability in section ii . in ref .
@xcite , the one - loop contributions via some approximations are too small to cause any instability of the classical vacuum in the valid region of perturbation theory .
therefore , these contributions do not change our results significantly .
we would like to thank prof .
pyungwon ko for discussions .
this work is supported by the brain korea 21 project and grant no .
1999 - 2 - 111 - 002 - 5 from the interdisciplinary research program of the kosef .
kim , b. kyae and j.d .
park , hep - ph/0007008 . .
u. mahanta , hep - ph/0004128 .
j. garriga , o. pujol@xmath183s and t. tanaka , hep - th/0004109 . . .
s. ichinose , phys . lett . * 284b * , 234 ( 1992 ) ; nucl . phys . *
b395 * , 433 ( 1993 ) .
dewitt , phys .
162 * , 1195 ( 1967 ) ; . c. cs@xmath184ki , m.l .
graesser and g.d .
kribs , hep - th/0008151 .
riess et al . , astron .
j. * 116 * , 1009 ( 1998 ) ; p.m. garnavich et al .
, astrophys .
j. * 509 * , 74 ( 1998 ) .
m.e . peskin and d.v .
schroeder , an introduction to quantum field theory " ( addison - wesley publishing company , 1995 ) . ; . .
u. mahanta , hep - ph/0011148 . | assuming the goldberger - wise mechanism , we investigated the effective potential at the one - loop level in the randall - sundrum theory .
we found the lower and upper bounds of the radion mass @xmath0 and the radion - sm coupling constant , @xmath1 and @xmath2 for the one - loop level potential .
these bounds were determined from two constraints of @xmath3 and @xmath4 gev , which can produce strong bounds of @xmath5 and @xmath6 .
it is phenomenologically important that the one - loop allowed upper bound of the radion mass is about five times larger than the tree - level one , but the radion is still lighter than the kaluza - klein modes .
# 1#2#3#4#1 , phys
. lett . *
# 2b * , # 3 ( # 4 ) # 1#2#3#4#1 , nucl .
phys . *
b#2 * , # 3 ( # 4 ) # 1#2#3#4#1 , phys . rev .
* d#2 * , # 3 ( # 4 ) # 1#2#3#4#1 , phys . rev
. lett .
* # 2 * , # 3 ( # 4 ) # 1#2#3#4#1 , mod . phys . lett .
* a#2 * , # 3 ( # 4 ) # 1#2#3#4#1 , phys . rep . * # 2 * , # 3 ( # 4 ) # 1#2_#1_#2 # 1#2_#1_#2 /#1#1/ |
given an interaction lagrangian characterized by a dimensionless coupling constant @xmath1 , one can usually guess whether perturbation theory can be used by comparing @xmath2 with unity .
if it is much less than unity , there is a good chance that the perturbation technique could be used .
effective chiral lagrangian gives on the other hand a power series expansion in momenta and hence the above rule can not be applied .
it is important to set up a set of rules to test whether perturbation theory could be used .
chiral perturbation theory ( chpt ) @xcite is a well - defined perturbative procedure for the effective chiral lagrangian allowing one to calculate systematically low energy phenomenon involving soft pions .
it is now widely used to analyze low energy pion physics even in the presence of resonance as long as the energy region of interest is sufficiently far from the resonance . in this scheme , the unitarity relation is satisfied perturbatively order by order .
the standard procedure of testing chpt calculation of the pion form factor @xcite , which claims to support the perturbative scheme , is shown here to be unsatisfactory .
this is so because the calculable terms are extremely small , less than 1.5% of the uncalculable terms at an energy of 0.5 gev or lower whereas the experimental errors are of the order 10 - 15% .
this obvious fact escapes the attention of many due to the complexity of the calculations .
we show how this situation can be dealt with without asking for a new measurement of the pion form factor with a precision much better than 1.5% .
this can be done using dispersion relation which establishes in a model independent way a relation between the real and imaginary part of the amplitude . from these results
, we set up a procedure to test whether it is possible to use the perturbation technique for a given lagrangian .
because the vector pion form factor @xmath3 is an analytic function with a cut from @xmath4 to @xmath5 , the @xmath6 times subtracted dispersion relation for @xmath3 reads : @xmath7 where @xmath8 and , for our purpose , the series around the origin is considered .
because of the real analytic property of @xmath3 , it is real below @xmath4 . by taking the real part of this equation
, @xmath9 is related to the principal part of the dispersion integral involving the @xmath10 apart from the subtraction constants @xmath11 .
the polynomial on the r.h.s . of eq .
( [ eq : ff1 ] ) will be referred in the following as the subtraction constants and the last term on the r.h.s . as the dispersion integral ( di ) .
the evaluation of di as a funtion of @xmath12 will be done later . notice that @xmath13 is the coefficient of the taylor series expansion for @xmath3 , where @xmath14 is the nth derivative of @xmath3 evaluated at the origin .
the condition for eq .
( [ eq : ff1 ] ) to be valid was that , on the real positive s axis , the limit @xmath15 as @xmath16 .
the coefficient @xmath17 of the taylor s series is given by : @xmath18 where @xmath19 .
the meaning of this equation is clear : under the above stated assumption , not only the coefficient @xmath11 can be calculated but all other coefficients @xmath17 can also be calculated .
the larger the value of @xmath20 , the more sensitive is the value of @xmath17 to the low energy values of @xmath10 . in theoretical work such as in chpt approach , to be discussed later , the number of subtraction is such that to make the di converges .
the elastic unitarity relation for the pion form factor is @xmath21 where @xmath22 is the elastic p - wave pion pion phase shifts . below the inelastic threshold of @xmath23 where @xmath24 is the pion mass
, @xmath3 must have the phase of @xmath22 @xcite .
it is an experimental fact that below @xmath25 the inelastic effect is very small , hence , to a good approximation , the phase of @xmath3 is @xmath26 below this energy scale .
@xmath27 and @xmath28 where @xmath26 is the strong elastic p - wave @xmath29 phase shifts .
because the real and imaginary parts are related by dispersion relation , it is important to know accurately @xmath30 over a large energy region .
below 1.3 gev , @xmath30 can be determined accurately because the modulus of the vector form factor @xcite and the corresponding p - wave @xmath29 phase shifts are well measured @xcite except at very low energy .
it is possible to estimate the high energy contribution of the dispersion integral by fitting the asymptotic behavior of the form factor by the expression , @xmath31 where @xmath32 is the @xmath0 mass squared . using eq .
( [ eq : ieu ] ) and eq .
( [ eq : reu ] ) , @xmath30 and @xmath9 are determined directly from experimental data and are shown , respectively , in fig.1 and fig.2 .
in the following , for definiteness , one assumes @xmath33 as @xmath34 on the cut , i.e. @xmath3 does not grow as fast as a linear function of @xmath12 .
this assumption is a very mild one because theoretical models assume that the form factor vanishes at infinite energy as @xmath35 . in this case , one can write a once subtracted dispersion relation for @xmath3 , i.e. one sets @xmath36 and @xmath37 in eq .
( [ eq : ff1 ] ) . from this assumption on the asymptotic behavior of the form factor ,
the derivatives of the form factor at @xmath38 are given by eq .
( [ eq : an ] ) with n=1 and m=0 .
in particular one has : @xmath39 where the standard definition @xmath40 is used .
eq.([eq : rms ] ) is a sum rule relating the pion rms radius to the magnitude of the time like pion form factor and the p - wave @xmath29 phase shift measurements . using these data ,
the derivatives of the form factor are evaluated at the origin : @xmath41 where the upper limit of the integration is taken to be @xmath42 . by fitting @xmath10 by the above mentioned asymptotic expression ,
the contribution beyond this upper limit is completely negligible .
the only experimental data on the derivatives of the form factor at zero momentum transfer is the rms radius of the pion , @xmath43 @xcite .
this value is very much in agreement with that determined from the sum rules .
in fact the sum rule for the rms radius gets overwhelmingly contribution from the @xmath0 resonance as can be seen from fig.1 . the success of the calculation of the r.m.s .
radius is a first indication that causality is respected and also that the extrapolation procedures to low energy for the p - wave @xmath29 phase shifts and for the modulus of the form factor are legitimate .
dispersion relation for the pion form factor is now shown to be well verified by the data over a wide energy region . using @xmath30 as given by eq .
( [ eq : ieu ] ) together with the once subtracted dispersion relation , one can calculate the real part of the form factor @xmath9 in the time - like region and also @xmath3 in the space like region .
because the space - like behavior of the form factor is not sensitive to the calculation schemes , it will not be considered here .
the result of this calculation is given in fig.2 .
as it can be seen , dispersion relation results are well satisfied by the data .
the i - loop chpt result can be put into the following form , similar to eq .
( [ eq : ff1 ] ) : @xmath44 where @xmath45 subtraction constants are needed to make the last integral on the rhs of this equation converges and @xmath46 with @xmath47 calculated by the @xmath48 loop perturbation scheme . similarly to these equations ,
the corresponding experimental vector form factor @xmath49 and @xmath50 can be constructed using the same subtraction constants as in eq .
( [ eq : peri ] ) but with the imaginary part replaced by @xmath51 , calculated using eq .
( [ eq : ieu ] ) .
the one - loop chpt calculation requires 2 subtraction constants . the first one is given by the ward identity , the second one is proportional to the r.m.s .
radius of the pion . in fig .
1 , the imaginary part of the one - loop chpt calculation for the vector pion form factor is compared with the result of the imaginary part obtained from the experimental data . it is seen that they differ very much from each other .
one expects therefore that the corresponding real parts calculated by dispersion relation should be quite different from each other . in fig.2 the full real part of the one loop amplitude
is compared with that obtained from experiment . at very low energy one can not distinguish the perturbative result from the experimental one due to the dominance of the subtraction constants . at an energy around @xmath52
there is a definite difference between the perturbative result and the experimental data .
this difference becomes much clearer in fig .
3 where only the real part of the perturbative di , @xmath53 , is compared with the corresponding experimental quantity , @xmath54 .
it is seen that even at 0.5 gev the discrepancy is clear .
supporters of chpt would argue that chpt would not be expected to work at this energy .
one would have to go to a lower energy where the data became very inaccurate .
this argument is false as can be seen by comparing the ratio @xmath55 .
it is seen in fig .
4 that _ everywhere _ below 0.6 gev this ratio differs from unity by a factor of 6 - 7 due to the presence of non perturbative effects .
similarly to the one - loop calculation , the two - loop results are plotted in fig .
( 1 ) - fig .
( 4 ) @xcite .
although the two - loop result is better than the one - loop calculation , because more parameters are introduced , calculating higher loop effects will not explain the data because in chpt both the form factor and scattering amplitude which enter in the imaginary part calculation are dominated by a polynomial behavior .
it is seen that perturbation theory is inadequate for the vector pion form factor even at very low momentum transfer .
this fact is due to the very large value of the pion r.m.s . radius or a very low value of the @xmath0 mass ( see below ) .
two unitarized models which are relevant are as follows .
the first model is obtained by introducing a zero in the calculated form factor to get an agreement with the experimental r.m.s . radius .
the pion form factor is now multiplied by @xmath56 where @xmath32 is the @xmath0 mass squared @xcite .
the experimental data can be fitted with a @xmath0 mass equal to @xmath57 and @xmath58 .
these results are in excellent agreement with the data @xcite . the second model , which is more complete at the expense of introducing more parameters , is based on the two - loop chpt calculation with unitarity taken into account .
it has the singularity associated with the two loop graphs . using the same inverse amplitude method as was done with the one - loop amplitude , but generalizing this method to two - loop calculation
, hannah has recently obtained a remarkable fit to the pion form factor in the time - like and space - like regions .
his result is equivalent to the ( 0,2 ) pad approximant method as applied to the two - loop chpt calculation @xcite .
both models contain ghosts which can be shown to be unimportant @xcite . as can be seen from figs . 1 , 2 and 3 the imaginary and real parts of these two models are very much in agreement with the data .
a small deviation of @xmath10 above @xmath59 is due to a small deviation of the phases of @xmath3 in these two models from the data of the p - wave @xmath29 phase shifts .
let us examine in details the one - loop chpt calculation of the vector pion form factor @xmath3 @xcite : @xmath60 where @xmath61 and the r.m.s .
radius of the vector form factor is related to @xmath62 by the definition @xmath63 .
the function @xmath64 is given by : @xmath65 for@xmath66 ; for other values of s , @xmath67 can be obtained by analytic continuation . the unitarised version of eq . ( [ eq : pertv ] ) , obtained by the inverse amplitude , the pad approximant or the n / d methods , is given by @xcite : @xmath68 it is obvious that eq .
( [ eq : vu ] ) has the breit - wigner resonance character while that from eq .
( [ eq : pertv ] ) does not , although their amplitude and first derivative are identical at @xmath38 .
furthermore , if the parameter @xmath69 was fixed by the the r.m.s .
radius , the @xmath0 mass squared , @xmath32 , would come out to be slightly low compared measured @xmath0 mass . neglecting this last problem which is unimportant here
, the taylor s series expansion around @xmath38 reveals that eq .
( [ eq : vu ] ) gives rise to a coefficient of the @xmath70 term as @xmath71 which is much larger than that coming from the third term of eq .
( [ eq : pertv ] ) , @xmath72 .
this is the signal of the failure of the perturbation method .
in other words , chpt should work if the r.m.s .
radius is much smaller than its experimental value or @xmath73 .
this last condition means that the physical @xmath0 mass is far too small for the perturbation theory to be valid
. we can generalize our criteria for the validity of the perturbation method to the nth loop result .
for this purpose we have to add one more step in our calculation in order to make it free from criticisms : eq .
( [ eq : vu ] ) could in principle contain unwanted poles due to the unitarisation procedure .
this can be eliminated by subtracting out the ghost pole which can be conveniently done by writing down a dispersion relation for @xmath3 with the same subtraction constants as used in the perturbative series eq .
( [ eq : pertv ] ) , but with the imaginary part given by eq .
( [ eq : vu ] ) @xcite .
provided that the ghost removing procedure did not change very much the unitarised amplitude , one could then compare the perturbative calculation with the modified unitarized result .
if their difference was negligible , one could be sure that the perturbative scheme could then be used .
because the calculable quantities of the vector pion form factor in the chpt scheme are too small ( well within experimental errors ) compared with the uncalculable ones , unless some unitarisation is made , it may be better to give up this perturbative scheme .
one would then gain in the transparency of the physics .
although we have not made here a detailed study of the processes @xmath74 the loop contributions can be shown to be completely negligible compared with the subtracted terms and are therefore not relevant . in conclusion ,
higher loop perturbative calculations do not solve the unitarity problem .
the perturbative scheme has to be supplemented by the well - known unitarisation schemes such as the inverse amplitude , n / d and pad approximant methods @xcite . for a review , see j. f. donoghue , e. golowich and b. r. holstein , _ dynamics of the standard model _ ( cambridge univ . press , cambridge , 1992 ) .
s. weinberg , physica 96 a ( 1979 ) 327 .
j. gasser and h. leutwyler , ann .
( n.y . ) * 158 * ( 1984 ) 142 .
j. gasser and h. leutwyler , nucl .
b250 * 465 ( 1985 )
. + ibid . *
b250 * ( 1985 ) 539 .
j. gasser and ulf - g mei@xmath75ner , nucl .
b357 * ( 1991 ) 90 . g. colangelo _
et al . _ ,
d * 54 * , ( 1996 ) 4403 ; j. bijnens _ et al _
, j. high energy phys .
_ 05 _ , ( 1998 ) 014
. k. m. watson , phys . rev .
* 95 * , ( 1954 ) 228
. l. m. barkov _
b256 * , ( 1985 ) 365 .
aleph collaboration , r. barate _
et al . _ , z. phys .
c * 76 * , ( 1997 ) 15 .
s. d. protopopescu _ et al .
d * 7 * , ( 1973 ) 1279 .
b. hyams _ et al .
. phys . * b64 * , ( 1973 ) 134 .
p. eastabrooks and a. d. martin , nucl .
b79 * , ( 1974 ) 301 .
na7 collaboration , s. r. amendolia _ et al .
_ , nucl . phys . *
b277 * , ( 1986 ) 168 .
t. n. truong , phys .
lett . * 61 * , ( 1988 ) 2526 .
t. n. truong , phys .
* 67 * , ( 1991 ) 2260 .
le viet dung and tran n. truong , report no hep - ph/9607378 .
t. hannah , phys . rev . *
d55 * , ( 1997 ) 5613 .
d thesis , aarhus university ( 1998 ) .
t. hannah and tran n. truong ( unpublished ) t. n. truong , phys .
* b 313 * , ( 1993 ) 221 . h. lehmann , phys . lett . * b 41 * , ( 1972 ) 529 , a. dobado , m.j .
herrero and t. n. truong , phys . lett .
* b 235 * , ( 1990 ) 129 , 134 , s. willenbrock , phys . rev .
* d43*,(1991 ) 1710 and references cited therein , a. dobado and j. r. pelaez , phys .
* d56 * ( 1997 ) 3057 , j. a. oller , e. oset and j. r. pelaez , phys .
lett . * 80 * , ( 1998 ) 3452 .
fig.1 : the imaginary part of the vector pion form factor @xmath76 , given by eq .
( [ eq : ieu ] ) , as a function of energy in the @xmath77 unit .
the solid curve is the the experimental results with experimental errors ; the long - dashed curve is the two - loop chpt calculation , the medium long - dashed curve is the one - loop chpt calculation , the short - dashed curve is from the modified unitarized one - loop chpt calculation fitted to the @xmath0 mass and the experimental r.m.s . radius , and the dotted curve is the unitarized two - loop calculation of hannah @xcite . fig . 2 : the real parts of the pion form factor @xmath78 , given by eq . ( [ eq : reu ] ) as a function of energy .
the curves are as in fig .
the real part of the form factor calculated by the once subtracted dispersion relation using the experimental imaginary part is also given by the solid line .
fig . 4 : the ratio of the one - loop chpt to the corresponding experimental quantity , @xmath55 , defined by eq .
( [ eq : dperti ] ) , as a function of energy , is given by the solid line ; the corresponding ratio for the two - loop result is given by the dashed line .
the ratio of the unitarized models to the experimental results is unity ( not shown ) .
the experimental errors are estimated to be less than 10% . | the vector pion form factor is used as an example to analyze this question . given the experimental radius of the pion , the crucial question is whether perturbative methods could be used for the effective chiral lagrangian to calculate the pion form factor .
our analysis shows that the pion rms radius is far too large ( or the related @xmath0 resonant mass is too low ) for the perturbation theory to be valid .
epsf -50pt |
fullerenes are stable carbon molecules with spherical , ellipsoidal or cylindrical hollow shape . the existence of these carbon structures was theoretically predicted by different authors between 1965 and 1975 @xcite , but it was not until 1985 that kroto _
et al . _ observed @xmath6 @xcite .
fullerene molecules , together with graphene and carbon nanotubes , are of considerable interest to researchers from many scientific areas and have multiple possible applications , specially on nanotechnology .
their unique mechanical and electronic properties , such as low weight , high strength , flexibility and thermal stability , make this carbon structures quite suitable for the creation of nano - devices @xcite .
one of the most important features of fullerene molecules is that they can harbor atoms in their interior constituting endohedral compounds , denoted by @xmath4 @xcite , whose main advantage is the isolation of the enclosed atom from its environment .
endohedral atoms can be produced in laboratories @xcite , and many technological applications , such as drugs - delivery agents , molecular containers , hydrogen or lithium storage , are now in development @xcite .
previous research mostly aimed to investigate the shell structure and spectral properties of @xmath7 , in which the enclosed atom is located right at the geometrical center of @xmath6 , and the fullerene molecule is modeled by a short - range spherical shell with an attractive potential @xcite .
nevertheless , for larger fullerene molecules the equilibrium position of the confined atom is no longer the geometrical center of the molecule @xcite .
the purpose of the present work is to calculate and describe the physically expected differences between the electronic states of the atom confined by the fullerene cage and their free state .
we focus on the systems @xmath4 , where @xmath8 , @xmath1 , @xmath2 , and @xmath9 .
the chosen @xmath10 isomers were @xmath11 and @xmath12 because they are the only two isomers that have been prepared and characterized as pristine structures @xcite .
each of these fullerene molecules was modeled by a short range shell according to molecule symmetry .
@xmath13 was modeled by a spherical shell , while both @xmath10 isomers were described by an ellipsoidal shell with proper parameters .
the fundamental question underlying our work could be stated in the following way : is the endohedral compound formed by a neutral atom inside a neutral fullerene ( @xmath4 ) , or does the valence electron of the encapsulated atom localize into the fullerene shell , giving rise to a zwitterion - like state ( @xmath5 ) ? in order to understand the transference of the valence electron to the fullerene cage @xmath14 , we analyze the electronic density , the projections onto free atomic states , and the weights of partial angular waves as a function of the attraction that the fullerene exerts upon the valence electron .
it is important to mention that the short range spherical ( or ellipsoidal ) potential for the fullerene cage is indeed a very simple approach ; even so , this model explains and provides a good description of the system and its properties .
there are plenty of studies of the electronic structure as well as photoionization of @xmath7 where the fullerene molecule is modeled as a spherical short range shell @xcite , which show an excellent agreement with another theoretical studies as well as with the experimental photoionization cross section or electron elastic scattering cross section @xcite .
this paper is organized as follows . in the second section
, we present a complete description of the theoretical method ; the third section is focused on the computation , results and discussion .
finally , our conclusions are given in the fourth section .
in a first step we used the structure of the fullerene molecules to obtain the position of the confined atom as the minimum of the classical @xmath3-body potential as explained in sec .
[ obtained x - position ] .
once the position of the atom is determined , the electronic structure of the enclosed atom is calculated modelling the fullerene as a spherical or ellipsoidal square well according to the symmetry of the fullerene molecule .
we model the interaction between the @xmath15 atom and the carbon fullerene using the non - bonded classical lennard - jones potential with each carbon atom of the @xmath16 . @xmath17 where @xmath18 is the position of the confined atom , @xmath19 the position of the @xmath20-th carbon from ref .
@xcite , and the pair @xmath21 , @xmath22 are the lennard - jones parameters adjusted for the @xmath23 interaction , displayed in table [ tljp ] ( @xmath24 are used throughout this paper ) .
the equilibrium position @xmath25 of the @xmath15 confined atom is calculated minimizing this potential , @xmath26 .[tljp ] lennard - jones parameters for @xmath23 interactions .
( a ) from ref .
@xcite , ( b ) from ref .
@xcite and ( c ) from ref .
[ cols="^,^,^,^",options="header " , ] once the equilibrium position of the enclosed atom was determined , we focused on the valence electron states . in order to study
how the spectral properties of @xmath0 , @xmath1 and @xmath2 atoms are modified by the carbon cage , our calculations included values of the depth of the attractive shell @xmath27 in the @xmath28 $ ] range .
the first four energy levels of the encapsulated atoms for each fullerene are depicted in fig .
[ av ] as a function of @xmath27 , @xmath11 in full black line , @xmath12 in dash - dotted orange line , and @xmath29 in gray dash - dotted line .
even when the main spectrum difference is between the @xmath10 isomers and @xmath29 , this difference is small because the distance between the core atom @xmath15 and the fullerene shell is determinated by @xmath22 , and it is almost independent of the size of @xmath30 ( see tables [ tljp ] , [ tsph - ellipp ] , and [ td ] ) .
notice that when varying the depth of the shell , even for values of @xmath31 , the ground state of @xmath0 remains stable at the ionization energy of the free atom ground state @xmath32 , a behaviour that has also been observed for the @xmath6 case in ref .
@xcite . in other words , given that the ground electronic state of hydrogen is highly localized around the nucleus , it is not influenced by the outer confining well unless the fullerene becomes sufficiently attractive .
the other @xmath0 states , as well as the @xmath1 and @xmath2 states , are influenced by the attractive fullerene shell even for small @xmath33 values because all of these states are delocalized states .
( color online ) first four energy levels of the encapsulated atoms at @xmath11 ( black full line ) , @xmath12 ( orange dash - dotted line ) , and @xmath29 ( gray dash - dotted line ) . ]
the ritz - variational method provides not only the eigenvalues of the hamiltonian matrix , but also the eigenfunctions .
the probability density of finding the valence electron around a position given by @xmath34 is proportional to @xmath35 , where @xmath36 is defined in eq .
( [ epsi ] ) .
the probability of finding the valence electron on the @xmath37 plane for the ground ( @xmath38 ) and first excited wavefunctions ( @xmath39 ) of @xmath40 are depicted in fig .
[ fu2_d2 ] , where the enclosed atom is located in the positive @xmath41-axis at the positions given in table [ td ] .
this figure shows several interesting features .
first of all , the symmetry of the localization in the fullerene shell depends both on the distance between the enclosed atom and the geometrical center of the molecule ( origin of coordinates ) and on the symmetry of the atomic free ( @xmath42 ) states ( @xmath43-wave for ground state and @xmath44-wave for first excited state ) .
it is physically expected that an electron in an @xmath43-wave with no spatial - direction preference will be localized in the nearest part of the fullerene shell . in the case of an electron in a @xmath44-wave with spatial preference for
the @xmath41 positive and negative axis , denoted by @xmath45 and @xmath46 , this spatial preference and the tendency to localizing in the nearest part of the fullerene attractive shell will compete .
the decisive variable in this competition is the distance between the enclosed atom and the geometrical center of the fullerene . in the case of @xmath2 ,
whose nucleus is located closer to the geometrical center of the fullerene , the valence electron in the ground state will localize in the whole fullerene shell when the fullerene gets increasingly attractive .
an electron in the first excited state of @xmath2 will localize mostly on the upper and lower hemispheres " of the fullerene shell , in a @xmath44-wave symmetric shape . in the case of @xmath1 ,
whose nucleus locates the farthest from the geometrical center of the fullerene , the valence electron in the ground state will localize on the nearest part of the fullerene shell , which is the upper hemisphere " .
an electron in the first excited state of @xmath1 will also localize on the lower hemisphere " due to the @xmath44-wave symmetry of the initial state ( @xmath42 ) . as mentioned above
, the ground @xmath0 state remains unchanged until the fullerene shell becomes sufficiently attractive ( @xmath47 ) , and for @xmath33 values above this critical value , the electron becomes localized on the fullerene shell .
the first excited @xmath0 state presents a quite interesting phenomenon which coincides with @xmath33 values for the anticrossing energy levels , @xmath48 $ ] in fig .
[ av ] . by increasing the @xmath33 values in the range
@xmath49 $ ] the electron becomes localized in the fullerene , and for the anticrossing values @xmath48 $ ] the electron is confined into the atom and relocalized in the fullerene with a new symmetry . a similar effect was originally described and termed _ mirror collapse _ by connerade _
_ for the @xmath6 case in ref .
@xcite .
( color online ) valence electron probability density on @xmath37 plane of the ground ( @xmath38 ) and first excited wavefunctions ( @xmath39 ) of @xmath40 for increasing @xmath33 values from top to bottom . for @xmath50 , @xmath38 the values are @xmath51 , for @xmath39 the values are @xmath52 . for @xmath53 , @xmath38 the values are @xmath54 , for @xmath39 the values are @xmath55 . for @xmath56 , @xmath38 the values are @xmath57 , for @xmath39 the values are @xmath58 . ] with the aim of clarifying these features , we calculated as a function of @xmath33 the probability of finding the valence electron inside the fullerene cage , @xmath59 where @xmath60 is the inner volume of the fullerene .
we also calculated the probability of finding the valence electron in the positive upper hemisphere " of the fullerene shell , denoted by @xmath61 , @xmath62 where @xmath63 is the volume of the upper hemisphere " of the fullerene molecule .
the probability of finding the valence electron in the lower hemisphere " of the fullerene @xmath64 is defined in a similar manner .
these probabilities are shown in fig .
[ frhon1 ] for the ground state and in fig .
[ frhon2 ] for the first excited state .
it is clear that the valence electron of @xmath1 and @xmath2 becomes localized in the fullerene shell as soon as the interaction with the fullerene is non - zero , and this localization takes place in a soft and monotonic way .
the studied hydrogen electronic states show a different behaviour .
the ground state localization occurs in a step - like way at @xmath47 , as one can see from the staggered curves of @xmath65 , @xmath61 and @xmath64 in fig .
[ frhon1 ] .
the mentioned re - confinement of the excited state in a narrow @xmath33 range is exposed in the high peak of @xmath65 ( fig .
[ frhon2 ] ( a ) ) for @xmath48 $ ] in the case of @xmath10 isomers and in the @xmath66 $ ] range for @xmath13 .
as expected , for these values of @xmath33 , @xmath61 , and @xmath64 undergo an appreciable decrease . as it is shown in fig .
[ frhon1 ] and [ frhon2 ] , the main difference in electron localization is between @xmath10 isomers and @xmath13 . for @xmath13 ,
the nucleus of the enclosed atom locates at the largest distance from the lower hemisphere " of the fullerene ; as a consequence , the valence electron ground state for all the elements becomes localized only in @xmath67 ( fig .
[ frhon1 ] ) , and the first excited states ( @xmath44 wave for @xmath42 ) becomes localized mostly in @xmath67 and shows poor localization in @xmath68 ( fig .
[ frhon2 ] ) . for @xmath10 cases ,
since the @xmath2 nucleus locates the closest to the geometrical center of the fullerene , its states show a very equilibrated localization , which is more pronounced for @xmath39 ( fig .
[ frhon2 ] ( b ) ) .
the valence electron of @xmath1 in the ground state shows a complete localization in @xmath67 , and as is expected given the symmetry of the initial state , the excited state displays a more equilibrate localization between @xmath67 and @xmath68 .
( color online ) probability of finding the valence electron in the ground state , inside the fullerene cage @xmath65 ( first column from left to right ) , and inside the positive and negative z - axis shell of the fullerene , @xmath61 and @xmath64 ( second and third column ) . ]
( color online ) probability of finding the valence electron in the excited state , inside the fullerene cage @xmath65 ( first column from left to right ) , and inside the positive and negative z - axis shell of the fullerene , @xmath61 and @xmath64 ( second and third column ) , for the hydrogen case ( top ) and the lithhium an sodium case ( bottom row ) . ]
now , one could ask if the re - confinement of the excited state in the hydrogen case is a re - confinement into the atom or just into the inner volume of the fullerene . to answer this question we calculated the projections onto the free atomic states _
i.e. _ without interaction with the fullerene ( @xmath42 ) , @xmath69 where @xmath70 denote the ground and excited free atomic states .
the projection of the first excited state of @xmath0 onto the ground atomic free state is displayed in fig .
[ fcin2_h ] , which shows that the re - localization of the @xmath0 excited state is a re - confinement into the atom in a state that is quite similar to the atomic free ground state , and not only the return of the electron to the inner fullerene cage .
( color online ) projection of the first excited state onto the ground atomic free state for @xmath50 ( dark gray full line ) @xmath50 ( dark gray dash line ) and @xmath71 ( light gray dash line ) . ]
finally , we performed an analysis of the mixing of different angular partial waves .
for this purpose we followed kang _
_ @xcite and calculated the partial wave - weights of the ground and excited states . from the normalization condition of the wavefunction , using eq .
( [ eu_bs ] ) , and expanding the angular b - spline into legendre polynomials , @xmath72 it is possible to define a partial wave overlap matrix , with elements given by , @xmath73 and , based on this definition , the partial wave - weights of the state are obtained , @xmath74 the partial weights of @xmath43 , @xmath44 and @xmath75-waves as a function of @xmath33 are shown in fig .
[ fpl_c80 ] for the @xmath10 isomers .
as expected , the larger the distance between the nucleus of the enclosed atom and the geometrical center of the fullerene is , the more the partial waves get mixed .
therefore , the @xmath1 states present a higher mixing of partial angular waves ( see fig .
[ fpl_c80 ] ( c ) and ( d ) ) while @xmath2 shows the purest wave - states ; the valence electron ground state of @xmath2 is mostly an @xmath43-wave ( see fig .
[ fpl_c80 ] ( e ) ) and its excited states , a @xmath44-wave ( fig .
[ fpl_c80 ] ( f ) ) .
all of the hydrogen weights exhibit a quasi - critical behavior at @xmath76 ( see fig .
[ fpl_c80 ] ( a ) and ( b ) ) .
in particular , the first excited state of @xmath77 switches its principal partial angular weight from @xmath43 to @xmath44 as can be seen in fig .
[ fpl_c80 ] ( b ) .
this means that by increasing from zero the values of @xmath33 , the electron becomes localized in the fullerene as an @xmath43-wave .
after that , for @xmath33 values in a small interval around the anticrossing value @xmath78 , the electron is confined into the atom and relocalized in the fullerene , but this time as a @xmath44-wave .
( color online ) partial weights of @xmath43 , @xmath44 , and @xmath75-waves for @xmath10 isomers .
the different elements are in columns ( @xmath0 , @xmath1 and @xmath2 from left to right ) and the ground and excited states are displayed in the rows from top to bottom . ]
in this work we described how the valence electron of @xmath0 , @xmath1 and @xmath2 atoms enclosed by three different fullerene molecules ( @xmath11 , @xmath12 and @xmath29 ) localizes in the fullerene when the strength of the attractive molecule cage increases .
we used the structure of the fullerene molecules to calculate the equilibrium position of the endohedrally atom as the minimum of the classical @xmath3-body lennard - jones potential . from lj parameters ( table [ tljp ] ) we obtained that @xmath1 atoms locate farther than @xmath0 and @xmath2 atoms in all cases .
once the position of the enclosed atom was determined , the fullerene molecules were modeled by a short range attractive shell according to molecule symmetry , and the enclosed atoms were modeled by an effective one - electron potential .
@xmath13 was modeled by a spherical shell , while both @xmath10 isomers were described by an ellipsoidal shell with proper parameters .
we found that the localization of the valence electron on the fullerene shell depends on the distance between the nucleus of the enclosed atom and the geometrical center of the fullerene , as well as on the symmetry of the atomic free states _
i.e. _ without interaction with the fullerene cage .
as we expected , the main differences between fullerene molecules were observed when comparing @xmath13 and @xmath10 isomers .
the @xmath1 and @xmath2 valence electron becomes localized in the fullerene shell continuously and monotonically as soon as the interaction with the fullerene is non - zero .
thus , @xmath1 and @xmath2 atoms inside a fullerene cage present zwitterion - like states _
i.e. _ the endohedral compound has the form @xmath79 and @xmath80 .
the hydrogen electronic states present a different behavior . in order to understand it
, we analyzed the electron probability density , the projections onto free atomic states and the weights of partial angular waves .
the @xmath0 ground state remains unaltered until @xmath33 values get over @xmath81 ; beyond this value the electron becomes localized in the fullerene shell in a step - like way .
therefore , the endohedral compound in the case of hydrogen is formed by a neutral atom inside a neutral fullerene @xmath82 .
the first excited @xmath0 state presents a quite interesting phenomenon which coincides with @xmath33 values for the anticrossing energy levels . increasing @xmath33
while still remaining under the anticrossing values , the electron becomes localized in the fullerene as an @xmath43-wave . for anticrossing values the electron
is confined into the atom in a state similar to the atomic free ground state .
when @xmath33 takes values slightly above the anticrossing , the electron is re - localized in the fullerene shell as a @xmath44-wave .
this switching of the principal partial angular weight could be a possible way to define the avoided crossing @xmath33 value .
we would like to make some final remarks on one of the possible applications of the simple approach used in the present work .
so far , light induced conical intersection ( lici ) as first proposed by m. sindelka , n. moiseyev and l.s .
cederbaum was applied to diatomic molecules in running and standing laser waves @xcite . within the present approach it is possible to study the coupling between nuclear and electronic modes in endohedral compounds when a conical intersection @xcite between electronic levels is induced by laser waves .
work is in progress in this direction .
we acknowledge secyt - unc and conicet for partial financial support of this project . e.c . would like to thank alvaro cuestas for an exhaustive reading of the manuscript . | the localization of the valence electron of @xmath0 , @xmath1 and @xmath2 atoms enclosed by three different fullerene molecules is studied .
the structure of the fullerene molecules is used to calculate the equilibrium position of the endohedrally atom as the minimum of the classical @xmath3-body lennard - jones potential .
once the position of the guest atom is determined , the fullerene cavity is modeled by a short range attractive shell according to molecule symmetry , and the enclosed atom is modeled by an effective one - electron potential . in order to examine whether the endohedral compound is formed by a neutral atom inside a neutral fullerene molecule @xmath4 or
if the valence electron of the encapsulated atom localizes in the fullerene giving rise to a state with the form @xmath5 , we analyze the electronic density , the projections onto free atomic states , and the weights of partial angular waves . |
the density functional theory ( dft ) of hohenberg , kohn , and sham is the method of choice for describing the ground - state properties of materials @xcite .
however , in the initial derivation of the dft , the eigenvalues are lagrange multipliers introduced to orthogonalize the eigenvectors , which in their turn are used to compute the total energy and the charge density . in this formulation
the eigenvalues have therefore no physical meaning and should not be considered as excited states .
nevertheless , the dft in the local density approximation ( lda ) or in its spin resolved local density formulation ( lsda ) , has been used successfully to compute the excited states , namely , optical and magneto - optical properties , x - ray absorption and magnetic dichroism spectra .
the lda or lsda were indeed intended to compute the ground - state properties of materials , and their use during the last two decades has produced an excellent track record in the computation of these properties for a wide variety of materials , ranging from simple metals to complex semiconductor superlattices .
however , it is now believed that the dft can do more than computing the ground state properties .
this is because the kohn sham equations could be viewed as deriving form a simplified quasi - particle ( qp ) theory where the self - energy is local and time averaged , i.e. , @xmath1 , here @xmath2 is the local exchange and correlation potential as , for example , parameterized by von barth and hedin @xcite .
viewed in this way , the ks eigenvalues are then approximate qp energies and could be compared to experimental data .
this argument is supported by quasiparticle calculations within the so called gw approximation of hedin @xcite showing that the valence qp energies of semiconductors are in good agreement with these obtained using lda , and the conduction qp energies differ by approximately a rigid energy shift @xcite . in the literature
this shift is often called `` scissors - operator '' shift @xcite . in the last few years
spectroscopy is becoming the standard tool for measuring excited states of materials .
its owes its impressive advances to the availability of synchrotron tunable highly polarized radiation .
in particular , the measurement of optical , magneto - optical properties as well as magnetic x - ray dichroism are now becoming routine tasks for probing the structural and magnetic properties of materials .
considerable attention has been focused on transition - metal surfaces and and thin films due to their novel physical properties different from that of bulk materials and due to potential industrial applications such as magneto - optical recording , sensors , or technology based on giant magneto - resistance . in this respect
, theory is falling far behind experiment and it is becoming hard to give a basic interpretation of experimental data .
this paper , which is far from being a review paper about calculated excited states , tries to bridge the gap between experiment and theory by describing a rather quantitative method for computing excited states of materials .
this method uses the local density approximation and the linear muffin - tin orbital ( lmto ) method . in the first part of this paper
we introduce the density functional theory and the local density approximation and justify the use of lda eigenvalues as approximate excited states and relate them to quasiparticle energies .
in the second part we describe the construction of the lmto basis set within an all - electron full - potential approach @xcite which will be used to determine the momentum matrix elements .
we devote the third part to the determination of the momentum matrix elements . in the first part of the application section we present some examples of computation of semiconductors optical spectra @xcite , and leave out the optical properties of metals and magneto - optical properties of materials and refer the reader to ref . @xcite . in the second part of the applications we show some examples of x - ray magnetic dichroism calculations @xcite .
the density functional method of hohenberg and kohn @xcite which states that the ground state total energy of a system of n interacting electrons in an external potential @xmath3 is a functional of the electron density @xmath4 does not provide an analytical form of the functional @xcite .
this method remains numerically intractable without the kohn and sham introduction of the so called local density approximation @xcite in which the exchange and correlation functional @xmath5 appearing in the total energy : @xmath6 is given by @xmath7 where @xmath8 is the exchange - correlation energy of a uniform electron gas of density @xmath9 .
thus , kohn and sham constructed a set of self - consistent single - particle equations : @xmath10 where the density @xmath11 is given by : @xmath12 and @xmath13 @xmath14 instead of the true kinetic energy of the electron gas , kohn and sham used the homogeneous electron kinetic energy : @xmath15 this use of homogeneous - electron kinetic energy in the kohn sham equations redefined the exchange - correlation function to be :
@xmath16 it is then crucial to use a good basis - set for the description of the electronic structure of realistic systems . the augmented plane wave @xcite ( apw ) , and the korringa - kohn - rostoker @xcite ( kkr ) methods can be used , in principle , to solve exactly the kohn
sham equations , however these methods are numerically involved and their linearization , introduced by andersen is much preferable .
andersen linearization , has not only made the techniques for solving the band - structure problem transparent by reducing it essentially to the diagonalization of one - electron hamiltonian , and cuts the cost of computation by at least one order of magnitude .
the linearized versions of these two powerful methods are the linear augmented plane wave ( lapw ) and linear muffin - tin orbital ( lmto ) methods , respectively @xcite . in this paper
, we will only use the lmto method to study excited states of solids .
the reason for this choice is that the lmto method is the mostly used method in computational electronic structure .
this is due primarily to the use of atomic - sphere approximation ( asa ) which made the lmto method run fast even on today s cheap personal computers . due to this reduced computational cost ,
the lmto asa method became the method of choice of researchers without access to supercomputers .
the quasiparticle ( qp ) electronic structure of an interacting many - body system is described by the single - particle eigenstates resulting from the interaction of this single particle with the many - body electron gas of the system .
the single - particle eigenstate energies are the results of solving a schroedinger like equation containing the non - local and energy - dependent self - energy instead of the exchange - correlation potential appearing in kohn sham like equations : @xmath17 thus the self - energy @xmath18 contains all many - body effects .
almost all ab - initio qp studies were performed within the so - called @xmath19 approximation , where the self - energy @xmath18 is calculated within hedin s @xmath19 approximation .
this method consists of approximating the self - energy as the convolution of the lda self - consistent green function @xmath20 and the screened coulomb interaction @xmath21 within the random - phase approximation .
the qp eigenvalues are often obtained using first - order perturbation theory starting from lda eigenvalues and eigenvectors @xcite .
although there are early calculations starting from hartree - fock @xcite or tight - binding @xcite methods . nevertheless , the best results are based on a lda starting point @xcite .
thus the @xmath19 predicted optical excitations energies of semiconductors are within 0.1 ev form the experimental results and the surprizing fact is that the qp wave functions are almost identical to these produced within the lda @xcite ( the wave function overlap is more than 99% ) . for a general review of gw calculations
see the review by araysetianwan and gunnarsson @xcite or by aulbur , jnsson and wilkins @xcite .
it is clear that the quasiparticle schroedinger equation resembles to the kohn sham equation .
both equations describe a fictitious electron moving in a effective potential .
the difference is that the self - energy is nonlocal and energy dependent whereas the lda potential is local and averaged over time .
this resemblance can be further pushed by noticing that the dft can be used to obtain excitation energies .
for example , the ionization energy , @xmath22 , and the electron affinity , @xmath23 , are difference between ground state energies : @xmath24 were @xmath25 is the number of electrons of the system . and
since the dft gives the correct ground state energies it should produce , in principle , the correct ionization and electron affinity energies .
for metals , the addition or removal of an electron from the system costs the same energy , and hence the ionization energy is equal to the electron affinity . for insulators , the energy gap makes all the difference and hence breaks this symmetry . thus the energy band gap is given by : @xmath26 in practice , however , the calculation is often obtained within the lda and the energy band gap is calculated as the difference between the lowest conduction band and the highest valence band .
it was shown by sham and schlter @xcite and perdew and levy @xcite that the calculated energy gap differ from the true band gap by an amount @xmath27 even when the dft is used without the lda .
the @xmath27 value could range from @xmath28 in the case of silicon to @xmath29 in the case of germanium .
for most of the semiconductors , the @xmath19 calculations show that the lda eigenvalues differ form the @xmath19 quasiparticle energy by a constant @xmath27 which is almost independent of the @xmath30-point .
this finding is important and shows that the lda eigenvalues have some meaning and could be used to calculate excited states . so as stated in the introduction , the kohn
sham equations could be viewed as deriving form a simplified quasi - particle ( qp ) theory where the self - energy is made local and time averaged , i.e. , @xmath1 .
this approximation is certainly good for metals where we have a good data base for excited state calculated within the lda @xcite and where the agreement with experiment is good . for semiconductors ,
this approximation is not bad either , provided we know the value of the discontinuity of the exchange and correlation .
usually , this value is provided by @xmath19 calculations or by experiment .
in this section we describe the lmto basis - set used to calculate the excited states of solids .
we discuss the basis used for an all electron calculation where the potential is not supposed to be spherically symmetric nor of muffin - tin type .
the use of a general potential makes the study of open structures possible without having to resort to the so - called `` empty - sphere '' approximation . to define the basis - set , we divide the space into non overlapping spheres called `` muffin - tin '' spheres and a region between these spheres which we call interstitial region . inside the muffin - tin spheres
the schroedinger equation is solved at a fixed energy for each angular momentum @xmath31 and variational parameter @xmath32 ( which is defined later )
. the linearization amounts to the use of a linear combination of the solution @xmath33 of the schroedinger equation for a fixed energy and its energy derivative @xmath34 inside the muffin - tin spheres .
these linear combinations matche continuously and differentiably to an envelop function ( spherical function ) in the interstitial region .
the bloch wave function in the interstitial region is given by a linear combination of these hankel functions centered at each site : @xmath35 where @xmath36 stands for the number of the basis function quantum numbers ( these numbers are @xmath37 ) , where @xmath38 is the site number , @xmath39 groups the two angular quantum numbers , and @xmath40 is the linearization energy for a particular atom type @xmath41 and angular momentum number @xmath31 .
the envelop functions are defined as @xmath42 .
@xmath43 here @xmath44 is the neumann function and @xmath45 bessel function for the angular momentum @xmath31 , and @xmath46 are the spherical harmonics . to get the differentiability of the wave - function at the boundary of the muffin - tin spheres , we write the envelope function inside the muffin - tin spheres .
the envelope function for a muffin - tin sphere @xmath47 is given by : @xmath48 to produce smooth basis functions we require that the basis function is differentiable at the boundary of each muffin - tin sphere , i.e. , that a linear combination of @xmath49 and @xmath50 matches continuously and differentiably @xmath51 and @xmath52 at the boundary of the parent sphere and other spheres , respectively .
using these matching conditions at the muffin - tin spheres , the bloch wave function inside a muffin - tin sphere @xmath53 of the unit cell at the origin is given by @xcite : @xmath54 where @xmath55
@xmath56 @xmath57 to add the spin dependence to the basis - set , the bloch wave function is multiplied by the eigenvector of the pauli spin operator @xmath58 : @xmath59 such that : @xmath60 where @xmath61 is the quantization axis chosen in advance .
here we give a review of the determination of the dielectric response of a semiconductor due to the application of an electric field .
we expend the description of our published work @xcite by giving more details concerning the calculation of the momentum matrix elements .
an electromagnetic field of frequency @xmath62 , and a wave vector @xmath63 interacting with atoms in a crystal produces a response of frequency @xmath62 and a wave vector @xmath64 ( @xmath65 and @xmath66 being reciprocal lattice vectors ) .
the microscopic field of wave vector @xmath67 is produced by the umklapp processes as a result of the applied field @xmath68 @xmath69 where @xmath70 is the total field producing the non - diagonal elements in the microscopic dielectric function @xmath71
. the microscopic dielectric function in the random phase approximation is given by @xcite : @xmath72 here @xmath9 and @xmath73 are the band indexes , @xmath74 is the zero temperature fermi distribution , and @xmath75 is the cell volume .
the energies @xmath76 and the the crystal wave function @xmath77 are produced for each band index @xmath9 and for each wave vector @xmath78 in the brillouin zone .
the macroscopic dielectric function in the infinite wave length limit is given by the inversion of the microscopic dielectric function : @xmath79_{\bf 0,0 } } \\ \nonumber & = & \epsilon_{0,0}(\omega ) - \lim_{{\bf q }
\to { \bf 0 } } \sum_{{\bf g , g^\prime \ne 0 } } \epsilon_{0,\bf g}({\bf q } , \omega ) t^{-1}_{\bf g , g^\prime } ( { \bf q } , \omega ) \epsilon_{\bf g^\prime,0}({\bf q } , \omega ) \ ; \label{localfield}\end{aligned}\ ] ] where @xmath80 is the inverse matrix of @xmath81 containing the elements @xmath82 with @xmath83 and @xmath84 . the first term of this equation is the interband contribution to the macroscopic dielectric function and the second term represent the local - field correction to @xmath85 .
the most recent ab - initio pseudopotential calculation found that the local - field effect reduces the static dielectric function by at most 5% @xcite .
previous calculations with the same method have also found a decrease of @xmath86 by about the same percentage @xcite . for insulators
the dipole approximation of the imaginary part of the first term of equation ( [ localfield ] ) is given by @xcite : @xmath87 here @xmath88 is the velocity operator , and in the lda @xmath89 ( @xmath90 being the momentum operator ) , and where @xmath91 . the matrix elements @xmath92 are calculated for each projection @xmath93 , @xmath94 or @xmath95 and @xmath96 , with the wave function @xmath97 expressed in terms of the full - potential lmto crystal wave function described by equations ( [ wavemuffin ] ) and ( [ planewave ] ) .
the * k*-space integration is performed using the tetrahedron method @xcite with a large number of irreducible * k * points the brillouin zone .
the irreducible * k*-points are obtained from a shifted * k*-space grid from the high symmetry planes and @xmath98 point by a half step in each of the @xmath99 , @xmath100 , and @xmath101 directions .
this scheme produces highly accurate integration in the brillouin zone by avoiding high symmetry points . to calculate these matrix elements we first defined a tensor operator of order one out of the momentum operator @xmath102 .
the muffin - tin part of the momentum matrix elements is calculated using the commutator @xmath103=2 \nabla_{\mu } $ ] so that : @xmath104 where @xmath105 are the usual gaunt coefficients , and @xmath106 is the radius of the muffin - tin sphere of atom @xmath38 .
in the interstitial region the plane - wave representation of the wave function ( see equation [ planewave ] ) makes the calculation straightforward , but a special care has to be taken for the removal of the extra contribution in the muffin - tin spheres .
however , we find it much easier and faster to transform the interstitial matrix elements as an integral over the surface of the muffin - tin spheres using the commutation relation of the momentum operator and the hamiltonian in the interstitial region .
the calculation of the interstitial momentum matrix elements is then similar to the calculation of the interstitial overlap matrix elements .
the @xmath107 case has been already derived by chen using the korringa , kohn and rostoker greens - function method @xcite .
we have tested that both the plane - wave summation and the surface integration provide the same results .
@xmath108 a hankel function can be integrated over a volume by knowing its integral over the bounding surface : @xmath109 the surface of the interstitial consists of the exterior of the muffin - tin spheres and the unit cell boundary . over the surface of the muffin tins : the surface area is @xmath110 and the normal to the sphere points inward @xmath111 at a muffin - tin sphere boundary @xmath112 the bloch wave function is given by : @xmath113 where @xmath114 and @xmath115 let w denotes the wronskian @xmath116 we define then @xmath117 and @xmath118 @xmath119 \nonumber \\ & & \rb_{\ell m,\ell_i m_i}({\mathbf{\tau}}\sminus{\mathbf{\tau}}_i,\kappa_i,\bvc k ) \bigr)\end{aligned}\ ] ] then @xmath120 \\ \nonumber & & \rb_{\ell m,\ell_i m_i}({\mathbf{\tau}}\sminus{\mathbf{\tau}}_i,\kappa_i,\bvc k ) \bigr ) + \delta(f , i,\kappa_i)\end{aligned}\ ] ] where @xmath121 \nonumber \\ & & + \sum_{\mu } \hat\bvc e_{\mu}(\tau_f ) \biggl [ \\ \nonumber & & + b_{\ell_f+1 m_i+\mu,\ell_f m_f } ( { \mathbf{\tau}}_f\sminus{\mathbf{\tau}}_i,\kappa_i,\bvc k ) \gauntcof { \ell_f } { m_i } { \ell_f\splus1 } { m_f\splus\mu } 1 \mu \\
\nonumber & & - b_{\ell_f-1 m_f-\mu,\ell_i m_i } ( { \mathbf{\tau}}_f\sminus{\mathbf{\tau}}_i,\kappa_i,\bvc k ) \gauntcof { \ell_f } { m_f } { \ell_f\sminus1 } { m_f\splus\mu } 1 \mu \kappa^2 \biggr ] \end{aligned}\ ] ] equation ( [ ehrenreich ] ) can not be used directly to determine the optical properties of semiconductors , when the gw approximation or the scissors operator is used to determine the electronic structure .
the velocity operator should be obtained from the effective momentum operator @xmath122 which is calculated using the self - energy operator , @xmath123 , of the system @xcite :
@xmath124 gw calculations show that the quasiparticle wave function is almost equals to the lda wave function @xcite . based on this assumption
, it can be easily shown @xcite that in the case of the scissors operator , where all the empty states are shifted rigidly by a constant energy @xmath27 , the imaginary part of the dielectric function is a simple energy shift of the lda dielectric function towards the high energies by an amount @xmath27 , i.e. , @xmath125 .
the real part of the dielectric function is then obtained from the shifted @xmath126 using kramers - kronig relations .
the expression of @xmath127 is given by : @xmath128 @xmath129 is very similar to @xmath130 except that one of the interband gap @xmath131 is substituted by the qp interband gap @xmath132 .
to test for the accuracy of the calculation within the lda the f - sum rule : @xmath133 where @xmath134 is the number of valence bands , should be always checked to ensure the accuracy of the calculations .
it is easily seen that the dielectric function @xmath135 calculated using the scissors - operator shift does not satisfy the sum rule ( @xmath136 is the free - electron plasmon frequency ) : @xmath137 because ( i ) @xmath138 satisfies this rule , and ( ii ) @xmath135 is obtained by a simple shift of @xmath139 by the scissors - operator @xmath27 towards higher energies .
the non simultaneous satisfaction of both the f - sum rule and the integral sum rule within the scissors approximation shows the limitation of this approximation .
while the scissors operator approximation describes nicely the low lying excited states , which is seen in the good determination of the static dielectric function and the low energy structures , i.e. @xmath140 and @xmath141 , in the imaginary part of the dielectric function , it seems to fail for the description of the higher excited states .
this is not surprising because the higher excited states which are free electrons like are most probably well described within lda and need no scissors - operator shift .
this is supported by the fact that the the energy - loss function , -im@xmath142 , within the lda has it maximum roughly at the free electron plasmon frequency whereas within the scissors approximation its maximum is shifted to higher energies . for our purpose
the scissors - operator shift remains a good approximation for the description of the low - lying excited states of semiconductors and their optical properties .
we have used our fp - lmto method and the formalism outlined above to calculate the optical properties of materials @xcite . in general
our results are often in good agreement with the experimental results . for semiconductors ,
however , good agreement with experiment is only achieved when the so called scissors - operator shift is used .
figure 1 presents our relativistic calculation of the imaginary part of the dielectric function of gaas compared to the experimental results of ref .
@xcite .
is only slightly underestimated while @xmath141 is overestimated .
notice that the shifted dielectric function by 1.3 ev , which produces the correct band gap , overestimates the peak positions by about 0.3 ev .
excitonic effect should shift these peaks to lower values in agreement with experiment . ]
the lda relativistic results underestimates the band gap by about 1.3 ev .
when the imaginary part of the dielectric function is shifted to higher energies by 1.3 ev the results the @xmath140 and @xmath141 peaks are overestimated in our calculation .
one needs to shift the spectrum by less than the band gap as done in ref .
@xcite to produce good agreement with experiment .
it seems then that the optical band gap is less than the band energy gap ( 1.5 ev ) .
the optical band gap is produced by interband transitions to the low lying conduction states .
excitonic effects are therefore important and are responsible for the reduction of the energy gap of semiconductors .
it is interesting to notice though that the static dielectric function are in good agreement with experiment for gaas , si , and ge when the shift correspond to the energy band gap obtained from photoluminescence @xcite .
more interesting are the wide band - gap materials where the lda calculated static dielectric function is in good agreement with experiment despite that the band gap is still underestimated by lda .
correcting the band gap using the scissors operator makes the static dielectric much small than the measured value . as an example of wide gap material , we present in figure 2 and 3 the imaginary part of the dielectric function of gan for the cubic ( 3 ) and wurtzite structure ( b4 ) . .
calculated static dielectric function @xmath143 for gan compared to pseudopotential ( pp ) results and experiment . for the wurtzite structure we have calculated @xmath144 for a polarization parallel to the @xmath145 plan and @xmath146 which is perpendicular . [ cols= " < ,
< , < , < " , ] [ epsilon0 ] table i shows that our lda dielectric constant calculations are in agreement with available experimental results and the pseudo - potential ( pp ) results @xcite including local - field effects ( an error about our calculation is reported in ref .
@xcite ; our value for @xmath144 is not 4.48 but 5.54 and the pp value should then be 4.48 ) .
it is interesting to notice that static dielectric is in good agreement for the for all the nitrides @xcite while the band gap is underestimated .
the scissors - operator shift fails to explain the static dielectric function of large gap semiconductor .
recently , both local - field effects and electron - hole interaction were included on an ab - initio computation of the dielectric function of few semiconductors @xcite by extending the semi - empirical hanke and coworkers approach @xcite which is based on the solution of the bethe - salpeter equation @xcite .
the excitonic effects seem to improve significantly the agreement between theory and experiment .
however for large band - gap semiconductors , such as diamond , the inclusion of the excitonic effects seem to underestimate the optical band gap by as much as 1 ev @xcite .
it is not clear from these calculations whether the static dielectric function for wide - band gap semiconductors is improved when excitonic effects are included .
more theoretical work along these lines is needed to fully understand the dielectric function of wide - gap semiconductors .
x - ray absorption spectroscopy ( xas ) probes selectively each core orbital of each atomic species in a material .
two decades ago the theoretical work of erskine and stern show that the x - ray absorption could be used to determine the x - ray magnetic circular dichroism ( xmcd ) in transition metals when left and right circularly polarized x - ray beams are used @xcite .
more recently these ideas were implemented experimentally and xas was used to determine the local magnetic properties of each magnetic atomic orbital in a magnetic compound @xcite .
thus the circular magnetic x - ray dichroism is an important tool for the investigation of magnetic materials @xcite , especially through the use of sum rules for the direct determination of the local orbital and spin contributions to the total magnetic moment @xcite .
thole and co - workers show that the circular - magnetic - x - ray dichroism is related to the magnetic moment of the photo - excited atom when the core electron is excited to the conduction states that are responsible of the magnetic properties of the material . on the theoretical side , ebert and his co - workers @xcite
have developed a fully - relativistic local - spin - density - approximation approach that was used with success to calculate the xmcd at the k - edge of fe , the l@xmath147-edge of gadolinium , and fe and co multilayers .
wu et al used slab linear augmented plane wave method to study the l@xmath148 xmcd of fe @xcite .
brouder and co - workers uses multiple - scattering theory to solve the schrdinger using spherical potentials and spin - orbit coupling as a perturbation in the final state @xcite .
recently ankudinov and rehr used a method based on a non - relativistic treatment of propagation based on high order multiple scattering theory and spinor - relativistic dirac - fock treatment of the dipole matrix elements to calculate the fe k edge and gd l@xmath149 edge xmcd @xcite .
the calculation of the x - ray absorption for left and right circularly polarized x - ray beams is implemented within the local - density approximation ( lda ) by means of all - electron full - relativistic and spin - polarized full - potential linear muffin - tin orbital method ( lmto ) .
the core electrons are spin - polarized and their electronic states are obtained by solving the full - dirac equation , whereas for the valence electrons the spin - orbit coupling is added perturbatively to the the semi - relativistic hamiltonian .
the total hamiltonian is then solved self - consistently . to calculate the polarization dependent cross - section
we consider the case where the internal field polarizes the spins along the magnetization easy axis .
with respect to this axis we defined the left- and right - circular polarization , which correspond to the photon helicity ( @xmath150 ) ( @xmath151 ) respectively and the following dipole interaction : @xmath152 .
the absorption cross - section @xmath153 for left ( @xmath154 ) and right ( @xmath155 ) circular polarized x - ray calculated at the relativistic j@xmath156 ( @xmath157 ) core level and in the dipole approximation is given by : @xmath158 using lda in conjunction with the relativistic full - potential lmto technique .
represent the k - edge x - ray absorption of fe , for left and right circularly polarized light , compared to the experimental results .
the agreement at low energy with experiment is good and start degrading at higher energies above the mean absorption peak .
it is of interest to point out that the magnetic x - ray dichroism at the k - edge which is due to the spin polarization and the spin - orbit in the final state is very small in the case of fe .
the difference between the right and left circularly polarization of the light is not even visible on the graph .
however , the x - ray magnetic circular dichroism can be measured and figure 4 shows a good agreement of the calculated dichroic signal with the experimental results of shtz @xcite . at the l@xmath148 edge of @xmath159 transition metals
the x - ray magnetic dichroism is much important because it is meanly due to the presence of the strong spin - orbit coupling in the initial @xmath160 states ( in the case of fe the spin - orbit splitting between the @xmath161 and @xmath162 is about 13 ev ) . in figure 5
we show the calculated x - ray absorption and xmcd at the co in ptco ordered alloy @xcite .
-edge of fe compared to the experimental spectrum of grange _
et al._. ] to compare the results with experiment we have to take into account the effect of the core hole and the experimental resolution .
this is done by convoluting the calculated spectra by a lorentzian of widths of 0.9 ev and 1.4 ev for the l@xmath163 and l@xmath149 edges , respectively , in addition a gaussian broadening of 0.4 ev is added to take into account the experimental resolution .
the calculation of the x - ray magnetic circular dichroic signal ignoring the electron - hole recombination effect provides a semi - quantitative agreement with the experimental spectra .
hence , we believe that the core hole effect represented here by a lorentzian broadening plays a significant role in determining the correct l@xmath164 branching ratio for 3@xmath165 transition metals . the underestimation of the @xmath166 branching ratio remains a challenge for theorists and further theoretical development along the line proposed by schwitalla and ebert @xcite is needed to bring the theory at the level of the experiment . for the 4@xmath165-transition metals ,
the core hole is deeper , and the agreement with experiment of the xmcd is satisfactory .
figure 6 shows the calculated xmcd at the site of pt of the copt ordered alloy .
-edge of pt compared to the experimental spectrum of grange _
et al._. ] in contrast to what is obtained for co , the results for the pt site show a much better agreement with experiment , due to the fact that the core hole effect is less intense ( core hole much deeper than that of co ) . for the pt atom we used both a lorentzian ( 1 ev ) and a gaussian ( 1 ev ) to represent the core hole effect and a gaussian of 1 ev width for the experimental resolution .
the experimental and theoretical @xmath167 and @xmath168 edges are separated by a spin - orbit splitting of the @xmath160 core states of @xmath169 and @xmath170 ev respectively .
the width of both @xmath167 and @xmath168 edges is comparable to experiment , but the calculated @xmath167 edge is much larger .
this produces a calculated integrated branching ratio of 1.49 which is much smaller than the experimental ratio of 2.66 . here
again the theory is underestimating the branching ratio .
we have reviewed the fp - lmto method and the implementation of the optical properties and x - ray magnetic dichroism within the local density approximation .
we have showed that the momentum matrix elements can be evaluated as a muffin - tin contribution and a surface term .
the method has been successfully used to compute the optical properties of metals @xcite , semiconductors @xcite and magneto - optical properties @xcite of transition metals alloys , as well as x - ray magnetic circular dichroism @xcite with high precision . for small - gap semiconductors
a scissors - operator shift should be used to reproduce the static and dynamic dielectric function @xcite .
excitonic effects seem to be important in reproducing the correct optical energy gap @xcite . for wide - gap semiconductors the local - density approximation ( lda )
static dielectric function is in good agreement with experiment and no scissors - operator shift is required despite the underestimation of the band gap by lda @xcite . for the computation of the x - ray magnetic circular dichroism the agreement with experiment
is rather good @xcite .
however , the so called branching ratio is underestimated by the theory . more theoretical work where the electron core - hole interaction is taken into account
is needed to bring the theory at the quality level of experiment @xcite .
part of this work was done while one of us ( m.a ) was at ohio state university and were supported by nsf , grant number dmr-9520319 .
supercomputer time was granted by cnusc ( project gem1917 ) on the ibm sp2 and by the universit louis pasteur de strasbourg on the sgi o2000 supercomputer .
p. hehenberg et w. kohn , phys
. rev . * 136 * , b864 ( 1964 ) ; w. kohn et l. j. sham , phys .
* 140 * , 1133 ( 1965 ) .
u. von barth and l hedin , j. phys .
c * 5 * , 1629 ( 1972 ) .
l. hedin , phys . rev . * 139 * , a796 ( 1965 ) .
m. s. hybertsen and s. g. louie , phys .
b * 32 * , 7005 , ( 1985 ) ; _ ibid _ , * 34 * , 5390 , ( 1986 ) .
r. w. godby , m. schlter , and l. j. sham , phys .
b * 37 * , 10159 ( 1988 ) .
levine and d. allan , phys .
b * 43 * , 4187 ( 1991 ) ; phys .
* 66 * , 41 ( 1991 ) .
see the article by j. m. wills _
_ in this book .
o. k. andersen , phys .
b * 12 * , 3060 ( 1975 ) .
j. petalas , s. logothetidis , s. boultadakis , m. alouani , and j. m. wills , phys .
b * 52 * , 8082 ( 1995 ) . m. alouani and j. m. wills , phys .
b * 54 * , 2480 ( 1996 ) . for the calculation of magneto - optical properties of materials see a. delin , phd thesis , uppsala 1998 ; p. ravindran _ et al . _ * 59 * 15680 ( 1999 ) .
r. ahuja , s. auluck , j. m. wills , m. alouani , b. johansson , and o. eriksson , phys .
b * 55 * , 4999 ( 1997 ) .
c. s. wang and j. callaway , phys .
b * 9 * , 4897 ( 1979 ) ; d. g. laurent , j. callaway , and c. s. wang , phys . rev .
b * 19 * , 5019 ( 1979 ) .
uspenski , e. g. maksimov , s. u. rashkeev , and i. i. mazin , z. phys .
b * 53 * , 263 ( 1983 ) .
m. alouani , j. m. koch , and m. a. khan , j. phys .
f * 16 * , 473 ( 1986 ) .
m. amiotti _
b * 45 * , 13285 ( 1992 ) .
s. v. halilov and yu a. uspenskii , j. phys : condens .
matter * 2 * , 6137 ( 1990 ) ; s. v. halilov , _ ibid _ * 4 * , 1299 ( 1992 ) .
p. m. oppeneer , t. maurer , j. sticht , and j. kbler , phys .
b * 45 * , 10924 ( 1992 ) ; t. kraft , p. m. oppeneer , v. n. antonov , and h. eschrig , _ ibid _ * 52 * , 3561 ( 1995 ) .
k. w. wierman , j. n. hifiker , r. f. sabiryanov , s. s. jaswal , r. d. kirby , j. a. woollam , phys .
b * 55 * , 3093 ( 1997 ) .
m. alouani , j.m .
wills , and j.w .
wilkins , phys .
b * 57 * , 9502 ( 1998 ) .
w. grange , i. galanakis , m. alouani , m. maret , j. p.
kappler , a. rogalev , phys .
b ( accepted ) .
i. galanakis , s. ostanin , m. alouani , h. dreysse , j. m. wills , phys .
b ( submitted ) . j. c. slater , * 51 * , 151 ( 1937 ) ; t. l. loucks , _ augmented plane wave method _ , ( w. a. benjamin , inc . , new york 1967 ) j. korringa , physica * 13 * , 392 ( 1947 ) ; w. kohn and n. rostoker , phys . rev . * 94 * , 1111 ( 1954 ) .
r. w. godby , m. schlter , and l. j. sham , phys .
* 56 * , 2415 , ( 1986 ) ; phys .
b * 35 * , 4170 , ( 1987 ) ; _ ibid _ , * 36 * , 6497 ( 1987 ) , and * 37 * , 10159 ( 1987 ) . g. strinati , h. j. mattausch , and w. hanke , phys
. rev .
b * 25 * , 2867 ( 1982 ) .
sterne and j. c. inkson , j. phys .
c * 17 * , 1497 ( 1984 ) .
m. rohlfing , p. krger , and j. pollmann , phys .
b * 48 * , 17791 ( 1993 ) .
f. araysetianwan and o. gunnarsson , rep
. prog . phys . * 61 * , 237 - 312 ( 1998 ) .
w. g. aulbur , l. jnsson , and j. w. wilkins , _quasiparticle calculations in solids_ , to be published in solid state physics ; edited by h. ehrenreich .
l. j. sham and m. schlter , phys .
51 * , 1888 ( 1983 ) ; phys .
b * 32 * , 3883 ( 1985 ) .
j. p. perdew and m. levy , phys .
* 51 * , 1884 ( 1983 ) s. l. adler , phys . rev . * 126 * , 413 ( 1962 ) ; n. wiser , phys
. rev . * 129 * , 62 ( 1963 ) .
s. baroni and r. resta , phys .
b * 33 * , 7017 ( 1986 ) . h. ehrenreich and m. l. cohen , phys . rev . * 115 * , 786 ( 1959 ) .
o. jepsen et o. k. andersen , solid state commun .
* 9 * , 1763 ( 1971 ) ; g. lehmann et m. taut , phys .
* 54 * , 469 ( 1972 ) .
a. b. chen , phys .
b * 14 * , 2384 ( 1976 ) .
r. del sole and r. girlanda , phys .
b * 48 * , 11789 ( 1993 ) .
d. e. aspnes and a. a. studna , phys .
b * 27 * , 985 ( 1983 ) .
j. chen , z. h. levine , and j. w. wilkins , appl .
. lett . * 66 * , 1129 ( 1995 ) .
s. albrecht , l. reining , r. del sole , and g. onida , phys . rev
* 80 * , 4510 ( 1998 ) . l. x. benedict , e. l. shirley , and r. b. bohn , phys .
lett . * 80 * , 4514 ( 1998 ) ; phys .
b * 57 * , r9385 ( 1998 ) .
g. strinati , h. j. mattausch , and w. hanke , phys .
* 45 * , 290 ( 1980 ) ; h. j. mattausch , w. r. hanke , and g. strinati , phys .
b * 27 * , 3735 ( 1983 ) ; n. meskini , h. j. mattausch , and w. hanke , solid stat .
* 48 * , 807 ( 1983 ) ; m. del castello - mussot and l. j. sham , phys .
b * 31 * , 2092 ( 1985 ) .
j. l erskine and e. a. stern , phys .
b * 12 * , 5016 ( 1975 ) .
g. van der laan _
et al . _ ,
b * 34 * , 6529 ( 1986 ) .
lett . * 58 * , 737 ( 1987 ) .
p. carra and m. alterelli phys .
lett . * 64 * , 1286 ( 1990 ) . c. t. chen _
et al . _ ,
lett . * 75 * , 152 ( 1995 ) and references therein .
et al . _ ,
lett . * 66 * , 2595 ( 1991 ) . for atoms : b. t. thole _
et al . _ ,
* 68 * , 1943 ( 1992 ) , p. carra et al . , phys .
lett * 70 * , 694 ( 1993 ) , for solids : a. ankudinov and j. j. rehr , phys .
b * 51 * , 1282 ( 1995 ) .
h. ebert , p. strange , and b. l. gyorffy , j. appl .
phys . * 63 * , 3055 ( 1988 ) ; p. strange and b. l. gyorffy , phys .
b * 52 * , r13091 ( 1995 )
. g. y. guo _
b * 50 * , 3861 ( 1994 ) ; e. ebert , rep .
. phys . * 59 * 1665 ( 1996 ) .
brouder , m. alouani , and k. h. bennamann , phys .
b * 54 * , 7334 ( 1996 ) and references therein .
a. ankudinov and j. j. rehr , phys .
b * 56 * , 1712 ( 1997 ) .
n. v. smith , c. t. chen , f. sette , and l. f. mattheiss , phys .
b * 46 * , 1023 ( 1992 ) .
r. wu , d. wang , and a. j. freeman , phys .
lett . * 71 * , 3581 ( 1993 ) ; r. wu and a. j. freeman , phys .
lett . * 73 * 1994 ( 1994 ) .
j. schwitalla and h. ebert , phys . rev .
lett . * 80 * , 4586 ( 1998 ) . | the most popular electronic structure method , the linear muffin - tin orbital method ( lmto ) , in its full - potential ( fp ) and relativistic forms has been extended to calculate the spectroscopic properties of materials form first principles , i.e , optical spectra , x - ray magnetic circular dichroism ( xmcd ) and magneto - optical kerr effect ( moke ) .
the paper describes an overview of the fp - lmto basis set and the calculation of the momentum matrix elements .
some applications concerning the computation of optical properties of semiconductors and xmcd spectra of transition metal alloys are reviewed .
= cmmib10 = cmmib7 = cmmib5 = = = = cmbsy10 = cmbsy7 = cmbsy5 = = = * * # 1(#1 ) # 1(#1 ) # 1(#1 ) # 1(#1 ) # 1 # 1 # 1 # 1 # 1\{#1 } # 1\{#1 } # 1\{#1 } # 1\{#1 } # 1e^#1 # 1#2e^#1(#2 ) # 1#2e^#1 # 1#2e^#1 # 1#2e^#1 # 1#2e^#1 # 1#2e^#1[#2 ] # 1#2e^#1 # 1#2e^#1 # 1#2e^#1 # 1#2e^#1 # 1#2e^#1\{#2 } # 1#2e^#1 # 1#2e^#1 # 1#2e^#1 # 1#2e^#1 # 1(#1 ) # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1\{#1 } # 1 # 1 # 1 # 1 # 1,#2,#3y_#1#2(#3 ) # 1,#2,#3c_#1#2(#3 ) # 1,#2,#3y_#1#2(#3 ) # 1,#2,#3y_#1#2^(#3 ) # 1,#2,#3c_#1#2(#3 ) # 1,#2,#3c_#1#2^(#3 ) # 1,#2j_#1(#2 ) # 1,#2j_#1(#2 ) # 1,#2,#3j_#1(#2,#3 ) # 1,#2n_#1(#2 ) # 1,#2n_#1(#2 ) # 1,#2,#3n_#1(#2,#3 ) # 1,#2h_#1(#2 ) # 1,#2h_#1(#2 ) # 1,#2,#3h_#1(#2,#3 ) # 1,#2,#3k_#1(#2,#3 ) # 1,#2_#1e^i # 1#1^ # 1@xmath0@xmath0 # 1|#1|
# 1|-0.14em|#1|-0.14em| # 1#20.05em^#1 - 0.14em/-0.10em_#2 # 1#2-.15em # 1 # 2 # 1#2#3#4#5#6 ( # 1 # 2 # 3 # 4 | # 1 # 3 # 5 # 6 ) # 1#2#3#4#5#6 ( ) # 1,#2,#3,#4,#5,#6 ; ( -)^#3#1#6 ( 2#51 ) ^12 # 1#2#1#2 # 1#2#3 ( # 1 # 2 # 3 ) # 1#2#3 ( # 1 # 2 # 3 ) # 1#2#3#4 ( # 1|#2 # 3 # 4 ) # 1#2#3#4#5#6 # 1#2#3#4#5#6 ( # 1 , # 2 ; # 3 , # 4 ; # 5 , # 6 ) # 1#2#3(#1;#2;#3 ) # 1,#2,#3;#4;y_#1#2#3(#4 ) # 1,#2,#3;#4;_#1#2#3(#4 ) # 1,#2,#3;#4;c_#1#2#3(#4 ) # 1,#2,#3;#4;_#1#2#3(#4 ) # 1*#1 * # 1*#1 * # 1#1 h l o p s # 1#2d^#1 - 1.2pt#2 # 1#1 | # 1#1 | # 1#1 | # 1#1 | # 1| # 1 # 1| # 1 # 1| # 1 # 1| # 1 # 1#2#1 | # 2 # 1#2#1 | # 2 # 1#2#1 | # 2 # 1#2#1 | # 2 # 1#1 # 1#1 # 1#1 # 1#1
= 11 # 1@y to
= 12 # 1,#2;#3,#4;#5 ; _ # 1f_#2 ( | # 5 ) # 1,#2;#3,#4;#5,#6;#7;g^#1,#2_#3,#4 ( # 7 | ) # 1#1 # 1#2 ( # 1 | # 2 ) # 1#2 ( # 1 | # 2 ) # 1#2 ( # 1 | # 2 ) # 1#2 ( # 1 | # 2 ) # 1#2#1mod_#2 |
a considerable body of evidence from diverse sources leads to the conclusion that star - forming disc galaxies such as the milky way accrete @xmath0 of gas each year ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) , and have built up their observed discs gradually over the last @xmath1 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
what remains unclear is from what reservoir this gas is drawn , and how it enters the thin gas disc within which stars are formed .
a central question is the temperature of the reservoir : is this low enough ( @xmath2 ) for the reservoir to contain largely neutral gas , or comparable to the virial temperature , @xmath3 , of the gravitationally bound groups within which milky - way type galaxies currently reside ? in recent years there has been much enthusiasm for so - called cold - mode accretion of gas that has failed to be shock heated to the virial temperature @xcite .
cosmological simulations suggest that cold - mode accretion is the dominant process at redshifts @xmath4 , but gradually becomes less important . a powerful argument against its currently being
the dominant process is the persistent failure of 21-cm surveys to identify significant bodies of intergalactic hi in the nearby universe @xcite .
in particular the galaxy s `` high - velocity '' hi clouds , which as late as 1999 were argued to be distant and massive @xcite , are now known to be at distances @xmath5 and have masses @xmath6 too small for these clouds to be a cosmologically significant reservoir of gas @xcite .
@xcite already inferred from the presence of interstellar absorption in the spectra of high - latitude stars that the galactic disc must be embedded in pervasive medium of temperature @xmath7 .
the empirical case for such `` coronal gas '' was greatly strengthened by the copernicus and fuse missions , which detected ions such as ovion high - latitude sight lines to distant uv sources , in particular sight lines that pass close to high - latitude hi clouds @xcite .
the highly ionised gas detected must be collisionally ionised and is most readily interpreted as material at the interface between coronal gas with @xmath8 and clouds of cooler , partly neutral hi .
cosmology strongly suggests that galaxies should be embedded in virial - temperature coronae .
first , standard cosmology predicts that only a minority of baryons are contained in stars and cool interstellar gas ( e.g. * ? ? ?
* ; * ? ? ?
* ) . in rich clusters of galaxies the `` missing '' baryons
are directly detected through their x - ray emission @xcite . in lower - density environments , such as the local group , it is thought that the surface - brightness of x - ray emission is too low to be detected by current instrumentation @xcite .
second , three lines of argument indicate that star - formation is an inefficient process in which as much gas is ejected from a star - bursting system as is converted into stars : ( i ) we actually see winds blowing off star - forming discs @xcite ; ( ii ) the spectra of several quasars show blue - shifted absorption - line systems indicative of massive winds flowing away from the star - bursting host galaxy @xcite ; ( iii ) in clusters of galaxies of order half the metals synthesised by stars are in the intergalactic medium @xcite .
our premise in this paper is that most of the baryons originally associated with the milky way s dark matter comprise a corona of gas at the virial temperature , and that the gas that sustains star formation in the disc is drawn from this corona .
the question is how gas makes the transition from a pressure supported corona to the centrifugally supported disc . for more than thirty years
x - ray astronomers have studied the virial - temperature coronae of rich groups and clusters of galaxies .
most of these coronae have central cooling times that are significantly shorter than their ages , and it is natural to ask whether we can understand the milky - way s corona by extrapolating results for cluster `` cooling flows '' to lower masses .
it seems , however that the dynamics of these systems is qualitatively different from the dynamics of the milky way s corona because the central galaxies of rich clusters do not have massive stellar discs .
the extent to which gas in rich clusters is cooling ( as opposed to radiating ) is controversial , but it is now widely accepted that radiative losses by the inner corona are largely offset by mechanical feedback from the central black hole @xcite .
it is important to understand the origin of this qualitative difference in the dynamics of the coronae of star - forming galaxies of the `` blue cloud '' and `` green valley '' in colour - luminosity space @xcite and that of the coronae of massive galaxies within the `` red sequence '' .
sensitive hi observations of star - forming disc galaxies reveal that these galaxies keep @xmath9 to @xmath10 per cent of their hi a kiloparsec or more above or below their disc plane ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
most of this gas is thought to have been driven out of the disc by supernova - powered bubbles @xcite and must consist of clouds that are moving on essentially ballistic trajectories because if it were in hydrostatic equilibrium , its vertical density profile would be very much steeper than that observed by virtue of its low temperature @xcite .
the gas is expected to return to the disc within @xmath11 , so the phenomenon of extraplanar hi is evidence for galactic fountains @xcite .
hi observations of extraplanar gas in the milky way are hard to interpret on account of our location within the plane , but the leiden argentina
bonn ( lab ) survey @xcite , which mapped local hi emission with high sensitivity , is consistent with the milky way having a distribution of extraplanar hi that is similar to the distributions of hi studied around external nearby galaxies .
the extraplanar gas of nearby galaxies has two key properties : ( i ) the mean rotation speed of the gas declines quite rapidly with distance from the plane ; ( ii ) the gas shows net motion towards the galaxy s symmetry axis .
fraternali & binney ( 2006 ; hereafter fb06 ) and fraternali & binney ( 2008 ; hereafter fb08 ) fitted models of a galactic fountain to observations of ngc891 and ngc2403 and showed that these models can only account for the observed rotation rates and inward motion of the extraplanar gas if the observed hi clouds accrete gas such that the mass of a cloud exponentiates on a timescale @xmath12 .
such an accretion rate simultaneously accounts for data from the two rather different galaxies modelled and yields rates of accretion onto the discs , @xmath13 and @xmath14 that happen to be similar to the rates at which star - formation consumes gas in these galaxies .
the accreted gas is required to have significantly smaller specific angular momentum about the galaxy s symmetry axis than thin - disc gas .
fb08 suggested that the gas swept up by the fountain s clouds comes not from the corona but from cool streams embedded within it .
they were driven to this conclusion by the shortness of the time it takes coronal gas to flow past a cloud this time is very much shorter than the cooling time of coronal gas . in this paper we argue that notwithstanding the disparity between the cooling and flow times of the coronal gas , the interaction between a cold cloud and coronal gas leads to cooling of the coronal gas and accretion of it onto the star - forming disc . in section [ sec : theory ]
we give analytical arguments why cloud - corona interaction must lead to cooling of coronal gas rather than evaporation of hi . in section [ sec : simuls ] we present hydrodynamical simulations of the flow past a cloud , which confirm the analytic arguments . in section
[ sec : discuss ] we discuss the implications of these simulations both for further observations and the theory of galaxy formation .
section [ sec : conclude ] sums up .
[ fig : ctime ]
in model 2 of @xcite , the density of the corona decreases with radius as @xmath15 and at @xmath16 has electron density @xmath17 .
the corona is assumed to be isothermal with @xmath18 .
figure [ fig : ctime ] shows the cooling time of gas with metallicities @xmath19}=-1 $ ] and @xmath20 that is in pressure equilibrium with plasma of this temperature and the mean density of this corona between 8 and @xmath21 . at @xmath22 , the cooling time is @xmath23 .
once the temperature has fallen to @xmath24 , the cooling time has dropped by nearly forty to @xmath25 , which is less than the dynamical time at the solar galactocentric radius , @xmath26 . by the time the temperature has dropped by a further factor of two to @xmath27 ,
the cooling time has fallen by more than a further order of magnitude and is a mere @xmath28 .
thus although the cooling time of ambient gas in the lower corona is long , any diminution in temperature will dramatically shorten the cooling time . in the model of @xcite
the radiating plasma flows inwards at a rate @xmath29 .
the work done by compression offsets radiative losses , so the plasma temperature remains @xmath30 .
the mass of an hi halo is rather ill - defined : in an external galaxy it is not clear at what value of @xmath31 we should place the boundary between disc and halo hi , and in the galaxy there is a similar ambiguity in the line - of - sight velocity that divides halo from disc hi .
if in ngc 891 the disc - halo boundary is conservatively place at @xmath32 , the mass of the hihalo is @xmath33 ( fb06 ) .
@xcite conclude that @xmath34 per cent of the galaxy s hi is halo gas , and that the total himass within @xmath21 ( the region within which most star formation occurs ) is @xmath35 , so the mass of the hi halo within @xmath21 is @xmath36 . the hi halo is largely confined to the region @xmath37 , so we compare the mass of the hi halo at @xmath38 to the mass of coronal gas in the cylindrical annulus @xmath39 , @xmath37 .
the coronal mass is @xmath40 , so within this volume there is over twice as much hi as coronal gas , although the coronal gas will occupy nearly all the space . the clouds that make up the hi halo typically take @xmath41 to travel from their launch point in the plane through the halo and back to the plane .
consequently , in one gyr @xmath42 of hi is passed through the @xmath43 of gas in the lower corona .
hence the temperature of the coronal gas would be halved if just 4.5 per cent of the gas that passes through the corona in a gyr were to mix with the coronal gas .
this drop in temperature would bring the cooling time of the coronal gas down to @xmath44 , the local orbital time .
the hi clouds of the halo plough through the corona at speeds @xmath45 that are less than but comparable to the sound speed of the coronal gas . consequently the flow around these clouds is likely to be in a high reynolds number regime , and each cloud must be decelerated by ram pressure of order @xmath46 , where @xmath47 is the coronal density . in these circumstances
the cloud loses its momentum on a timescale equal to @xmath48 times the time @xmath49 for the cloud to move its own length ( fb08 ) .
clouds have diameters of a few tens of parsecs and travel a few kpc , while @xmath50 .
hence clouds must surrender a significant fraction of their momentum to coronal gas @xcite .
above we showed that in a gyr @xmath51 of hi passes through the @xmath52 of the lower corona .
clearly the coronal gas can not absorb a significant fraction of the momentum of more than 20 times its mass of hi ( fb08 ) .
moreover , if the hi disc were losing angular momentum to the corona at the rate this calculation implies , it would be contracting on a gyr timescale .
the natural resolution of these problems is that the disc _ accretes _ most of the coronal gas that hi clouds encounter .
then the momentum lost by hi clouds would be returned to the disc , and it would not build up in the corona .
as was mentioned in section 1 , coronal gas in dark - matter halos more massive than those of spiral galaxies has been extensively studied for four decades .
cooling within these halos shows no tendency to produce a cold stellar disc
the coldest gas is near the centre , where the cooling time is shortest , and whatever gas cools out of the corona feeds the central black hole rather than forming a star - forming disc . the argument of the last subsection may help us to understand why in less massive dark halos cooling coronal gas flows into the disc rather than onto a central black hole : so long as the halo has a star - forming disc , that disc sustains its star formation by reaching up and grabbing coronal gas .
discs are disrupted by major mergers , which occur rather frequently .
analytic arguments and hydrodynamical simulations of cosmological clustering suggest that when the disc of a relatively low - mass galaxy is disrupted by a major merger , it will quickly re - form from filaments of cold inflowing gas ( e.g. * ? ? ?
thereafter it will sustain itself by grabbing coronal gas . when a more massive galaxy experiences a major merger , the disrupted star - forming disc is less likely to re - form .
if it does not form from cold inflowing gas , it will not form from coronal gas , because in the absence of a star - forming disc , catastrophically cooling coronal gas will feed the central black hole and reheat the corona .
thus star - formation permanently ceases if there is insufficient cool gas to re - form a star - forming disc after a merger . in this picture
it is likely that cooling at the centre of the corona of a disc galaxy leads to episodic reheating of the central corona , just as in a classical cooling flow .
that is , we suggest that the corona of a star - forming disc galaxy accretes onto central black hole _ as well as _ onto the disc . for this to be a viable proposal , the central reheating associated with accretion by the black hole
must not undermine the ability of the star - forming disc to grab coronal gas from the part of the corona that lies above it
. this condition could be satisfied if the agn outburst were sufficiently small and sufficiently directed perpendicular to the galactic plane .
this is a topic for a later paper , however . as a cloud moves through the ambient coronal gas , turbulence in the boundary layer at the interface of the hot and cold fluids
must cause gas to be stripped from the cloud at some rate . to calculate this rate from ab - initio physics is a daunting task because one would have to consider plasma instabilities in addition to hydrodynamical ones such as the kelvin - helmholtz instability , and turbulent energy will be cascading to very small scales . in view of these difficulties ,
the obvious way forward is to parametrise the problem by hypothesising that mass is stripped from the a cloud of mass @xmath53 at a rate @xmath54 .
a turbulent wake of stripped gas will run back through the corona from the moving cloud . in this wake turbulence will mix the stripped gas with ambient gas . a key quantity is the cooling time of the plasma that results from this mixing .
we estimate this under the assumption that mixing occurs so quickly that radiative losses during mixing can be neglected .
let @xmath55 be a short length of the wake , which has cross - sectional area @xmath56 and originally contained a mass @xmath57 of coronal gas .
in the time @xmath58 that it took the cloud to pass through the length , a mass @xmath59 of gas was stripped from it .
after the cold gas has mixed and come into thermal equilibrium with the coronal gas , the temperature of the resulting fluid is @xmath60 where @xmath61 and @xmath62 are the coronal and cloud temperatures , respectively .
the natural unit for the cross - sectional area of the wake is the characteristic cross - section @xmath63 of the cloud .
we write @xmath64 , where @xmath65 .
thus @xmath66 we simplify this expression under the assumption that the cloud is in approximate pressure equilibrium with the corona , so @xmath67 , and have @xmath68 since @xmath69 is the fraction of the cloud s mass that is stripped in the time taken for the cloud to travel its own length , and @xmath65 , the second term in the numerator of equation ( [ eq : trat ] ) must be small compared to unity and may be neglected .
the second term in the denominator is larger by a factor @xmath70 , so @xmath71 may be significantly lower than @xmath61 . on account of the steepness of the cooling - time curve plotted in fig .
[ fig : ctime ] , this result implies that the cooling time in the wake may fall below the ambient cooling time by a factor of several . from this back - of - envelope calculation
we draw the following conclusions * material in the wake will become hi on the timescale that it takes the parent hi cloud to fly its trajectory if the mass - loss rate @xmath72 exceeds the critical value @xmath73 for @xmath74 this condition is that in the time taken to travel its own length the cloud lose at least a fraction @xmath75 of its mass . *
if the mass - loss rate of clouds falls below this critical value , what mass is stripped from the cloud will be integrated into the corona .
this will lower the cooling time of the ambient corona , but not lead to prompt accretion of the wake onto the disc .
this drop in the temperature of the corona near the disc will lower the critical mass - loss rate required for subsequent wakes to cool promptly . * our estimate of @xmath76
was obtained under the assumption that we can neglect cooling during mixing .
although the mixing process is likely to be fast , the cooling rate is extremely large at temperatures that lie within a factor 30 of the cloud s temperature .
hence it is likely that a more exact calculation would produce a lower estimate of @xmath76 .
we have also neglected compression of the coronal gas as it flows around the cloud .
however , the effect of compression on the cooling time of ambient gas is unclear because , while the cooling time decreases with increasing density at constant temperature , it increases with temperature , and compression will be associated with a ( largely adiabatic ) rise in @xmath77 . *
the effective value of @xmath78 is uncertain . on one hand ,
as the simulations will show , the cloud flattens in its direction of motion , making @xmath79 . on the other hand ,
the distribution of cloud material is likely to be concentrated in a network of thin sheets .
the effective value of @xmath78 for an individual sheet could be small .
* dimensional analysis indicates that @xmath72 will lie near @xmath76 : the rate of stripping is essentially determined by the rate at which coronal gas hits the leading surface of the cloud and strips a comparable mass from the cloud .
quantitatively , @xmath80 , where @xmath81 .
when we eliminate @xmath82 and @xmath47 in favour of @xmath53 and @xmath61 we find that the characteristic mass - loss parameter is @xmath83 * if we take the mass - loss rate to be given by the previous item , we find that the mass of a cloud that will completely mix with the corona after travelling distance @xmath84 is @xmath85 equivalently , the distance travelled prior to destruction is predicted to be @xmath86 times the size of the cloud .
numerical simulations of a cloud of gas at @xmath87 moving through coronal gas at an initial speed @xmath88 will illustrate these points and give insight into both typical mass - loss rates and the critical mass - loss rate required for prompt cooling of the wake .
the simulations are idealised in that they neglect gravitational acceleration and the variation of the coronal density along the cloud s trajectory .
however , the galactocentric radius of the clouds varies by @xmath89 per cent , so the density variation is not extreme .
the range of relevant velocities depends on the extent to which the corona corotates with the disc , which is currently unclear .
the velocity we have chosen to explore is at the low end of the relevant range because it is of order the velocity at which clouds have to be ejected from the disc in order to reach heights of a few kiloparsecs .
larger velocities would clearly lead to more rapidly interaction with the corona than that explored here .
it is important to be clear about what can and can not be learnt from the simulations .
first any simulation incorporates limited physics and limited spatial resolution . in the real world
the fluid is a magnetised , almost perfectly conducting , largely collisionless plasma
. the shear flow at the cloud - corona interface will draw out and strengthen the field lines originally in the plasma .
this drawing out will make the velocity distribution of particles anisotropic , which will excite plasma instabilities .
these instabilities will both heat and mix the plasma . by contrast the tendency of the field lines to follow stream lines will strongly inhibit mixing . none of this complex physics is included in our simulations . instead , in the simulations
ablation and mixing are driven by the kelvin - helmholtz instability , which gives rise to ripples in the fluid interface .
these ripples develop into vortices , which shed smaller vortices , which themselves shed vortices . in a numerical simulation numerical viscosity
prematurely truncates this hierarchy of vortices on a scale of a few times the grid spacing @xmath90 . on scales finer than @xmath90 , the fluid
is represented as perfectly mixed , whereas it is in reality a roughly fractal foam of high- and low - density regions .
the condition for the simulations to be a reliable guide to the large - scale structure of the flow is that the large - scale dynamics of this foam is equivalent to the dynamics of the locally homogeneous fluid that is actually represented .
this is a reasonable proposition , but we have to expect that the values of the flow s macroscopic parameters , such as ambient pressure and cooling rate , at which a simulation of given resolution most closely approximates reality are likely to vary with resolution .
that is , we anticipate a phenomenon analogous to renormalisation in quantum field theory , where the appropriate value of the bare electron mass , for example , is a function of the largest wavenumber summed over .
the simulations do not include thermal conduction . on the smallest scales conduction must play an important role in homogenising a mixture of cloud and coronal gas , and in the simulations this role is effectively covered by numerical mixing and diffusion .
consequently , our failure to model thermal conduction is most worrying on intermediate and larger scales . in a magnetised plasma heat
is largely conducted along field lines .
the field lines in upstream ambient gas are inevitably disconnected from field lines in the cloud .
any connection between these sets of field lines must occur in the turbulent wake .
this fact is likely to severely limit the effectiveness of thermal conduction .
the physical properties of the clouds and the corona considered here are such that the clouds would be stable against conductively - driven evaporation if they were stationary ( see * ? ? ?
* ) , so it is very unlikely that conduction plays an important role in the energetics of our problem . in principle conduction lowers the rate at which a moving cloud is ablated , by damping the kelvin - helmholtz instability @xcite , but this effect will be unimportant if conduction is magnetically suppressed to values small compared to the spitzer or saturated values . in light of these remarks , the aims of the simulations are as follows * to estimate the mass - loss rate @xmath72 for comparison with @xmath91 .
it is advantageous to do this in the absence of radiative cooling for then ( a ) the calculations are slightly faster , and ( b ) one can identify the cloud with gas at @xmath92 since any ablated material will be heated to and remain at higher temperatures .
we will show that the mass - loss rate is reasonably independent of @xmath90 . * to show that for any given metallicity of the gas , there is a critical ambient pressure @xmath93 above which the mass of cool gas increases with time through condensation in the wake and below which the wake tends to evaporate .
we shall find that although our values of @xmath93 vary with both metallicity and @xmath90 , they lie within the range of values that occur in practical cases . hence it is plausible that the true value of @xmath94 lies below the actual ambient pressures , so real wakes give rise to condensation and accretion .
@xmath95 $ ] & @xmath96 $ ] & @xmath97 $ ] & grid size + + z_0 & 25 & no met .
& @xmath98 & @xmath99 & @xmath100 & @xmath101 & c , m , f + z_1 & 25 & -3.0 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & c , m , f + z_2 & 25 & -2.0 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & c , m , f + z_3 & 25 & -1.5 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & c , m , f + z_4 & 25 & -1.0 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & c , m , f + z_5 & 25 & -0.5 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & m , f + z_6 & 25 & 0.0 & @xmath98 & @xmath102 & @xmath103 & @xmath104 & m + + t_4_z_3 & 25 & -1.5 & @xmath98 & @xmath105 & @xmath100 & @xmath106 & m + t_4_z_4 & 50 & -1.0 & @xmath98 & @xmath105 & @xmath100 & @xmath106 & m + t_3_z_4 & 50 & -1.0 & @xmath98 & @xmath105 & @xmath107 & @xmath108 & m + the parameters of the simulations are listed in table [ tab : simuls ] . in all simulations
the initial cloud velocity was @xmath109 and the initial radius was @xmath110 .
the temperature of the corona is restricted to a narrow range around @xmath111 by the requirement that the corona be bound to the galaxy and yet be extensive enough to contain a cosmologically significant mass ( * ? ? ?
* e.g. ) , so in all simulations we made the corona s temperature @xmath111 . in simulations of the low - pressure sequence the total particle density of the corona was @xmath112 ( except in the z_6 simulation in which it was reduced to @xmath113 ) , implying @xmath114 , while in the high - pressure simulations it was twice as great .
these values are both lower than the density @xmath17 at @xmath115 in model 2 of @xcite , and may be compared with the total particle density @xmath116 at @xmath16 above the plane adopted by @xcite . in all but two simulations
the initial cloud temperature was @xmath117 ; in the last two ( high - pressure ) simulations the cloud temperature was lowered to @xmath118 since at the higher temperature ( and therefore lower density contrast ) the cloud was totally disrupted by @xmath119 .
the cloud mass ranged from @xmath104 to @xmath120 depending on the pressure of the corona and the temperature of the cloud .
the jeans mass of the standard cloud ( @xmath121 pc , @xmath122 k ) is @xmath123 so our neglect of self gravity is amply justified .
the calculations were performed on two - dimensional , cartesian grids of three sizes : @xmath124 ( c ) , @xmath125 or @xmath126 for simulations run to @xmath119 ( m ) , and @xmath127 ( f ) .
ghost cells are used to ensure that the pressure gradient vanishes at the grid boundaries .
every configuration was simulated twice , once with cooling on and once with cooling off .
when radiative cooling is permitted , it follows the prescription of @xcite .
the metallicity of the cloud is always the same as that of the ambient medium , and is varied from zero up to solar .
we used the eulerian code echo , which is flux - conserving and uses high - order shock - capturing schemes ; a detailed description and tests of it can be found in @xcite . since one of the dimensions perpendicular to the cloud s velocity has been suppressed , we are in effect simulating flow around an infinite cylindrical cloud that is moving perpendicular to its long axis .
the cylinder initially has a circular cross section of radius @xmath82 . from the simulations we obtain quantities per unit length of the cylinder .
we relate these to the corresponding quantities for an initially spherical cloud of radius @xmath82 by multiplying the cylindrical results by the length @xmath128 within which the mass of the cylinder equals the mass of the spherical cloud .
we first study the simulated flows in the absence of radiative cooling , in order to assess the importance of the limited resolution of the simulations .
figure [ fig : tplot ] shows the temperature distribution on the grid after @xmath129 and @xmath119 with radiative cooling turned off .
the cloud has been flattened into a pancake by ram pressure from the medium it is moving into .
the shear flow over the leading face of the pancake is causing vortices to be shed from the pancake s edges that are analogous to the vortices shed by an aeroplane wing . in the highly turbulent wake behind the cloud
, the temperature fluctuates around @xmath130 depending on the fraction of the gas in each cell that comes from the cloud rather than the corona .
figure [ fig : nocoolablate ] quantifies the speed at which ablation reduces the cloud s mass by plotting the mass of gas below @xmath24 versus time for simulations of three resolutions .
we see that over @xmath129 @xmath131 of gas ( @xmath132 per cent of the cloud s mass ) are heated to above @xmath24 , regardless of the grid resolution ; it seems that even the coarsest grid has sufficient resolution to model satisfactorily the stripping of gas from the leading edge of the cloud . in the absence of radiative cooling ,
any gas that is stripped from the cloud will eventually be heated to above @xmath24 as it mixes with coronal gas .
the details of the smallest vortices involved in the mixing process are resolution - dependent in the sense that at higher resolutions some gas remains cold for slightly longer before numerical mixing on the grid scale eliminates it .
the cloud s mass - loss rate is resolution - independent because all stripped gas will be heated within a couple of large - scale eddy - turnover times , regardless of resolution . the mass - loss rate increases gently with time because it depends on the area of the cloud s leading face , which increases with time as the cloud is squashed into a thinner and thinner pancake , a process that is apparent in fig .
[ fig : tplot ] . the dotted straight lines in fig .
[ fig : nocoolablate ] show the critical mass - loss rate defined by equation ( [ eq : defsac ] ) for @xmath133 and @xmath134 .
we see that , as predicted above on dimensional grounds , the measured mass - loss rate lies near @xmath76 .
actually the simulations must underestimate @xmath135 because mass is lost from the cloud s edges , which have total length @xmath136 in cylindrical geometry and @xmath137 in the spherical case .
thus we expect the numerical rates to be @xmath138 of the true rate and in three dimensions @xmath78 would lie close to unity .
these results enable us to estimate the minimum size that a cloud must have if it is to survive a typical passage through the corona . from fig .
10 of fb06 we have that trajectories last @xmath41 , and in this time a cloud travelling at @xmath139 will move @xmath140 . setting @xmath141 and @xmath142 in equation ( [ eq : mofl ] )
we find that clouds with masses less than @xmath143 will completely mix with the corona before they can return to the plane .
it is important to know how the mass - loss rate scales with cloud mass and thus cloud radius . since mass is lost from the leading surface of the cloud , we would expect the absolute mass - loss rate to scale as @xmath144 and therefore the specific mass - loss rate @xmath145 .
we have verified this dependence by simulating the evolution of a cloud of the standard density but radius doubled to @xmath146 . when radiative cooling is switched on , the strength of the dependence of the cooling rate upon @xmath77 that is apparent in fig .
[ fig : ctime ] , substantially increases the difficulty of the simulations and the uncertainties surrounding their results because now the structure of the turbulent wake is crucial .
indeed , evaporation is favoured over cooling by more effective dispersal of stripped material through a large volume of the corona .
the higher the resolution delivered by the code , the greater the dynamic range of the hierarchy of vortices , and the more effective is the dispersal of stripped material , with the consequence that increased resolution favours evaporation over condensation .
each panel of fig .
[ fig : cool-0 ] is the analogue of fig .
[ fig : nocoolablate ] but with radiative cooling turned on ; in the three panels the metallicity increases from zero ( top panel ) through @xmath147}}=-1 $ ] ( middle panel ) to @xmath20 ( bottom panel ) .
the simulations with [ fe / h ] @xmath148 -1 do not differ significantly from that at zero metallicity . in all three panels the mass of cool gas starts by increasing rather than decreasing , and the rate of increase naturally increases with metallicity .
when @xmath147}}=-0.5 $ ] , the increase continues throughout the @xmath129 simulated , but at lower metallicities the mass of cool gas eventually starts to decrease . as anticipated
, the results are unfortunately dependent on numerical resolution : in general there is a tendency for the mass of cold gas to decrease as the resolution increases , although the medium and high - resolution simulations give rather similar results .
figure [ fig : cool - hp ] shows the effect of doubling the particle density of the corona to @xmath149 : even at metallicity @xmath147}}=-1 $ ] the mass of cool gas now increases throughout the @xmath119 simulated , and in fact in the latter half of the simulation cold gas accumulates at an accelerating rate .
this result should be contrasted with that shown by the middle panel of fig .
[ fig : cool-0 ] , which shows that at the same metallicity but half the density , the mass of cool gas starts to decrease after @xmath150 .
if we extrapolate the behaviour shown in fig .
[ fig : cool - hp ] to the whole milky way halo , assuming as its total mass the value given in sect .
2.1 , we obtain a global accretion rate of @xmath151 .
figure [ fig : cooltplot ] shows the temperature distribution at @xmath152 in a simulation with cooling of plasma with @xmath147}}=-0.5 $ ] and ambient particle density @xmath112 . comparing this figure with the upper panel of fig .
[ fig : tplot ] we see that cooling makes the wake longer and less laterally extended .
we have also run simulations with lower values for the coronal gas density and found that accretion is still present for @xmath153 provided that the metallicity is about solar .
figure [ fig : x - vx ] shows for the high - pressure simulation t_3_z_4 the distribution of cool gas along the cloud s direction of travel ( top panel ) and its distribution in velocity along the same direction ( bottom panel ) at the start of the simulation ( black curve ) and after @xmath119 when radiative cooling is ( blue curve ) or is not ( red curve ) included . from the upper panel
we clearly see the effectiveness of cooling in enhancing the mass of cool gas .
we also see that @xmath154 percent of the cold gas lies more that @xmath155 behind the cloud . in the lower panel around @xmath109
we clearly detect the deceleration of the main body of the cloud , but more striking is the width of the velocity range over which small amounts of cool gas are distributed .
a few times @xmath156 is accelerated to higher velocities than the cloud s .
a slightly larger mass of gas is decelerated to negative velocities .
this velocity distribution implies that gas circulates around vorticies at speeds that are comparable to the speed of the cloud s forward motion .
.[fig : vwidth ] figure [ fig : vwidth ] shows the distribution of cold gas in @xmath157 , the direction of travel , and in @xmath158 for the high - pressure simulation t_3_z_4 with and without cooling .
turning on cooling greatly broadens the extent in both @xmath157 and @xmath158 of low - density cool gas . in both simulations
, the region of highest gas density runs from @xmath159 at the cloud s location down to @xmath160 @xmath161 behind the cloud .
thus hydrodynamics leads to a steep velocity gradient , @xmath162 in the part of the wake that will dominate 21-cm emission .
@xcite assume that a gradient in the measured line - of - sight velocity along complex a , @xmath163 , is entirely produced by the galaxy s gravitational field .
in light of our simulations this assumption should be treated with some caution .
the lower panel of fig .
[ fig : vwidth ] shows that the velocity width of the wake increases with distance from the cloud for @xmath164 from the cloud as a result of the lower envelope ( velocities in the opposite direction to the cloud s motion ) moving downward towards @xmath165 , while the upper boundary remains at @xmath166 .
this is further evidence that turbulent eddies impart peculiar velocities comparable to the cloud s velocity . just behind the cloud material
has to flow faster than the cloud in order to flow into the space vacated as the cloud moves on .
further back similar vortices carry cold gas away from the cloud with similar velocities . in the region @xmath167
to @xmath168 behind the cloud , the turbulence damps quite rapidly . from these simulations we conclude the following .
* there is clear evidence that the simulations limited resolution is adequate for determining the rate at which gas is stripped from a cloud , although the two - dimensional nature of the simulations suggests that they will underestimate the stripping rate by a factor @xmath132 . *
the stripping rate tends to increase with time as a result of the cloud being flattened by ram pressure . *
the stripping rate is best determined from simulations in which cooling is turned off , and it is found to lie close to the rate expected on dimensional grounds . in
[ sec : theory ] we saw that this rate is very similar to the critical rate at which cooling takes over from evaporation . * whether stripped gas is evaporated or leads to the condensation of coronal gas depends on the structure of the turbulent wake .
consequently the finite spatial resolution of the simulations makes it impossible to determine with confidence the combinations of pressure and metallicity which divide evaporation from condensation .
however the simulations confirm analytic arguments , which imply that for a given cloud there is a critical path in the space spanned by the coronal pressure and metallicity which divides situations in which , at high pressure or metallicity , coronal gas condenses in the wake and those in which , at lower pressure or metallicity , the gas stripped from the cloud is evaporated by the corona . both the simulations and
analytic arguments suggest that the parameters of the galactic corona lie close to this critical path .
* the neutral gas that trails the cloud is strongly influenced by hydrodynamics forces and can not be considered to be on an orbit .
there is a large velocity gradient along its high - density ridge in velocity - position space . within @xmath169 from the cloud
the velocity width of the stream is of order of the cloud s velocity .
the picture developed here of the connection between the galactic fountain and accretion onto the star - forming disc differs materially from that proposed by fb08 : in that paper hi clouds grew in mass as they moved through the corona ; here each cloud loses mass , but the mass of cold gas in a cloud and its wake taken together increases with time .
a natural question is whether the present picture predicts a similar overall picture of 21-cm emission to that proposed by fb08 .
if it does , it will fit the data reasonably well .
fb08 fitted 21-cm data cubes for ngc891 and ngc2403 .
individual clouds are for the most part not resolved in these data , so the observed emission is due to many superposed clouds . in the fb08 model ,
each cloud places a blob of emission in the cube , centred on its sky - position and line - of - sight velocity , and smeared by the angular and velocity resolution of the survey .
if the picture developed here is correct , the emission of each cloud is extended in velocity and is elongated on the sky .
very sensitive 21-cm data would be required to see the full extent of the trail of an individual cloud the antenna temperature will drop by two orders of magnitude within @xmath161 and @xmath170 of the cloud s position and velocity . however , the integrated emission from many individually undetectable trails must contribute to the hi `` beards '' of nearby galaxies . in a forthcoming study
we will modify the pseudo - data cubes of fb08 to include the extended emission from the trails of clouds .
we suspect that these improved data cubes will have similar observable properties to those of fb08 because the velocity centroid of each cloud s total emission will be unchanged .
clearly it is to surveys of our own galaxy that we must turn for evidence that clouds have hi - rich wakes .
maps of the 21-cm emission of individual high - velocity clouds generally display a tadpole - like structure :
the cloud is elongated and the point of highest surface brightness lies towards one end ( e.g. * ? ? ?
* ; * ? ? ?
* ) . in the scenario proposed here ,
tadpole - like structures can be explained by the interaction between the cloud and the ambient ( coronal ) medium .
this interaction leads to the formation of trailing material behind the cloud . even in our galaxy
most emission from extraplanar gas is unresolved in the sense that along any direction in the lab survey , emission is detected over a wide band in velocity around 0 .
however , a scan through the data cube , one heliocentric velocity at a time , reveals numerous elongated structures along which there is a systematic trend in velocity
. these could well be the wakes of relatively massive clouds .
clouds with masses well below the threshold for detection of their 21-cm emission can be detected through the absorption lines to which they give rise in the ultraviolet spectra of background sources .
absorption - line studies suggest that large hi complexes are associated with numerous small hiclouds @xcite
. could these small clouds be the knots of cold gas visible in the wake of fig .
[ fig : cooltplot ] ? if this interpretation is correct , the small clouds would be found only on one side of the large complex , and they would have a mass spectrum that was restricted to masses very much smaller than that of the complex . in the simulations the metallicity of the cloud is the same as that of the corona , whereas real - world clouds will be more metal - rich than the corona by a factor up to 10 . whether gas condenses or evaporates depends on the cooling rate of gas that is roughly a 5050 mixture of gas stripped from the cloud and coronal gas , so the most realistic simulations are those in which the universal metallicity is about half that of real clouds ; that is the simulations with @xmath147}}\sim-0.5 $ ] .
we must obviously ask how valid a guide two - dimensional simulations will be to the real , three - dimensional problem . reducing the dimensionality of the problem
reduces the number of high - wavenumber modes relative to the immediately driven low - wavenumber modes , thus making the power - spectrum of turbulence less steep .
this argument suggests that turbulent mixing will be more effective in two dimensions than in three , with the consequence that our simulations have a tendency to over - estimate the pressure or metallicity of the transition from evaporation to condensation .
however , this conclusion must be considered very tentative at this time .
the ablation of clouds that move through a low - density medium has been discussed by , among others , @xcite , @xcite , @xcite and @xcite .
in addition to studying the ablation of pressure - bounded clouds as here , some of these authors have considered also gravitationally bound clouds and included thermal conduction in addition to radiative cooling .
@xcite find that thermal conduction stabilises a moving cloud by reducing the magnitude of the velocity gradient in the boundary layer where the cloud meets the ambient medium ; a smaller velocity gradient leads to slower growth of the kelvin - helmholtz instability .
most previous simulations assume the flow to be axisymmetric around the velocity of the cloud s motion . in such simulations material
becomes trapped on the assumed symmetry axis , where the radial velocity must vanish by symmetry .
@xcite also simulated clouds with the symmetry assumed here and found that the main results were independent of the adopted geometry .
@xcite found ablation to proceed faster than we do and interpreted their ablation timescale as the inverse of the kelvin - helmholtz growth rate , which is larger than @xmath91 by @xmath171 .
the difference between their ablation rate and ours probably arises from their use of a different criterion for identifying cloud gas : they took this to be the mass within an appropriate sphere , whereas we have defined cloud gas to be cool gas , regardless of where it resides . in other respects our results are in accordance with earlier findings .
the simulations most comparable to ours are those of @xcite , who simulated the ablation of cool clouds that fall towards the disc through the corona .
consequently , their characteristic coronal density , @xmath172 , was a factor 2 to 4 lower than ours .
their simulations used a three - dimensional grid , so their grid spacing was coarser and they could not provide evidence of numerical convergence .
their clouds , which had similar initial masses to ours , fragmented in a similar fashion . in most simulations
the total mass of hiin the computational volume declined with time , although in some simulations an upturn in the hi mass is evident at late times as the cloud approaches the plane .
our numerical results are entirely consistent with theirs , although our physical motivation is different and , crucially , our parameter regime extends to the higher coronal densities expected near the plane .
recent attempts to model extraplanar hi include those of @xcite and @xcite .
@xcite investigated the equilibria of differentially rotating distributions of gas in flattened gravitational potentials .
they showed that by making the specific entropy and angular momentum of the gas vary appropriately within the meridional plane , the kinematics of the gas can be made consistent with the data for hi around ngc891 .
they noted that dynamical equilibrium required the gas to be too hot to be neutral , so the gas within their model could not be the observed as hi itself , but suggested it might be coronal gas within which hi clouds were embedded as almost stationary structures ( see also * ? ? ? * ) .
@xcite studied the settling of hot gas within the gravitational potentials of spherical galaxy - sized dark halos .
the gas was initially spinning with a rotation velocity that was independent of radius ( so specific angular momentum @xmath173 ) . a prompt cooling catastrophe caused cold gas to accumulate in a centrifugally supported disc , and clouds of cold gas
were subsequently seen to be falling through hotter gas towards this disc .
this simulation is interesting but one has to worry that the cool clouds are numerical artifacts .
it is well known that standard smooth - particle hydrodynamics artificially stabilises contact discontinuities @xcite .
moreover , observed high - velocity clouds all have masses that lie below the resolution limit of the simulations of kaufmann et al .
, so the clouds they see have no counterparts in reality . finally , @xcite showed that unless the specific entropy profile of coronal gas is unexpectedly flat , a combination of buoyancy and thermal conductivity ( which was not included in the kaufmann et al .
simulations ) suppresses thermal instability .
there is abundant evidence that in galaxies like ours , star - formation powers a fountain that each gigayear carries @xmath174 of hi to heights in excess of @xmath155 above the plane .
several lines of argument strongly suggest that galaxies like ours are surrounded by gas at the virial temperature coronae .
the density of the corona , which must vary with position , is very uncertain , especially in the region above the star - forming disc .
however , there are indications that this density is such that the local coronal cooling time is of order a gigayear .
at least half of the baryons in the universe are believed to reside in coronae and their extensions to intergalactic space .
models of the chemistry and stellar content of the galactic disc require the disc to accrete @xmath175 of low - metallicity gas .
the corona is the only reservoir of baryons that is capable of sustaining an infall rate of this order for a hubble time .
therefore there is a strong prima - facie case that star formation in the disc is sustained by cooling of coronal gas .
the dynamical interaction of hi clouds of the fountain with coronal gas is inevitable . for any plausible coronal density
, the ram pressure arising from motion through the corona leads to non - negligible loss of momentum by fountain clouds .
if this momentum were retained by the corona rather than returned to the disc , the corona would rapidly become rotation - dominated .
we have not pursued this possibility because we think it is more likely that coronal gas that absorbs momentum from fountain clouds is shortly thereafter accreted by the disc .
we suggest that the absorption proceeds as follows : ( i ) coronal gas strips gas from the leading edge of the cloud as a result of kelvin - helmholtz instability ; ( ii ) in the turbulent wake of the cloud , the stripped gas mixes with a comparable mass of coronal gas ; ( iii ) as a result of this mixing the cooling time becomes shorter than the cloud s flight time and coronal and stripped gas together form knots of hi that trail behind the cloud and fall onto the disc within a dynamical time .
this scenario is suggested by a combination of analytic and observational arguments , and supported by hydrodynamical simulations .
analytic arguments imply that for a given coronal pressure and metallicity there is a critical rate of mass loss by a cloud , @xmath76 , such that at lower mass - loss rates , stripped gas will be evaporated by the corona , and the total mass of hi will decrease during a cloud s flight .
by contrast , when the mass - loss rate exceeds @xmath76 , stripped gas will lead to condensation of coronal gas , so the mass of hi increases over time .
dimensional arguments suggest that the actual mass - loss rate must lie close to @xmath76 .
we have used grid - based hydrodynamical simulations of the flight of a cloud to check the analytic arguments . any hydrodynamical simulation
is severely limited by its finite spatial resolution .
however , we present evidence from simulations in which radiative cooling has been switched off that our simulations have sufficient resolution to provide reliable estimates of the mass - loss rate .
this rate is such that clouds with masses @xmath176 ( eq . [ eq : mcrit ] ) will be totally disrupted before they return to the disc .
simulations that include radiative cooling confirm the existence of a critical mass - loss rate @xmath76 that depends on coronal pressure and metallicity in the expected manner . on account of the restricted resolution of our simulations
, we can not give a definitive value for @xmath76 within the local corona .
however , we can argue that if the critical mass - loss rate lay above the actual mass - loss rates , the coronal density and metallicity would rise secularly at the expense of the star - forming disc . as a consequence , the @xmath76 would decrease until it fell below the actual mass - loss rate .
that is , the conditions at the base of the corona have a tendency to adjust until they lead to accretion by the disc .
although the present model differs materially from that of fb08 in that we envisage the masses of clouds decreasing rather than increasing over time , it seems likely that when the model is used to simulate data cubes for the galaxies studied by fb08 , similar agreement with the data will be achieved .
this is because fb08 correctly give the dependence on time of both the mass of hi associated with a given cloud , and its velocity centroid .
however , a revision of the fb08 models is required to test this conjecture .
the simulations make detailed predictions for what should be seen in studies of the high - latitude hi distribution in the galaxy .
trails behind clouds should show large gradients in mean velocity that are dominated by hydrodynamical rather than gravitational forces .
high - sensitivity data should reveal small quantities of gas distributed around the mean velocity by of order the cloud s velocity .
the trail should be studded by knots of cold gas .
the idea that star - forming discs reach up into the surrounding corona and grab the relatively pristine gas required to sustain their star formation for a hubble time , makes a good deal of sense cosmologically by explaining how discs can remain star - forming as long as they are not disrupted by a major merger . if after a major merger there are significant streams of cold gas , this gas can seed a new star - forming disc , but in the absence of cold seed - gas , star formation ceases because coronal gas can only condense onto a disc that is already star - forming .
@xcite argued that thermal evaporation by coronal gas of filaments of cold gas determines whether a central cusp reforms when an early - type galaxy experiences a major merger .
similarly , when a late - type galaxy experiences a major merger , thermal evaporation of filaments of cold gas can prevent the gas disc reforming . as a halo proceeds up the clustering hierarchy , the effectiveness of thermal evaporation increases , and at some point a gas disc is prevented from forming after a major merger . following this event
, there is a dramatic reduction in the rate at which coronal gas cools , and the density and x - ray luminosity of the corona increase rapidly .
the higher they get , the lower the chance that at the next major merger a cool - gas filament will reach the galactic centre and renew the stellar cusp .
thus there is a direct connection between the process discussed here and the stark contrast in the detectability with x - rays of the coronae of star - forming and early - type galaxies @xcite .
the progressive reduction with cosmic time in the typical halo mass of galaxies making the transition from the blue cloud to the red sequence ( `` downsizing '' ) , follows from the decrease in the abundance of cold infalling gas with both cosmic time and halo mass - scale that is expected on analytic grounds and observed in ab - initio cosmological simulations . it would be interesting to add this idea to semi - analytic models of the evolution of the galaxy population
we thank an anonymous referee for his / her valuable comments .
most of the numerical simulations were performed using the bcx system at cineca , bologna , with cpu time assigned under the inaf - cineca agreement 2008 - 2010 .
fm gratefully acknowledges support from the marco polo program , university of bologna . | it is argued that galaxies like ours sustain their star formation by transferring gas from an extensive corona to the star - forming disc . the transfer is effected by the galactic fountain cool clouds that are shot up from the plane to kiloparsec heights above the plane .
the kelvin - helmholtz instability strips gas from these clouds .
if the pressure and the the metallicity of the corona are high enough , the stripped gas causes a similar mass of coronal gas to condense in the cloud s wake .
hydrodynamical simulations of cloud - corona interaction are presented .
these confirm the existence of a critical ablation rate above which the corona is condensed , and imply that for the likely parameters of the galactic corona this rate lies near the actual ablation rate of clouds . in external galaxies trails of hi behind individual clouds will not be detectable , although the integrated emission from all such trails should be significant .
parts of the trails of the clouds that make up the galaxy s fountain should be observable and may account for features in targeted 21-cm observations of individual high - velocity clouds and surveys of galactic hi emission .
taken in conjunction with the known decline in the availability of cold infall with increasing cosmic time and halo mass , the proposed mechanism offers a promising explanation of the division of galaxies between the blue cloud to the red sequence in the colour - luminosity plane .
[ firstpage ] hydrodynamics turbulence
ism : kinematics and dynamics galaxy : kinematics and dynamics galaxy : structure galaxies : formation intergalactic medium cooling flows |
localizing scalars and fermions in brane - world field theoretic models is conceptually straightforward @xcite . on the other hand ,
gauge field localization is tricky .
one either does not allow gauge fields to penetrate into the bulk @xcite , or has to deal with charge universality @xcite , i. e. , the property that the effective 4d gauge coupling must be independent of the bulk wave function of a charged particle .
the latter property implies that the gauge field zero mode is constant along extra dimensions ( unlike zero modes of scalars and fermions , which typically decay away from the brane ) .
the independence of the gauge field zero mode of extra - dimensional coordinates may lead to infrared problems .
indeed , it has been pointed out in refs .
@xcite that charged bulk fields induce one - loop scattering amplitude of the zero mode gauge bosons , see fig .
[ figure1 ] , which diverges in 5d lineary with the size of extra dimension @xmath0 , provided that the effective mass of the charged field does not grow indefinitely away from the brane .
this is easy to understand : suppose one keeps the size of the loop finite and moves it towards @xmath1 , where @xmath2 is the extra coordinate .
since the gauge field zero mode is constant in @xmath2 , and the effective mass of the charged field is constant at large @xmath2 , the loop contribution is independent of @xmath2 , and the integration over @xmath2 gives the volume of extra dimension .
likewise , the one - loop correction to the 4d propagator of the gauge field boson in the zero mode state is also proportional to @xmath0 .
this observation , however , does not necessarily mean that a theory has unacceptable ir pathologies . indeed ,
since the zero mode is strongly non - local along extra dimension , it can not be produced alone by a local source .
the ir pathology is indeed likely to be there in models with a gap between the zero and higher modes in the gauge field sector @xcite : at low 4d momenta heavy states decouple , and the only relevant degree of freedom is the zero mode .
however , the gap is absent in other models of gauge field localization , notably , generalizations @xcite of the rsii set up @xcite .
arguments have been given some time ago @xcite , suggesting that models of the latter type may be free of ir pathologies . in this paper
we calculate and compare objects of direct physical significance in models with and without gap . since in the brane - world scenario one
is interested primarily in the processes with particles ( say , charged fermions ) residing on the brane , the object of particular relevance is the brane - to - brane gauge field propagator .
as we are interested in ir properties , we are going to study its behavior as @xmath3 , where @xmath4 is 4d momentum .
we will calculate corrections to the brane - to - brane gauge field propagators , which are due to fermion loops in the bulk ; we consider massless fermions for simplicity . for concreteness
, we consider models of ref .
@xcite ( with the gap ) and ref .
@xcite with @xmath5 ( gapless ) .
our results confirm that the models with the gap between the gauge zero mode and higher modes are ir pathological .
namely , we will see that one loop correction to the brane - to - brane gauge field propagator behaves as @xmath6 at low 4d momenta ( the bare propagator is proportional to @xmath7 , as usual ) . on the other hand ,
the one loop correction in the gapless model is free of pathology : the one - loop correction behaves as @xmath7 , so it merely introduces the wave function renormalization . .
[ figure1],scaledwidth=50.0% ] the organization of this paper is as follows . in sec .
[ domainwallaction ] we consider 5d spinor qed with gauge field localized on the domain wall .
we introduce the model in sec .
[ thedwmod ] and in sec .
[ domainwallcorrections ] we explicitly calculate one loop fermion correction to the vector brane - to - brane propagator . in sec .
[ rsii-1 ] we consider 6d spinor qed in the background of rsii-1 metric with one compact extra dimension @xmath8 and one infinite extra dimension @xmath2 .
we calculate the vector propagator in sec .
[ vecgreenrsii ] . in sec .
[ mslsrsii ] we derive one loop contribution of @xmath8-homogeneous kk fermions into the vector brane - to - brane propagator . in sec .
[ kkrsii ] we discuss a position depenent cutoff scheme for calculating the one loop contribution of @xmath8-inhomogeneous kk excitations of fermions to the vector propagator , and calculate this correction explicitly .
we conclude in sec . [ discussion ] .
technical details are collected in appendices .
in this section we consider euclidean 5d brane - world model with fermion and vector fields propagating in the bulk @xcite .
the action of the model is @xmath9 , \label{dwaction}\ ] ] where indices @xmath10 label 5d space , @xmath11 , and @xmath12 is a field configuration which ensures the localization of vector zero mode on the brane ; the parameter @xmath13 is related to the brane thickness .
this mechanism of gauge field localization on a higher dimensional deffect is analogous to those considered in ref .
@xcite . for simplicity
we assume that the fermions propagate in the flat 5d bulk . 5d and
4d couplings are related by @xmath14 it is convenient to introduce the new field @xmath15 , which is related to @xmath16 as follows : @xmath17 the lagrangian of the vector field @xmath15 can be written as follows : @xmath18\!\!b_\nu- \\
-\frac{1}{2}b_z\partial^2_\mu b_z- b_z\!\left(\frac{\phi'}{\phi}- \partial_z\right)\!\!\partial_\mu b_\mu , \label{vecactresc } \end{split}\ ] ] where greek indices refer to 4d , @xmath19 .
we set the gauge @xmath20 , then the vector field @xmath21 is transverse , @xmath22 .
kk modes @xmath23 of the vector field @xmath21 obey , scaledwidth=50.0% ] @xmath24 where @xmath25 is 4d mass .
the mass spectrum is determined by the quantum - mechanical potential @xmath26 in this potential the vector field has one bound state ( see fig . [ figure3 ] ) , which is actually the zero mode , @xmath27 this mode is normalized with unit measure , in accordance with the lagrangian ( [ vecactresc ] ) .
( [ veceq1 ] ) has also solutions @xmath23 , which correspond to the continuous spectrum starting from @xmath28 .
therefore , the zero mode @xmath29 is separated from higher modes @xmath23 by non - zero mass gap @xmath30 .
now let us consider the propagator of the vector field @xmath21 from brane to bulk , @xmath31 .
this propagator obeys @xmath32 the solution to eq .
( [ kinkvecgrf2 ] ) is @xmath33 , where ( see appendix [ greedw ] ) @xmath34 and @xmath35 it is worth noting that brane - to - bulk propagator of the vector field @xmath36 is @xmath37 where @xmath38 vector fields @xmath16 and @xmath15 coincide on the brane @xmath39 , their propagators from brane to brane coincide as well @xmath40 at low energy , @xmath41 , only zero mode of gauge field is relevant , and the propagator ( [ zeromodeveckink ] ) takes the form @xmath42 it follows from eqs .
( [ chi ] ) and ( [ grenormkinkaterm ] ) that at small @xmath4 @xmath43 hence the bulk vector green s function @xmath44 decreases slowly towards @xmath45 in the ir regime .
we will also use the propagator @xmath46 in the momentum space @xmath47 we find from eqs .
( [ chi ] ) and ( [ greenkinkfourier ] ) , that as @xmath48 and @xmath49 , then the brane - to - bulk vector propagator in the momentum space tends to @xmath50 we note that eq .
( [ fourierdw ] ) differs considerably from the vector propagator in the flat 5d space @xmath51 in sec .
[ domainwallcorrections ] we show that the term @xmath52 in eq .
( [ fourierdw ] ) leads to ir pathology of the one - loop vector brane - to - brane propagator .
, scaledwidth=47.0% ] in this section we calculate the one loop fermion contribution to the vector propagator from brane to brane ( see fig .
[ figure2 ] ) .
the coupling of the fields @xmath53 and @xmath54 is @xmath55= \int d^4 x\ , dz\ , \g_5 \overline{\psi } \ , \gamma^\mu a_{\mu}\psi . \label{intdomainwall}\ ] ] one loop brane - to - brane vector green s function is @xmath56 where tree level propagator is @xmath57 , and one loop contribution @xmath58 is given by @xmath59 .
\label{loopcorrrsii-1 } \end{split}\ ] ] green s function of massless 5d fermions @xmath60 has the form @xmath61 where @xmath62 , @xmath63 is 5d momentum squared , @xmath64 .
therefore , the one loop fermion contribution ( [ loopcorrrsii-1 ] ) can be written as follows @xmath65 where @xmath66 is given by eq .
( [ greenkinkfourier ] ) and @xmath67 is vacuum polarization operator in flat 5d space with 4d indices : @xmath68 } { q^2(q - p)^2}.\label{piabtildekink}\ ] ] we make use of dimensional regularization and obtain @xmath69 where 5d momentum squared is @xmath70 let us consider the one loop correction @xmath71 in the low energy limit , @xmath3 .
we substitute eqs .
( [ greenkinkfourier ] ) and ( [ piabfourier ] ) into eq .
( [ loopcorrection11 ] ) and integrate ( [ loopcorrection11 ] ) over @xmath72 , then we get @xmath73
. \label{piabloopcorrkink } \end{split}\ ] ] the factor @xmath74 in eq .
( [ piabloopcorrkink ] ) is due to the vector zero mode ( see eq .
( [ zeromodeveckink2 ] ) ) .
functions @xmath75 and @xmath76 are defined by @xmath77 one obtains in ir regime @xmath78 @xmath79 @xmath80 where @xmath81 is the uv cutoff scale .
it follows from eq .
( [ pitransverse ] ) that in the low energy limit @xmath49 the function @xmath75 is proportional to @xmath82 .
the reason for this singularity is as follows .
we have pointed out in sec .
[ thedwmod ] that brane - to - bulk propagator @xmath66 has a peculiar term @xmath52 in the denominator as @xmath83 ( see eq . ( [ fourierdw ] ) ) .
this means that the integral ( [ qsdw ] ) for @xmath84 has the following ir contribution : @xmath85 this integral is saturated in ir region at @xmath86 , hence @xmath87 which coincides with eq .
( [ pitransverse ] ) .
therefore , the one loop contribution of massless fermions to the vector brane - to - brane propagator is @xmath88 while the tree level propagator @xmath89 is propotional to @xmath7 ( see eq .
( [ zeromodeveckink ] ) ) .
this means that the model is inconsistent in the ir limit .
one way to interpret the result ( [ g1dwmslss ] ) is to claim that the theory is in the strong coupling regime at @xmath90 , where the one loop correction exceeds the tree level propagator .
we conclude that the model with mass gap between zero and massive vector kk modes is pathological in ir , as expected .
in this section we consider a model without mass gap between zero and massive vector kk modes .
namely , we calculate the one loop contribution of massles fermions to the brane - to - brane vector propagator in the framework of rsii-1 model with one compact and one infinite extra dimensions
. 6d mectric of euclidean rsii-1 model is @xmath91 where indices @xmath10 denote the coordinates of 6d spacetime , @xmath92 .
greek indices @xmath93 label 4d subspace , @xmath19 , @xmath94 is the compact extra dimension , @xmath95 $ ] , and @xmath96 refers to extra dimension of infinite size @xmath97 $ ] .
the warp factor @xmath98 is given by @xmath99 from geometric point of view , the metric ( [ ads6metric ] ) describes 5-brane with one compact dimension located at the point @xmath39 of bulk space .
let us consider the action of @xmath100 gauge theory with massles fermions in the background ( [ ads6metric ] ) @xmath101= \int d^4 x\ , dz\ , d\theta \sqrt{g}\big[\frac{1}{4}f_{mn}f^{mn}+ \\
+ \,i\overline{\psi}\gamma^{\bar{m}}e_{\bar{m}}^m(\nabla_m - i \g_6 a_m)\psi \big ] , \label{u(1)action } \end{split}\ ] ] where indices @xmath102 label the tangent space and @xmath103 is spinorial covariant derivative .
couplings and fields have the following mass dimensions : @xmath104=-1 $ ] , @xmath105=2 $ ] and @xmath106=5/2 $ ] .
the size of compact extra dimension @xmath107 is assumed to be @xmath108 , where @xmath109 is the energy of interest . in secs .
[ vecgreenrsii ] and [ mslsrsii ] we study kk excitations of fermions and bosons which are homogeneous along the compact extra dimension @xmath8 .
we discuss in sec .
[ kkrsii ] quantum corrections to the photon propagator coming from the fermion states inhomogeneous along @xmath8 .
, scaledwidth=50.0% ] let us find brane - to - bulk vector green s function .
the vector part of the action ( [ u(1)action ] ) is @xmath110=\int d^4x\ , dz\ , d\theta \ ,
w^2(z)\,\frac{1}{4 } f_{mn}^2 , \label{vecact}\ ] ] where indices @xmath10 are contracted with flat metric .
the action ( [ vecact ] ) is analogous to that considered in domain wall set up ( see vector part of the action ( [ dwaction ] ) ) expect for the form of the warp factor @xmath98 .
we introduce the vector field @xmath15 in the same way as in sec .
[ thedwmod ] ( see eq .
( [ vecresckink ] ) ) @xmath111 we consider the field @xmath15 independent of @xmath8 with @xmath112 and choose the gauge @xmath20 . the equation of motion for kk mode @xmath23 of the field @xmath113 coincides with eq .
( [ veceq1 ] ) up to redefinition @xmath114 . in rsii-1 set up the spectrum is determined by quantum - mechanical potential ( see fig .
[ figure4 ] ) @xmath115 vector field @xmath23 has a zero mode @xmath116 which has to do the delta function well in @xmath117 .
this zero mode corresponds to the constant field @xmath118 .
the latter is homogeneous along the large extra dimension @xmath2 and represents the photon localized on the brane .
in contrast to the domain wall set up , there is no mass gap , separating the zero mode @xmath29 from the continuum of states @xmath23 . the brane - to - bulk vector green s function @xmath119 obeys @xmath120 we obtain the solution to eq .
( [ vecbranetobulkgreenf ] ) in appendix [ rsiinprop ] .
it is given by @xmath33 , where @xmath121 the first term in eq .
( [ greenvecpz ] ) is the green s function of massles field in flat 5d spacetime . the second term of eq .
( [ greenvecpz ] ) is proportional to @xmath7 , therefore this is the zero mode contribution to the vector propagator at the tree level . in the ir regime @xmath122 , brane - to - bulk propagator of the field @xmath123 has the form @xmath124 in contrast to the domain wall case ( see eq .
( [ vecgfdwa ] ) ) @xmath46 of rsii-1 set up decreases more rapidly towards @xmath45 in ir limit .
it is worth rewriting the green s function @xmath46 in the momentum space : @xmath125 where @xmath126 is defined by eq .
( [ 5dpsquared ] ) . in this section
we derive one loop fermion correction to the vector green s function from brane - to - brane .
it is known that massless fermions are conformal , i. e. , upon rescaling of the fermion field @xmath127 the fermion action reduces to the flat - space form @xmath128\!=\!\!\int\!\ !
d^4 x\ , dz\ , d\theta \!\left ( i\overline{\psi}\gamma^{\mu}\partial_\mu \psi \!+\!i\overline{\psi}\gamma^{\theta}\partial_{\theta } \psi \!+\!i\overline{\psi}\gamma^{z}\partial_{z}\psi \right ) .
\label{fermrescact}\ ] ] where @xmath129 , @xmath130 and @xmath131 are euclidean 6d gamma matrices ; @xmath54 is eight - component dirac spinor @xmath132 where @xmath133 and @xmath134 are four - component spinors which have appropriate signs of 6d chirality .
consider now the kk mode expansion @xmath135 where @xmath136 and @xmath137 are @xmath8-homogeneous and @xmath8-inhomogeneous kk modes , respectively .
we have dropped 6d chirality indices @xmath138 in eq .
( [ kkfermionsrs ] ) for simplicity . integrating out the compact extra dimension ,
one obtains @xmath139\!=\!\int\!\ !
d^4x dz d \theta \sum_{n } \big(i \overline{\psi}^{n } \gamma^\mu \partial_\mu \psi^{n}+ \\ + \
i\overline{\psi } ^{n } \gamma^z \partial_z \psi^{n } + i m_n \overline{\psi}^{n }
\psi^{n}\big ) , \end{split}\ ] ] where @xmath140 is the mass of kk state .
let us consider the one loop correction to the vector propagator coming from the @xmath8-homogeneous state @xmath141 .
the interaction of @xmath136 and @xmath142 is described by the action @xmath143\!=\!\!\int\!\ !
d^4x\ , dz \ , \g_5 \ , \overline{\psi}^{0}\!(x , z ) \gamma^\mu
a_\mu(x , z)\psi^0(x , z ) , \label{intzerors}\ ] ] where @xmath144 is the 5d coupling @xmath145 with mass dimension @xmath146=-1/2 $ ] . the action ( [ intzerors ] ) is analogous to the vector - fermion coupling in the domain wall set up ( see eq .
[ intdomainwall ] ) .
since there are two types of four - component dirac spinors in 6d model , one can carry over eqs .
( [ intdomainwall ] ) - ( [ qsdw ] ) to the rsii-1 scenario up to the fermion doubling factor in eq .
( [ piabfourier ] ) @xmath147 upon substituting eqs .
( [ greenfourierrs ] ) and ( [ piabfourier ] ) into eq .
( [ loopcorrection11 ] ) , one gets @xmath148 , \label{piabloopcorrrs } \end{split}\ ] ] where 5d and 4d effective couplings are related by eq .
( [ g5g4 ] ) up to the redefinition @xmath149 .
the functions @xmath75 and @xmath76 are defined by eq .
( [ qsdw ] ) with @xmath150 given by eq .
( [ greenfourierrs ] ) . then , after integrating eq .
( [ qsdw ] ) over @xmath72 , one finds @xmath151 therefore , the one loop fermion contribution @xmath71 to the vector brane - to - brane propagator @xmath152 ( cf .
( [ greenvecpz ] ) at @xmath39 ) in ir regime is given by @xmath153 it is worth noting that both @xmath71 and @xmath152 are proportional to @xmath154 as @xmath3 .
this means that there is no ir singularity in the one loop vector brane - to - brane propagator of the rsii-1 set up .
this is in contrast to sec .
[ domainwallaction ] .
one can understand the ir behaviour of @xmath71 in a way analogous to that considered in sec .
[ domainwallcorrections ] .
namely , let us consider eq .
( [ qsdw ] ) in the case of rsii-1 model . from ( [ greenfourierrs ] ) it follows that if @xmath155 then @xmath156 .
hence the main ir contributions to @xmath75 and @xmath76 come from the integrals @xmath157 ^ 2\!\ !
( p_z^2+p^2)^{3/2 } , \,\ , \\
q_l(p ) \propto \!\!\ ! \int\limits_{-\infty}^{+\infty}\!\ !
dp_z \frac{p^2}{k^2 } \!\!\left[\frac{k p}{(p_z^2+p^2)^2}\right]^2\!\!(p_z^2+p^2)^{1/2 } p_z^2 , \label{loopcorrrsnaive } \end{split}\ ] ] these integrals are saturated at @xmath158 , therefore one has @xmath159 which coincides with eq .
( [ qrsii1 ] ) for @xmath122 .
thus , the analysis of the one loop vector propagators in the two brane world models shows that the ir behaviour of vector brane - to - brane propagators is in one - to - one correspondence with the existence of the mass gap .
while there is ir pathology in the model with the gap , the gapless model is ir healthy . in this section
we derive one loop contribution of @xmath8-inhomogeneous fermions to the vector brane - to - brane propagator .
since the masses of corresponding kk excitations are large , @xmath160 , production of heavy inhomogeneous fermion modes is forbidden at the tree - level .
nevertheless , these modes may contribute to the vector brane - to - brane propagator at the one loop level .
let us recall that rsii-1 model is a 6d non - renormalizable spinor qed with warped extra dimension .
if the set up were 6d spinor qed on flat background , we would have to impose the following condition on gauge coupling @xmath161 and uv cutoff scale @xmath81 @xmath162 in warped space - time , it is appropriate to use the position dependent cutoff formalizm @xcite .
namely , the cutoff scale at given @xmath2 is @xmath163 .
in particular , one imposes position dependent upper bound on kk masses @xmath164 thus , the one - loop contribution of @xmath8-inhomogeneous kk states can be written as follows : @xmath165 where @xmath166 is the effective number of kk states , and @xmath167 ; @xmath168 is the one loop fermion integral of massive kk excitations ( compare with massless case eq .
( [ piabtildekink ] ) and ( [ piabfourier ] ) ) @xmath169 where @xmath170 .
since @xmath171 , we set @xmath172 in eq .
( [ piabtildekink2 ] ) , and then evaluate the integral .
in this way we obtain in ir regime @xmath173 @xmath174 let us use the variables @xmath167 and @xmath175 in eq .
( [ piabmassive ] ) . implementing the cutoff condition ( [ kkupperbound ] ) , we interchange the order of bulk space integration and kk summation in eq .
( [ piabmassive ] ) . then
, after integrating over @xmath72 , we have @xmath176\times \\ & \times g_a(p , z\!+\!z'/2,0 ) g_a(p , z\!-\!z'/2,0 ) , \label{piabmassive22 } \end{split}\ ] ] where @xmath177 , the masses of kk states are bounded now by @xmath178 .
hence , the number of kk states in ( [ piabmassive22 ] ) is @xmath179 integrating eq .
( [ piabmassive22 ] ) over @xmath180 and @xmath2 , one obtains @xmath181\!\mathcal{f}^{n}_1(p)+ \\ & + \eta_{\mu\nu } \mathcal{f}^{n}_2(p)\big ) , \end{split}\ ] ] where @xmath182 and @xmath183 are given by @xmath184 @xmath185 let us consider low momentum regime @xmath186 , then at @xmath187 one has @xmath188 therefore @xmath189 which leads to @xmath190 it follows from eq .
( [ l5l6 ] ) and eq .
( [ g5g4 ] ) that 6d and 4d couplings are related by @xmath191 therefore eq . ( [ thetainhomcontrb ] )
is finally written as @xmath192 this is indeed a small correction to the tree - level propagator , provided that @xmath193 , which coincides with eq .
( [ couplingcutoff ] ) .
we conclude that the entire rsii-1 scenario is viable at least at the one loop level .
in this paper we have considered two brane - world models with different mechanisms of gauge field localization on the brane .
we found that the fermion one loop correction to the vector brane - to - brane propagator has a pathological ir divergence in the framework of 5d massles spinor qed with gauge field localized on the domain wall , which makes this model inconsistent .
this result is consistent with the observation in ref .
we have also considered 6d massles spinor qed in the background of modified randall - sundrum metric .
we have explicitly calculated the one loop fermion cotribution to the vector brane - to - brane propagator in this framework in the low energy limit .
this contribution is healthy in ir , so one can consider the rsii-1 set up as consistent brane world scenario , at least at the one loop level . we conclude that ir are inherent in models with gauge field zero mode separated from heavier modes by a gap , while models without the gap may be healthy in ir .
we are indebted to d. s. gorbunov , a. l. kataev , m. y. kuznetsov , d. g. levkov , a. g. panin , v. a. rubakov and m. a. smolyakov for helpful discussions .
this work was supported in part by grants of russian ministry of education and science ns-5590.2012.2 and gk-8412 , grants of the president of russian federation mk-2757.2012.2 , and grants of rfbr 12 - 02 - 31595 mol a and rfbr 13 - 02 - 01127 a.
in this appendix we derive the vector propagator from the brane to bulk in the model of sec .
[ domainwallaction ] .
we take the solution to eq .
( [ kinkvecgrf2 ] ) in the form @xmath33 , where @xmath194 here @xmath195 and @xmath196 are linear combinations of odd and even solutions @xmath197 here @xmath198 @xmath199 where @xmath200 is hypergeometric function , parameters @xmath201 and @xmath202 are defined by @xmath203 with @xmath204 we impose the boundary condition far from the brane @xmath205 it follows from eqs .
( [ kinkvecevenoddexpansion ] ) and ( [ kinkvecboundcond ] ) that @xmath206 @xmath207 then eq .
( [ kinkvecevenoddexpansion ] ) can be written in the following form : @xmath208 , \\
g^{(-)}(p , z)=c^{(-)}[d(p ) g^{(\e)}(p , z)+ g^{(o)}(p , z ) ] .
\label{kinkvecevenoddexpansion2 } \end{split}\ ] ] matching condition @xmath209 at the brane position @xmath39 gives @xmath210 discontinuity of the derivative @xmath211 at the point @xmath39 yields @xmath212 thus , the propagator @xmath213 reads @xmath214 .
\label{gbrnblk}\ ] ] expanding @xmath215 and @xmath216 at large values of the variable @xmath217 , one has @xmath218 @xmath219
this gives @xmath220 substituting eqs .
( [ geven ] ) , ( [ godd ] ) and ( [ dpdefinition ] ) into eq .
( [ gbrnblk ] ) , and using the identity @xmath221 we obtain the final form of the brane - to - bulk vector propagator @xmath222
in this appendix we derive brane - to - bulk vector propagator @xmath213 in the framework of rsii-1 .
the latter obeys @xmath223 we take the solution to eq .
( [ rsgreq1 ] ) in the following form @xmath224 where @xmath225 matching condition of @xmath195 and @xmath196 and discontinuity of derivative at the point @xmath39 are @xmath226 @xmath227 this yields @xmath228 hence , we obtain @xmath229 v. a. rubakov and m. e. shaposhnikov , phys .
b * 125 * , 136 ( 1983 ) .
k. akama , lect .
notes phys . *
176 * , 267 ( 1982 ) [ hep - th/0001113 ] .
s. randjbar - daemi and m. e. shaposhnikov , phys .
b * 492 * , 361 ( 2000 ) [ hep - th/0008079 ] .
h. guo , a. herrera - aguilar , y. -x .
liu , d. malagon - morejon and r. r. mora - luna , arxiv:1103.2430 [ hep - th ] .
y. -x .
liu , x. -n .
zhou , k. yang and f. -w .
chen , phys .
d * 86 * , 064012 ( 2012 ) [ arxiv:1107.2506 [ hep - th ] ] . j. e. g. silva and c. a. s. almeida , phys .
d * 84 * , 085027 ( 2011 ) [ arxiv:1110.1597 [ hep - th ] ] .
a. a. andrianov , v. a. andrianov and o. o. novikov , arxiv:1210.3698 [ hep - th ] .
s. l. dubovsky and v. a. rubakov , int .
j. mod .
a * 16 * , 4331 ( 2001 ) [ hep - th/0105243 ] .
m. n. smolyakov , phys . rev .
d * 85 * , 045036 ( 2012 ) [ erratum - ibid .
d * 87 * , 029901 ( 2013 ) ] [ arxiv:1111.1366 [ hep - th ] ] .
m. n. smolyakov , phys .
d * 87 * , 104035 ( 2013 ) [ arxiv:1210.7978 [ hep - th ] ] .
i. oda , phys .
b * 496 * , 113 ( 2000 ) [ hep - th/0006203 ] .
t. gherghetta , e. roessl and m. e. shaposhnikov , phys .
b * 491 * , 353 ( 2000 ) [ hep - th/0006251 ] .
s. l. dubovsky , v. a. rubakov and p. g. tinyakov , jhep * 0008 * , 041 ( 2000 ) [ hep - ph/0007179 ]
. l. randall and r. sundrum , phys .
lett . * 83 * , 4690 ( 1999 ) [ hep - th/9906064 ] .
s. l. dubovsky and v. a. rubakov , hep - th/0204205 .
a. t. barnaveli and o. v. kancheli , sov .
j. nucl .
* 52 * , 576 ( 1990 ) [ yad .
fiz . * 52 * , 905 ( 1990 ) ] . | we discuss the one loop effects of massless fermion fields on the low energy vector brane - to - brane propagators in the framework of two qed brane - world scenarios .
we show that one loop photon brane - to - brane propagator has a power law pathologic ir divergences in the 5d qed brane - world model with mass gap between the vector zero mode and continuous states .
we also find that bulk fermions do not give rise to ir divergences in a photon brane - to - brane green s function at least at the one loop level in the framework of 6d qed brane model with gapless mass spectrum between vector zero mode and higher states . |
conventional molecular electronic structure methods such as density functional theory , hartree - fock theory , and coupled cluster theory start with the assumption that a single slater determinant ( sd ) provides a qualitatively good description of the molecule at hand @xcite .
while this assumption is valid for some molecules near equilibrium geometry , the static correlation which arises in other molecules , as well as for geometries far from equilibrium , requires the use of multireference ( mr ) methods .
these provide a qualitative description which is equivalent to multiple sds , thereby resolving the static correlation .
one of these mr methods is the exact diagonalization of the many - body hamiltonian in the full hilbert space , also known as full configuration interaction ( fci ) in quantum chemistry ( qc ) . because the many - body hilbert space grows exponentially with the number of single - particle states , only small systems , of up to 18 electrons in 18 orbitals ,
can be treated by fci . in 1999 , the density matrix renormalization group ( dmrg ) was introduced in qc @xcite .
this mr method allows to extend the system sizes for which numerically exact solutions can be found to about 40 electrons in 40 orbitals , depending on the nature of the system .
dmrg originated in 1992 in the field of condensed matter @xcite .
although it was originally introduced as a renormalization group flow for increasing many - body hilbert spaces , in 1995 it was realized that dmrg can be reformulated as the variational optimization of a particular wavefunction ansatz , the matrix product state ( mps ) @xcite .
this not only provided the theoretical validation that an energy obtained with dmrg is always an upper bound to the exact eigenvalue , but also shed light on dmrg from a quantum information perspective . non - critical quantum mechanical ground states are believed to obey the so - called area law for the entanglement entropy @xcite .
this implies that quantum correlation is local in such a ground state . for one - dimensional systems
, the boundary of a line segment consists of two points , and the entanglement entropy is a constant , independent of system length .
this is the reason why dmrg works extremely well for one - dimensional non - critical systems .
quantum information theory also induced the development of other so - called tensor network states ( tns ) , which capture the entanglement entropy well in higher dimensional and/or critical systems @xcite .
there even exists a continuous mps ansatz for quantum fields @xcite .
although the active orbital space of most molecular systems is far from one - dimensional , dmrg has been very useful for ab initio qc @xcite .
the variational upper bound to the true eigenvalue , obtained with dmrg , can be systematically improved by increasing the so - called bond or virtual dimension of the mps ansatz .
this provides a way to check the convergence of dmrg calculations . in ab initio qc methods which use fci ,
the fci solver can be replaced by dmrg .
ab initio dmrg allows for an efficient extraction of the reduced two - body density matrix ( 2-rdm ) @xcite .
the 2-rdm of the active space is required in the complete active space self - consistent field ( casscf ) method to compute the gradient and the hessian .
it is therefore natural to introduce a casscf variant with dmrg as active space solver , dmrg - scf @xcite .
this allows one to describe static correlation in large active spaces .
to add dynamic correlation as well , three dmrg - based methods have been introduced . _
( a ) _ with a little more effort , the 3-rdm and contracted 4-rdms can be extracted from dmrg as well .
these are required to apply second order perturbation theory to a casscf wavefunction , called caspt2 .
the dmrg variant is dmrg - caspt2 @xcite . _
( b ) _ based on a casscf wavefunction , a configuration interaction expansion can be introduced , called mrci .
recently , an approximate dmrg - mrci variant was proposed @xcite . _
( c ) _ yet another way is to perform a canonical transformation ( ct ) on top of an mr wavefunction .
when an mps is used as mr wavefunction , the method is called dmrg - ct @xcite .
in addition to ground states , dmrg can also find excited states . by projecting out lower lying eigenstates , or by targeting a specific energy @xcite
, the dmrg algorithm solves for a particular excited state . in these state - specific algorithms ,
the whole renormalized basis is used to represent one single eigenstate . in state - averaged dmrg ,
several eigenstates are targeted at once .
their rdms are weighted and summed to perform the dmrg renormalization step @xcite .
the renormalized basis then represents several eigenstates at once .
dmrg linear response theory ( dmrg - lrt ) can be used as well to find excited states .
once the ground state has been found , the mps tangent vectors to this optimized point can be used as an ( incomplete ) variational basis to approximate excited states @xcite . as the tangent vectors to an optimized sd yield the configuration interaction with singles ( cis ) , also called the tamm - dancoff approximation ( tda ) , for hartree - fock theory @xcite ,
the same names are used for dmrg : dmrg - cis or dmrg - tda . by linearizing the time - dependent variational principle for matrix product states @xcite , the dmrg random phase approximation ( dmrg - rpa )
is found @xcite , again in complete analogy with rpa for hartree - fock theory .
the variational optimization in an ( incomplete ) basis of mps tangent vectors can be extended to higher - order tangent spaces as well .
dmrg - cisd , or dmrg configuration interaction with singles and doubles , is a variational approximation to target both ground and excited states in the space spanned by the mps reference and its single and double tangent spaces @xcite . in ab initio qc
, two other tnss have been employed as well : the tree tns @xcite and the complete - graph tns @xcite . while they require a smaller virtual dimension to achieve the same accuracy , their optimization algorithms are less efficient , and as a result an mps is currently still the preferred choice for ab initio qc . in section
[ remarks ] , the dmrg algorithm is briefly introduced , and remarks specific to ab initio qc are discussed . in section [ symm ] , the implementation of particle number , spin , and abelian point group symmetries is presented .
an overview of the structure of chemps2 is given in section [ ourcode ] .
results on the low - lying states of the carbon dimer are presented in section [ c2 ] .
a summary is given in section [ summary ] .
atomic units are used in this work : @xmath6 j and @xmath7 m @xcite .
dmrg can be formulated as the variational optimization of an mps .
the mps ansatz with open boundary conditions is given by @xmath8^{n_1}_{\alpha_1 } a[2]^{n_2}_{\alpha_1;\alpha_2 } ...
a[l-1]^{n_{l-1}}_{\alpha_{l-2};\alpha_{l-1 } } a[l]^{n_l}_{\alpha_{l-1 } } \nonumber \\ & & ~ \ket{n_1 n_2 ... n_l}\end{aligned}\ ] ] where @xmath9 denotes the occupancy of orbital @xmath10 ( @xmath11 , @xmath12 , @xmath13 , or @xmath14 ) and the @xmath15 are the so - called bond or virtual indices . with increasing dimension @xmath16 of these virtual indices ,
a larger part of the hilbert space can be reached .
note that it is of no use to make virtual dimension @xmath17 larger than min@xmath18 , the minimum of the sizes of the partial hilbert spaces spanned by resp .
the first @xmath19 and the last @xmath20 orbitals .
the wavefunction @xmath21 does not uniquely define the ansatz , in analogy with a slater determinant . for the latter , a rotation in the occupied orbital space alone , or a rotation in the virtual orbital space alone , does not change the physical wavefunction .
only occupied - virtual rotations change the wavefunction . in an mps
, there is gauge freedom as well .
if for two neighbouring sites @xmath22 and @xmath23 , the left mps tensors are right - multiplied with the non - singular matrix @xmath24 @xmath25^{n_i}_{\alpha_{i-1};\alpha_i } = \sum\limits_{\alpha_j } a[i]^{n_i}_{\alpha_{i-1};\alpha_j } g_{\alpha_j;\alpha_i}\ ] ] and the right mps tensors are left - multiplied with the inverse of @xmath24 @xmath26^{n_{i+1}}_{\alpha_{i};\alpha_{i+1 } } = \sum\limits_{\alpha_j } g^{-1}_{\alpha_{i};\alpha_{j } } a[i+1]^{n_{i+1}}_{\alpha_{j};\alpha_{i+1}}\ ] ] the wavefunction does not change , i.e. @xmath27 : @xmath28^{n_i}_{\alpha_{i-1};\alpha_i } \tilde{a}[i+1]^{n_{i+1}}_{\alpha_{i};\alpha_{i+1 } } = \sum\limits_{\alpha_i } a[i]^{n_i}_{\alpha_{i-1};\alpha_i } a[i+1]^{n_{i+1}}_{\alpha_{i};\alpha_{i+1}}.\ ] ] chemps2 is a two - site dmrg algorithm , were at each so - called micro - iteration two neighbouring sites are simultaneously optimized .
suppose these sites are @xmath22 and @xmath23 .
the gauge freedom of the mps is used to bring it in a particular canonical form . for all sites to the left of @xmath22 ,
the mps tensors are left - normalized : @xmath29^{n_k}\right)^{\dagger}_{\alpha_k ; \alpha_{k-1 } } a[k]^{n_k}_{\alpha_{k-1};\beta_k } = \delta_{\alpha_k , \beta_k } \label{left - normalized}\ ] ] and for all sites to the right of @xmath23 , the mps tensors are right - normalized : @xmath30^{n_k}_{\alpha_{k-1};\alpha_k } \left(a[k]^{n_k}\right)^{\dagger}_{\alpha_k ; \beta_{k-1 } } = \delta_{\alpha_{k-1 } , \beta_{k-1}}. \label{right - normalized}\ ] ] combine the mps tensors of the two sites under consideration into a single two - site tensor : @xmath31^{n_i}_{\alpha_{i-1};\alpha_i } a[i+1]^{n_{i+1}}_{\alpha_{i};\alpha_{i+1 } } = b[i]_{\alpha_{i-1};\alpha_{i+1}}^{n_i;n_{i+1}}. \label{two - site - object - not - reduced}\ ] ] at the current micro - iteration of the dmrg algorithm , @xmath32 $ ] ( the flattened form of the tensor @xmath33 $ ] ) is used as an initial guess for the effective hamiltonian equation .
this equation is obtained by variation of the lagrangian @xcite @xmath34 ) \mid \hat{h } \mid \psi(\mathbf{b}[i ] ) } - \lambda \braket{\psi(\mathbf{b}[i ] ) \mid \psi(\mathbf{b}[i])}\ ] ] to the complex conjugate of @xmath32 $ ] : @xmath35 = \lambda \mathbf{b}[i ] .
\label{effhameq}\ ] ] the specific canonical choice of eqs .
( [ left - normalized])-([right - normalized ] ) ensured that no overlap matrix is present in this effective hamiltonian equation .
the lowest eigenvalue and corresponding eigenvector of this equation are searched . in chemps2 , this is done with our implementation of davidson s algorithm @xcite
. once found , it is decomposed with a singular value decomposition : @xmath36 _ { \left ( \alpha_{i-1 } n_i
\right ) ; \left ( n_{i+1 } \alpha_{i+1 } \right ) } = \sum\limits_{\beta } u[i ] _ { \left ( \alpha_{i-1 } n_i
\right ) ; \beta } \kappa[i]_{\beta } v[i ] _ { \beta ; \left ( n_{i+1 } \alpha_{i+1 } \right)}\ ] ] note that @xmath37 $ ] is hence left - normalized and @xmath38 $ ] right - normalized . in the dmrg algorithm ,
the original sum over @xmath39 of dimension @xmath40 is truncated to @xmath41 , thereby keeping the @xmath41 largest @xmath42_{\beta}$ ] .
so far , we have looked at a micro - iteration of the dmrg algorithm .
this micro - iteration happens during left or right sweeps . during a left sweep , @xmath33 $ ]
is constructed , the corresponding effective hamiltonian equation solved , the solution @xmath33 $ ] decomposed , the singular value spectrum truncated , @xmath43 $ ] is set to @xmath37 \times \kappa[i]$ ] , @xmath44 $ ] is set to @xmath38 $ ] , and @xmath22 is decreased by 1 .
note that @xmath44 $ ] is right - normalized for the next micro - iteration as required .
this stepping to the left occurs until @xmath45 , and then the sweep direction is reversed from left to right .
based on energy differences , or wavefunction overlaps , between consecutive sweeps , a convergence criterium is triggered , and the sweeping stops .
one sweep is called a macro - iteration in dmrg .
the effective hamiltonian in eq . ( [ effhameq ] ) is too large to be fully constructed .
only its action on a particular guess @xmath32 $ ] is available as a function . in order to construct
@xmath46 $ ] efficiently for general quantum chemistry hamiltonians , several tricks are used . _
( a ) _ the one - body matrix elements @xmath47 are incorporated in the two - body matrix elements @xmath48 : @xmath49\ ] ] where @xmath50 is the targeted particle number . _
( b ) _ suppose we want to optimize sites @xmath22 and @xmath23 , and that @xmath51 are the corresponding @xmath52 left renormalized basis states .
renormalized operators such as @xmath53 with @xmath10 and @xmath54 both smaller than @xmath22 are constructed and stored on disk @xcite . for the second quantized operators
@xmath55 and @xmath56 , the latin indices denote orbitals and the greek indices spin projections . _
( c ) _ once three second quantized operators are on one side of @xmath33 $ ] , they are multiplied with the matrix elements @xmath57 , and a summation is performed over the common indices to construct complementary operators @xcite : @xmath58 for two , three , and four second quantized operators on one side of @xmath33 $ ] , these complementary operators are constructed . a bare ( without matrix elements )
renormalized operator is only constructed for one or two second quantized operators on one side of @xmath33 $ ] . _
( d ) _ hermitian conjugation @xmath59 and commutation relations between the second quantized operators are also used to further limit the storage requirement for the renormalized partial hamiltonian terms .
there is also a one - site dmrg algorithm , in which only one mps site tensor is optimized at each micro - iteration , but this algorithm is more likely to get stuck in a local minimum . to help prevent the two - site dmrg algorithm from getting stuck in a local minimum
, a small amount of noise can be added to the solution @xmath33 $ ] , just before it is decomposed . this way , renormalized basis states corresponding to lost symmetries ( which should be there , but are not )
can be reintroduced @xcite .
the choice of orbitals and their ordering on the one - dimensional dmrg lattice have a significant influence both on getting stuck in local minima , as well as on how fast the variational energy @xmath60 converges with increasing @xmath16 @xcite .
the optimal choice and ordering are still under debate , although two rules of thumb are widely used .
active space orbitals in elongated molcular systems ( think about polyenes for example ) should be localized as much as possible to respect the area law for the entanglement entropy @xcite . for small molecules with a high point group symmetry ,
it is beneficial to put bonding and anti - bonding orbitals close to each other on the one - dimensional dmrg lattice , as they are most strongly correlated @xcite .
one possibility to settle this ongoing debate might be to look at the so - called two - orbital mutual information @xmath61 in the future @xcite .
this is a measure from quantum information theory for the amount of correlation between two orbitals , and is a two - point correlation function on the one - dimensional dmrg lattice .
a cost function can be associated with this measure , e.g. @xmath62 , which requires highly correlated orbitals to be close .
its gradient and hessian with respect to orbital rotations can be calculated by resp .
three- and four - point correlation functions on the one - dimensional dmrg lattice . these can be obtained efficiently @xcite . if local minima can be avoided , this yields a set of minimally entangled orbitals and their optimal ordering , from which extra rules of thumb can be drawn .
two extrapolation schemes exist to assess the convergence of the variational energy @xmath60 with increasing number of renormalized basis states @xmath16 .
the first is the scaling relation @xmath63 proposed by chan @xcite which is nowadays not often used .
the @xmath64 are constants which are determined by the fit .
the second and most widely used extrapolation scheme is based on the so - called maximal discarded weight @xmath65 during the last dmrg sweep for a certain value of @xmath16 : @xmath66 ^ 2_{\beta } \right\}.\ ] ] it proposes a linear relation between the variational energy @xmath60 and the discarded weight @xmath65 @xcite : @xmath67 by increasing @xmath16 stepwise , @xmath68 can be extrapolated .
the symmetry group of the hamiltonian can be used to label eigenstates by symmetry . to find an eigenstate with a particular symmetry , it is sufficient to restrict an optimization to the corresponding corner of the many - body hilbert space . for dmrg
, it is well understood how both abelian and non - abelian symmetries can be imposed @xcite .
each mps tensor and intermediary contracted tensor decompose into a clebsch - gordan coefficient and a reduced tensor .
the clebsch - gordan coefficient introduces a sparse block structure in the reduced tensor .
if the symmetry group of the hamiltonian is non - abelian , some irreducible representations ( irrep ) have a dimension larger than one , and then this factorization also presents an information compression , as the size of the full tensor is larger than the size of the reduced tensor .
in addition to the possibility of restricting an optimization to a particular symmetry corner of the many - body hilbert space , this sparsity and compression result in smaller requirements in disk , memory and computer time . in chemps2 , we have implemented three global symmetries for the mps wavefunction : @xmath69 total electronic spin , @xmath70 particle number , and abelian point group symmetry @xmath71 .
as we work real - valued in chemps2 , the latter are restricted to @xmath72 @xcite .
these global symmetries are imposed by requiring that the mps site tensors @xmath43^{n_i}_{\alpha_{i-1};\alpha_i}$ ] are irreducible tensor operators of the total symmetry group @xcite .
the local and virtual basis states ( @xmath73 and @xmath74 ) then have to transform according to the rows of the irreps of this symmetry group .
this is realized by rotating the basis states so that they can be represented by good spin ( @xmath75 and @xmath19 ) , spin projection ( @xmath76 and @xmath77 ) , particle number ( @xmath50 ) , and point group irrep ( @xmath78 ) quantum numbers .
the local basis states of orbital @xmath10 are labeled as @xmath79 where @xmath80 and @xmath81 are resp .
the trivial and orbital @xmath10 point group irreps .
@xmath14 corresponds to @xmath80 because for the abelian point groups with real - valued character tables , @xmath82 . in the same way ,
the virtual basis states are labeled as @xmath83 where the @xmath84 on the right - hand side allows to distinguish between seperate virtual basis states which belong to the same symmetry . due to the wigner - eckart theorem , each irreducible tensor operator @xmath43 $ ] factorizes into clebsch - gordan coefficients and a reduced tensor @xmath85 $ ] : @xmath86^{n_i}_{\alpha_{i-1};\alpha_i } = a[i]^{s s^z n i}_{j_l j_l^z n_l i_l \alpha_{i-1 } ; j_r j_r^z n_r i_r \alpha_i } \nonumber\\ & = \braket{j_l j_l^z s s^z | j_r j_r^z } \delta_{n_l + n ,
n_r } \delta_{i_l \otimes i , i_r } t[i]^{(s n i)}_{(j_l n_l i_l \alpha_l)(j_r n_r i_r \alpha_r ) } \quad \label{tensordecomp}\end{aligned}\ ] ] the @xmath69 , @xmath70 , and @xmath71 symmetries are imposed by their corresponding clebsch - gordan coefficients , and express nothing else than resp .
local allowed spin recoupling , local particle conservation , and local point group symmetry conservation .
the indices @xmath87 and @xmath88 keep track of the number of times an irrep occurs at a virtual bond . if the virtual dimension of a symmetry sector is @xmath89 , this would correspond to a dimension of @xmath90 in an mps which is not symmetry - adapted @xcite . if a clebsch - gordan coefficient is zero by symmetry , the corresponding blocks in @xmath85 $ ] do not need to be allocated , resulting in sparse block structure .
if @xmath19 or @xmath75 are not spin-0 , there is in addition data compression .
the desired global symmetry can be imposed on the mps by requiring that the left virtual index of the leftmost tensor in the mps chain consists of one irrep corresponding to @xmath91 , while the right virtual index of the rightmost tensor consists of one irrep corresponding to @xmath92 , the desired global spin , particle number , and point group symmetry .
this corresponds to the singlet - embedding strategy of sharma and chan @xcite .
the operators @xmath93 for orbital @xmath10 correspond to resp .
the @xmath94 row of irrep @xmath95 and the @xmath96 row of irrep @xmath97 @xcite . @xmath98 and @xmath99 are hence both doublet irreducible tensor operators . this fact permits exploitation of the wigner - eckart theorem also for renormalized operators and complementary operators , and to develop a code without any spin projections or @xmath69 clebsch - gordan coefficients .
contracting terms of the type ( [ tensordecomp ] ) and ( [ creaannih1])-([creaannih2 ] ) can be done by implicitly summing over the common multiplets and recoupling the local , virtual and operator spins .
an example is given in [ redtensors ] .
operators and complementary operators then formally consist of terms containing clebsch - gordan coefficients and reduced tensors . in our code
, however , only the reduced tensors need to be calculated and stored .
chemps2 uses the gnu scientific library @xcite to extract wigner 6-j and 9-j symbols for the recoupling .
no wigner 3-j symbols or clebsch - gordan coefficients are used in the program .
section [ seceffham ] can be reformulated with the reduced @xmath100-tensors from eq .
( [ tensordecomp ] ) and a reduced two - site object @xmath101 $ ] : @xmath102^{j(s_1 s_2 ) n_1 n_2 i_1 i_2}_{j_l n_l i_l \alpha_l ; j_r n_r i_r \alpha_r } = \delta_{n_l+n_1+n_2,n_r } \delta_{i_l \otimes i_1 \otimes i_2 , i_r } \sqrt{2j+1 } \nonumber \\ & ( -1)^{j_l+j_r+s_1+s_2 } \sum\limits_{j_m \alpha_m } \sqrt{2j_m+1 } \left\ { \begin{array}{ccc } j_l & j_r & j \\ s_2 & s_1 & j_m \end{array } \right\ } \nonumber\\ & t[i]^{s_1 n_1 i_1}_{j_l n_l i_l \alpha_l ; j_m ( n_l+n_1 ) ( i_l \otimes i_1 ) \alpha_m } \nonumber \\ & t[i+1]^{s_2 n_2 i_2}_{j_m ( n_l+n_1 ) ( i_l \otimes i_1 ) \alpha_m ; j_r n_r i_r \alpha_r}. \label{tttos}\end{aligned}\ ] ] eq . ( [ tttos ] ) is the analogue of eq .
( [ two - site - object - not - reduced ] ) .
the lagrangian can be written in terms of @xmath101 $ ] , the effective hamiltonian equation can be solved , and after convergence , eq . ( [ tttos ] )
can be backtransformed : @xmath103^{s_1 n_1 i_1 ; s_2 n_2 i_2 ; j_m}_{j_l n_l i_l \alpha_l ; j_r n_r i_r \alpha_r } = \delta_{n_l+n_1+n_2,n_r } \delta_{i_l \otimes i_1 \otimes i_2 , i_r } \nonumber \\ & \sqrt{2j_m+1 } ( -1)^{j_l+j_r+s_1+s_2 } \sum\limits_j \left\ { \begin{array}{ccc } j_l & j_r & j \\
s_2 & s_1 & j_m \end{array } \right\ } \nonumber \\ & \sqrt{2j+1 } s[i]^{j(s_1 s_2 ) n_1 n_2 i_1 i_2}_{j_l n_l i_l \alpha_l ; j_r
n_r i_r \alpha_r}. \label{stott}\end{aligned}\ ] ] per group of @xmath104 , we can perform a singular value decomposition : @xmath103^{s_1 n_1 i_1 ; s_2 n_2 i_2 ; j_m}_{j_l n_l i_l \alpha_l ; j_r n_r i_r \alpha_r } = \sum\limits_{\alpha_m } u[i]^{j_m n_m i_m}_{(j_l n_l i_l \alpha_l s_1 n_1 i_1);\alpha_m } \nonumber \\ & \lambda[i]^{j_m n_m i_m}_{\alpha_m } \left ( \sqrt{\frac{2j_m+1}{2j_r+1 } } v[i]^{j_m n_m i_m}_{\alpha_m;(j_r n_r i_r \alpha_r s_2 n_2 i_2 ) } \right).\end{aligned}\ ] ] after reshaping the indices to the normal form , it can be checked that @xmath37 $ ] is the reduced part of a left - normalized mps site tensor and that the term between brackets is the reduced part of a right - normalized mps site tensor .
the relation between @xmath105 $ ] and @xmath42 $ ] is given by @xmath106_{j_m n_m i_m \alpha_m } = \frac{\lambda[i]_{j_m n_m i_m \alpha_m}}{\sqrt { \sum\limits_{j_q n_q i_q \alpha_q } ( 2j_q+1 ) \lambda[i]_{j_q n_q i_q \alpha_q}^2 } } .\ ] ] the @xmath41 largest values of @xmath105 $ ] are kept .
chemps2 can be obtained from the cpc program library , and from its public git repository @xcite .
the source code contains comments in doxygen format .
a complete reference manual can be generated from these comments .
see ` readme ` on how to install the library and on how to generate the manual . in this section
, we give an overview of the basic structure of chemps2 so that new users can easily understand and alter the test runs to their own needs .
most molecular electronic structure programs have the ability to print matrix elements or to save them in binary format .
chemps2 requires two - body matrix elements with eightfold permutation symmetry , which do not break @xmath69 total electronic spin .
a ` chemps2::hamiltonian ` object should be created at the beginning of a calculation , and filled with the matrix elements of the problem at hand .
users can utilize their preferred molecular electronic structure program to generate the matrix elements .
the functions ` seteconst ` , ` settmat ` , and ` setvmat ` then fill the ` chemps2::hamiltonian ` object elementwise .
note that for @xmath107 we have assumed the physics notation .
this means that orbital @xmath10 at position @xmath108 ( denoted by @xmath109 ) scatters from orbital @xmath110 into orbitals @xmath111 and @xmath112 .
we have used psi4 @xcite to generate molecular orbital matrix elements .
two plugins can be found in the folder ` mointegrals ` , with corresponding instructions in ` readme ` .
one plugin allows to print matrix elements as text during a psi4 calculation , in a format which chemps2 is able to read .
the other plugin creates a ` chemps2::hamiltonian ` object during a psi4 calculation , fills it with the molecular orbital matrix elements , and stores it to disk in binary format .
the latter option requires linking of the chemps2 library to the psi4 plugin , but allows for reduced storage requirements . in the `
chemps2::problem ` object , users can specify the symmetry sector to which the calculations are restricted .
the ` chemps2::hamiltonian ` and the desired total electronic spin , particle number , and point group symmetry then completely determine a fci calculation . in order to do dmrg or dmrg - scf instead of resp .
fci or casscf , a convergence scheme for the subsequent sweeps should be set up .
the ` chemps2::convergencescheme ` object controls the dmrg sweeps .
it is divided into a number of consecutive instructions .
each instruction contains four parameters : the number of reduced renormalized basis states @xmath16 which should be kept , an energy threshold @xmath113 for convergence , the maximum number of sweeps @xmath114 , and the noise prefactor @xmath115 .
the parameters @xmath115 and @xmath16 are relevant for the micro - iterations . just before the decomposition of the reduced @xmath101$]-tensor , random noise
is added to it .
this random noise is bounded in magnitude by @xmath116 , where @xmath65 is the maximum discarded weight obtained during the previous left- or right - sweep .
after decomposition of the reduced @xmath101$]-tensor , its reduced schmidt spectrum @xmath105 $ ] is truncated to @xmath16 . the parameters @xmath113 and @xmath114 are relevant for the macro - iterations .
if after one macro - iteration ( left- plus right - sweep ) , the energy difference is smaller than @xmath113 , the sweeping stops and the next instruction is performed . if energy convergence is not reached after @xmath114 macro - iterations , the current instruction ends as well .
creation of a ` chemps2::dmrg ` object requires a ` chemps2::hamiltonian ` , a ` chemps2::problem ` , and a ` chemps2::convergencescheme ` .
each dmrg calculation starts by creating a new mps .
its virtual dimension @xmath16 is obtained from the first instruction of the ` chemps2::convergencescheme ` object . at each mps bond , this virtual dimension @xmath16 is distributed equally over all possible symmetry sectors , ensuring that the dimension of a certain symmetry sector does not exceed the corresponding fci dimension .
the so - created mps is filled with random noise .
the function ` solve ` performs the instructions of the convergence scheme .
afterwards , it returns the minimal variational energy encountered during all the performed micro - iterations . with the function ` calc2 dm ` , the reduced 2-rdms @xmath117 and @xmath118 are calculated : @xmath119 @xmath117 can be used to calculate the energy , the particle number @xmath50 , and the 1-rdm : @xmath120 and is needed for the dmrg - scf algorithm , while @xmath118 is important for spin - spin correlation functions .
the ` chemps2::dmrg ` object can also calculate excited states .
after the ground state @xmath121 has been determined , the desired number of excited states can be set once with the function ` activateexcitations ` . before `
solve ` is called to find the next new excitation @xmath122 , the function ` newexcitation ` should be called with the parameter @xmath123 .
this pushes back the current mps which represents @xmath124 , and sets the hamiltonian to @xmath125 our excited state dmrg algorithm is hence a state - specific algorithm , which projects out lower - lying states in the given @xmath126 symmetry sector .
an example can be found in ` tests / test5.cpp ` .
openmp parallelization is used in the ` chemps2::dmrg ` object to speed up _ ( a ) _ contractions involving tensors with a sparse block structure , for example the action of the effective hamiltonian on a particular guess , and _ ( b ) _ the construction of the ( often similar ) renormalized operators in between two micro - iterations .
a state - specific dmrg - scf algorithm is implemented in ` chemps2::casscf ` .
its creation requires a ` chemps2::hamiltonian ` object .
the number of occupied , active , and virtual orbitals per point group irrep should be given with the function ` setupstart ` before calling the scf routine .
the casscf routine which is implemented is the augmented hessian @xcite newton - raphson method from ref .
@xcite , with exact hessian .
it can be called with the function ` docasscfnewtonraphson ` , which requires the targeted symmetry sector , the convergence scheme , and the targeted root for the state - specific algorithm . when the gradient for orbital rotations reaches a predefined threshold , the routine returns the converged dmrg - scf energy .
an example can be found in ` tests / test6.cpp ` .
despite its simplicity at first sight , the carbon dimer provides a rich source of interesting physics .
the bond between the two carbon atoms is of the charge - shift type @xcite .
its strength tempts chemists to classify it as a quadruple bond @xcite , and recent research indicates how this fourth bond can be interpreted @xcite .
the @xmath1 core correlation is significant @xcite .
the low - lying bond dissociation curves are quasi - degenerate , and avoided crossings occur between states with the same spin and @xmath127 point group symmetry @xcite .
this happens for example between the @xmath2 and @xmath128 states , and between the @xmath129 and @xmath130 states .
fortunately , relativistic effects are small @xcite .
accurate data for the low - lying states , preferably at the fci level of theory for a given basis set , are useful to assess the accuracy of approximate molecular electronic structure methods .
the @xmath2 , @xmath131 , and @xmath132 bond dissociation curves of ref .
@xcite at the frozen core fci/6 - 31 g * level of theory are utilized to this end in several works @xcite .
the 12 lowest states of the carbon dimer are @xmath2 , @xmath133 , @xmath134 , @xmath135 , @xmath129 , @xmath131 , @xmath132 , @xmath136 , @xmath137 , @xmath138 , @xmath139 , and @xmath130 @xcite . in section [ c2resultssec ]
, we present the bond dissociation curves of these states at the dmrg(28o , 12e , d@xmath0=2500)/cc - pvdz level of theory . to estimate the contribution of @xmath1 core correlation to the @xmath2 bond dissociation curve , we compare energies at the dmrg(28o , 12e , d@xmath3=2500)/cc - pvdz , dmrg - scf(26o , 8e , d@xmath3=2500)/cc - pvdz , dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz , and dmrg - scf(34o , 8e , d@xmath3=2500)/cc - pcvdz levels of theory in section [ sec1scorecorr ] .
the cc - pcvdz basis augments the cc - pvdz basis with extra @xmath1 and @xmath140 functions to treat core and core - valence correlation @xcite .
for all calculations , the variational energies are converged to @xmath141 from the extrapolated value .
this implies that , for all practical purposes , we present data at the fci / cc - pvdz , casscf(26o , 8e)/cc - pvdz , fci / cc - pcvdz , and casscf(34o , 8e)/cc - pcvdz levels of theory .
since chemps2 can only handle abelian point groups , we use @xmath142 point group symmetry to obtain these 12 states : @xmath143 for the states ( [ firstofthed2hirreps])-([lastofthed2hirreps ] ) , we have calculated one extra state to check that no unexpected curve crossings occur . to discern the lowest three @xmath144 states , we have extracted the following fci coefficients from the dmrg object @xcite : @xmath145
when the fci coefficients are equal , the state has @xmath146 symmetry , and when the fci coefficients are each other s additive inverse , the state has @xmath147 symmetry . to discern the lowest three @xmath148 states , we have extracted the following fci coefficients from the dmrg object : @xmath149 when the fci coefficients are equal , the state has @xmath150 symmetry , and when the fci coefficients are each other s additive inverse , the state has @xmath151 symmetry .
an example is shown in fig .
[ fig3b1ucoeff ] . for the cc - pvdz basis ,
the @xmath139 state drops below the @xmath129 state at an interatomic distance between 3.6 and 3.8 @xmath152 .
the @xmath153 and @xmath154 fci coefficients allow to correctly label the @xmath148 ground state ( state 0 ) and first excited state ( state 1).,scaledwidth=45.0% ] the standard @xmath142 irrep order is not optimal to study the carbon dimer with dmrg .
as stated in section [ convergencesection ] , it is best to group bonding and anti - bonding orbitals together on the dmrg lattice .
the convergence behaviour of these two irrep orderings is shown in fig .
[ figirreporder ] .
we have used the latter ordering for our calculations . the orbital choice and ordering influence the convergence behaviour of dmrg .
the convergence behaviour of two irrep orderings is shown for the carbon dimer with interatomic distance 2.4 @xmath152 in the cc - pvdz basis .
the extrapolated energy was obtained from the ordering where bonding and anti - bonding orbitals are grouped , with the method described in section [ howtoextrapolateenergies].,scaledwidth=45.0% ] we have used the convergence scheme in tab .
[ tablec2convscheme ] for all the calculations of the carbon dimer .
the extrapolation scheme of eq .
( [ extrapolschemeeq ] ) is used to obtain energies which are correct up to 0.01 @xmath5 .
an example of such an extrapolation is shown in fig .
[ figextrapolationexample ] .
the energies shown in sections [ c2resultssec ] and [ sec1scorecorr ] are the extrapolated values .
the extrapolation scheme of eq .
( [ extrapolschemeeq ] ) is used to obtain energies which are correct up to 0.01 @xmath5 .
the example shown here is for the @xmath2 state of the carbon dimer at an interatomic distance of 2.35 @xmath152 in the cc - pvdz basis.,scaledwidth=45.0% ] .[tablec2convscheme ] convergence scheme for the carbon dimer calculations .
the symbols are explained in section [ convergenceschemesection ] . [ cols="^,^,^,^",options="header " , ] assessment of the importance of @xmath1 core correlation .
this effect is captured at the dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz level of theory.,scaledwidth=45.0% ] assessment of the importance of @xmath1 core correlation .
the relative energies with respect to the dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz calculations are shown.,scaledwidth=45.0% ] the extrapolated energies at the dmrg - scf(26o , 8e , d@xmath3=2500)/cc - pvdz , dmrg(28o , 12e , d@xmath3=2500)/cc - pvdz , dmrg - scf(34o , 8e , d@xmath3=2500)/cc - pcvdz , and dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz levels of theory are given in tab .
[ corecorrtable ] and are shown in fig .
[ corecorrabsolute ] .
the relative energies with respect to the dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz calculations are shown in fig .
[ corecorrrelative ] .
the @xmath1 core correlation is only captured at the dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz level of theory . without the necessary orbital freedom ,
the @xmath1 core correlation can not be captured .
the non - parallelity of the dmrg - scf(34o , 8e , d@xmath3=2500)/cc - pcvdz curve in fig .
[ corecorrrelative ] is of the order of 2 @xmath5 , far below the error due to basis set incompleteness . for small interatomic distances ,
the cc - pcvdz curves show a different behaviour than the cc - pvdz curves , as can be seen in fig .
[ corecorrrelative ] .
extra basis set freedom is required to capture the more complicated core dynamics in the united atom limit .
this can be understood as the transition from two light atoms , each with a doubly filled @xmath1 orbital , to one single heavy atom , with several orbitals tightly packed around the nucleus .
in section [ intro ] , we discussed how dmrg can be useful for ab initio quantum chemistry , and we gave an overview of dmrg - related methods .
these methods can be divided into two categories : dmrg can play the role of a large active space fci solver , or it can provide an approximate mps wavefunction , on which excitations can be built .
the dmrg algorithm was introduced in section [ remarks ] , where we discussed the use of complementary operators and how to overcome convergence difficulties .
both issues have to be addressed for dmrg to be an efficient and reliable approach for ab initio quantum chemistry . with symmetry - adapted dmrg ,
a huge performance gain can be obtained both in computation time and memory .
section [ symm ] introduced an mps ansatz which is an exact eigenstate of the symmetry group of the hamiltonian .
the wigner - eckart theorem allows the introduction of a sparse block structure in this ansatz . for non - abelian groups
, the wigner - eckart theorem also allows for data compression .
an overview of the high - level structure of chemps2 is given in section [ ourcode ] .
the required input for the ` chemps2::dmrg ` class and its output are discussed .
a dmrg - scf algorithm was implemented in ` chemps2::casscf ` .
section [ ourcode ] should help new users to understand the provided tests , and to alter them to their own needs . as an application
, we have calculated the 12 lowest bond dissociation curves of the carbon dimer at the dmrg(28o , 12e , d@xmath3=2500)/cc - pvdz level of theory .
in addition , we assessed the contribution of @xmath1 core correlation to the @xmath2 bond dissociation curve of the carbon dimer by comparing calculations at the dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz and dmrg - scf(34o , 8e , d@xmath3=2500)/cc - pcvdz levels of theory .
these results were presented in section [ c2 ] .
the low - lying bond dissociation curves of the carbon dimer were resolved with chemps2 to sub-@xmath5 accuracy .
the non - parallelity due to @xmath1 core correlation is of the order of 2 @xmath5 in the cc - pcvdz basis . in the future
, we would like to incorporate the two - orbital mutual information @xmath61 @xcite in chemps2 , as well as its gradient and hessian , to retrieve optimal orbitals and their corresponding ordering , as discussed in section [ convergencesection ] .
we are also working on an mpi implementation of chemps2 , in which the product @xmath46 $ ] is distributed over several processors .
each processor is then responsible for certain renormalized operators @xcite .
updated versions of chemps2 will be provided at its public git repository @xcite . the oxo - mn(salen )
complex @xcite is a great challenge for molecular electronic structure methods .
we are currently performing large active space dmrg - scf calculations with chemps2 to provide new insights in the relative order of the lowest singlet , triplet , and quintet states . understanding the active space structure of this complex and several of its transition states
will be of benefit for the experimentalists in our group @xcite .
s.w . received a ph.d .
fellowship from the research foundation flanders ( fwo vlaanderen ) .
w.p . acknowledges support from a project funded by the research foundation flanders ( fwo vlaanderen ) .
p.w.a . acknowledges support from nserc .
this work was carried out using the stevin supercomputer infrastructure at ghent university , funded by ghent university , the hercules foundation and the flemish government - department ewi .
note that during a sweep , we work with left - normalized tensors to the left and right - normalized tensors to the right of the current position . consider the following renormalized partial hamiltonian term in the graphical notation @xcite : @xmath155{diagram1.pdf } \label{example1 } } } \label{example1 } % \vcenter{\hbox{\scriptsize { % \setlength{\unitlength}{1 cm } % \begin{picture}(2.6,2 ) % \put(0.4,1.8){\circle{0.6 } } % \put(0.7,1.8){\line(1,0){0.65 } } % \put(1.4,1.7){$j_r j_r^z n_r i_r \alpha_r$ } % \put(0.18,1.7){a[k ] } % \put(0.4,1.3){\line(0,1){0.2 } } % \put(0.1,0.7){\line(1,0){0.6 } } % \put(0.1,1.3){\line(1,0){0.6 } } % \put(0.1,0.7){\line(0,1){0.6 } } % \put(0.7,0.7){\line(0,1){0.6 } } % \put(0.2,0.9){$\hat{a}_{k \sigma}$ } % \put(0.4,0.5){\line(0,1){0.2 } } % \put(0.4,0.2){\circle{0.6 } } % \put(0.7,0.2){\line(1,0){0.65 } } % \put(0.18,0.1){a[k ] } % \put(1.40,0.1){$\widetilde{j}_r \widetilde{j}_r^z \widetilde{n}_r \widetilde{i}_r \widetilde{\alpha}_r$ } % \put(0.1,1.0){\oval(1.0,1.6)[l ] } % \end{picture } % } } } \label{example1}\ ] ] with ( [ tensordecomp ] ) , it is easy to show that ( [ example1 ] ) can be written as @xmath156{diagram2.pdf } \label{againwe } } } % \vcenter{\hbox{\scriptsize { % \setlength{\unitlength}{1 cm } % \begin{picture}(2.0,2.4 ) % \put(0.1,0.9){\line(1,0){0.6 } } % \put(0.1,1.5){\line(1,0){0.6 } } % \put(0.1,0.9){\line(0,1){0.6 } } % \put(0.7,0.9){\line(0,1){0.6 } } % \put(0.2,1.1){l[k ] } % \put(0.4,1.5){\line(0,1){0.5 } } % \put(0.4,0.4){\line(0,1){0.5 } } % \put(0.4,2.0){\line(1,0){0.50 } } % \put(0.4,0.4){\line(1,0){0.50 } } % \put(0.95,1.9){$j_r n_r i_r \alpha_r$ } % \put(0.95,0.43){$\widetilde{j}_r ( n_r + 1)$ } % \put(0.95,0.17){$(i_r \otimes i_k ) \widetilde{\alpha}_r$ } % \end{picture } % } } } \label{againwe } \vspace{0.01\textwidth}\ ] ] with @xmath157 } % \put(0.4,1.5){\line(0,1){0.5 } } % \put(0.4,0.4){\line(0,1){0.5 } } % \put(0.4,2.0){\line(1,0){0.50 } } % \put(0.4,0.4){\line(1,0){0.50 } } % \put(0.95,1.9){$j_r n_r i_r \alpha_r$ } % \put(0.95,0.43){$\widetilde{j}_r ( n_r + 1)$ } % \put(0.95,0.17){$(i_r \otimes
i_k ) \widetilde{\alpha}_r$ } % \end{picture } % } } } & \vcenter{\hbox { \includegraphics[height=0.12\textwidth]{diagram2.pdf } } } % = \sum\limits_{\alpha_l } \vcenter{\hbox{\scriptsize { % \setlength{\unitlength}{1 cm } % \begin{picture}(3.7,2 ) % \put(1.1,1.8){\circle{0.6 } } % \put(1.4,1.8){\line(1,0){0.50 } } % \put(1.95,1.7){$j_r n_r i_r \alpha_r$ } % \put(0.9,1.7){t[k ] } % \put(1.1,1.2){\line(0,1){0.3 } } % \put(1.2,1.3){$0 0 i_0 $ } % \put(1.1,0.5){\line(0,1){0.3 } } % \put(1.2,0.55){$\frac{1}{2 } 1 i_k$ } % \put(1.1,0.2){\circle{0.6 } } % \put(1.4,0.2){\line(1,0){0.50 } } % \put(0.9,0.1){t[k ] } % \put(1.95,0.23){$\widetilde{j}_r ( n_r+1)$ } % \put(1.95,-0.03){$(i_r \otimes i_k ) \widetilde{\alpha}_r$ } % \put(0.8,1.0){\oval(1.0,1.6)[l ] } % \put(0.0,0.3){\rotatebox{90}{$j_r n_r i_r \alpha_l$ } } % \end{picture } % } } } \nonumber \\ = \sum\limits_{\alpha_l } \vcenter{\hbox { \includegraphics[height=0.12\textwidth]{diagram3.pdf } } } \nonumber \\ % & + ( -1)^{\widetilde{j}_r - j_r + \frac{1}{2 } } \sqrt{\frac{2 j_r + 1}{2 \widetilde{j}_r + 1}}\sum\limits_{\alpha_l } \vcenter{\hbox{\scriptsize { % \setlength{\unitlength}{1 cm } % \begin{picture}(3.7,2 ) % \put(1.5,1.8){\circle{0.6 } } % \put(1.8,1.8){\line(1,0){0.5 } } % \put(2.35,1.7){$j_r n_r i_r \alpha_r$ } % \put(1.3,1.7){t[k ] } % \put(1.5,1.2){\line(0,1){0.3 } } % \put(1.6,1.3){$\frac{1}{2 } 1 i_k$ } % \put(1.5,0.5){\line(0,1){0.3 } } % \put(1.6,0.5){$0 2 i_0 $ } % \put(1.5,0.2){\circle{0.6 } } % \put(1.8,0.2){\line(1,0){0.5 } } % \put(1.3,0.1){t[k ] } % \put(2.35,0.23){$\widetilde{j}_r ( n_r+1)$ } % \put(2.35,-0.03){$(i_r \otimes i_k ) \widetilde{\alpha}_r$ } % \put(1.2,1.0){\oval(1.0,1.6)[l ] } % \put(0.0,0.3){\rotatebox{90}{$\widetilde{j}_r ( n_r-1)$ } } % \put(0.4,0.3){\rotatebox{90}{$(i_r \otimes
i_k ) \alpha_l$ } } % \end{picture } % } } } & + ( -1)^{\widetilde{j}_r - j_r + \frac{1}{2 } } \sqrt{\frac{2 j_r + 1}{2 \widetilde{j}_r + 1}}\sum\limits_{\alpha_l } \vcenter{\hbox { \includegraphics[height=0.12\textwidth]{diagram4.pdf } } } \end{aligned}\ ] ] eq . ( [ example1 ] ) can hence be factorized into clebsch - gordan coefficients and a reduced spin-@xmath158 @xmath159-tensor .
the @xmath159-tensor has spin-@xmath158 because @xmath160 is a spin-@xmath158 operator .
it is shown in ref .
@xcite , that for two second quantized operators acting on different sites , the renormalized operator can be decomposed into two terms : one with a spin-0 reduced tensor and one with a spin-1 reduced tensor .
this follows from @xmath69 representation theory : @xmath161 .
s. rommer , s. stlund , class of ansatz wave functions for one - dimensional spin systems and their relation to the density matrix renormalization group , phys .
b 55 ( 1997 ) 21642181 . http://dx.doi.org/10.1103/physrevb.55.2164 [ ] .
s. daul , i. ciofini , c. daul , s. r. white , full - ci quantum chemistry using the density matrix renormalization group , int . j. quantum chem .
79 ( 6 ) ( 2000 ) 331342 .
http://dx.doi.org/10.1002/1097-461x(2000)79:6<331::aid-qua1>3.0.co;2-y [ ] .
a. o. mitrushenkov , g. fano , f. ortolani , r. linguerri , p. palmieri , quantum chemistry using the density matrix renormalization group , j. chem .
115 ( 15 ) ( 2001 ) 68156821 . http://dx.doi.org/10.1063/1.1389475 [ ] .
g. k .- l .
chan , m. head - gordon , highly correlated calculations with a polynomial cost algorithm : a study of the density matrix renormalization group , j. chem .
116 ( 11 ) ( 2002 ) 44624476 . http://dx.doi.org/10.1063/1.1449459 [ ] .
o. legeza , j. rder , b. a. hess , controlling the accuracy of the density - matrix renormalization - group method : the dynamical block state selection approach , phys .
b 67 ( 2003 ) 125114 . http://dx.doi.org/10.1103/physrevb.67.125114 [ ] .
g. k .- l .
chan , m. head - gordon , exact solution ( within a triple - zeta , double polarization basis set ) of the electronic schrdinger equation for water , j. chem .
phys . 118 ( 19 ) ( 2003 ) 85518554 .
http://dx.doi.org/10.1063/1.1574318 [ ] .
a. o. mitrushenkov , r. linguerri , p. palmieri , g. fano , quantum chemistry using the density matrix renormalization group ii , j. chem . phys .
119 ( 8) ( 2003 ) 41484158 .
[ ] .
g. k .- l .
chan , m. kllay , j. gauss , state - of - the - art density matrix renormalization group and coupled cluster theory studies of the nitrogen binding curve , j. chem .
121 ( 13 ) ( 2004 ) 61106116 . http://dx.doi.org/10.1063/1.1783212 [ ] .
o. legeza , j. slyom , quantum data compression , quantum information generation , and the density - matrix renormalization - group method , phys .
b 70 ( 2004 ) 205118 .
http://dx.doi.org/10.1103/physrevb.70.205118 [ ] .
g. moritz , b. a. hess , m. reiher , convergence behavior of the density - matrix renormalization group algorithm for optimized orbital orderings , j. chem .
122 ( 2 ) ( 2005 ) 024107 .
http://dx.doi.org/10.1063/1.1824891 [ ] .
g. k .- l .
chan , t. v. voorhis , density - matrix renormalization - group algorithms with nonorthogonal orbitals and non - hermitian operators , and applications to polyenes , j. chem . phys .
122 ( 20 ) ( 2005 ) 204101 . http://dx.doi.org/10.1063/1.1899124 [ ] .
g. moritz , m. reiher , construction of environment states in quantum - chemical density - matrix renormalization group calculations , j. chem .
124 ( 3 ) ( 2006 ) 034103 .
http://dx.doi.org/10.1063/1.2139998 [ ] .
j. hachmann , w. cardoen , g. k .- l .
chan , multireference correlation in long molecules with the quadratic scaling density matrix renormalization group , j. chem .
125 ( 14 ) ( 2006 ) 144101 . http://dx.doi.org/10.1063/1.2345196 [ ] .
j. hachmann , j. j. dorando , m. avils , g. k .- l .
chan , the radical character of the acenes : a density matrix renormalization group study , j. chem .
127 ( 13 ) ( 2007 ) 134309 .
http://dx.doi.org/10.1063/1.2768362 [ ] .
k. h. marti , i. m. ondk , g. moritz , m. reiher , density matrix renormalization group calculations on relative energies of transition metal complexes and clusters , j. chem .
128 ( 1 ) ( 2008 ) 014104 . http://dx.doi.org/10.1063/1.2805383 [ ] .
d. zgid , m. nooijen , the density matrix renormalization group self - consistent field method : orbital optimization with the density matrix renormalization group method in the active space , j. chem .
phys . 128 ( 14 ) ( 2008 ) 144116 . http://dx.doi.org/10.1063/1.2883981 [ ] .
d. ghosh , j. hachmann , t. yanai , g. k .- l .
chan , orbital optimization in the density matrix renormalization group , with applications to polyenes and beta - carotene , j. chem .
128 ( 14 ) ( 2008 ) 144117 . http://dx.doi.org/10.1063/1.2883976 [ ] .
t. yanai , y. kurashige , d. ghosh , g. k .- l .
chan , accelerating convergence in iterative solution for large - scale complete active space self - consistent - field calculations , int .
j. quantum chem .
109 ( 10 ) ( 2009 ) 21782190 . http://dx.doi.org/10.1002/qua.22099 [ ] .
y. kurashige , t. yanai , high - performance ab initio density matrix renormalization group method : applicability to large - scale multireference problems for metal compounds , j. chem .
130 ( 23 ) ( 2009 ) 234114 . http://dx.doi.org/10.1063/1.3152576 [ ] .
t. yanai , y. kurashige , e. neuscamman , g. k .- l .
chan , multireference quantum chemistry through a joint density matrix renormalization group and canonical transformation theory , j. chem .
132 ( 2 ) ( 2010 ) 024105 .
http://dx.doi.org/10.1063/1.3275806 [ ] .
w. mizukami , y. kurashige , t. yanai , communication : novel quantum states of electron spins in polycarbenes from ab initio density matrix renormalization group calculations , j. chem .
133 ( 9 ) ( 2010 ) 091101 .
http://dx.doi.org/10.1063/1.3476461 [ ] .
g. barcza , o. legeza , k. h. marti , m. reiher , quantum - information analysis of electronic states of different molecular structures , phys . rev .
a 83 ( 2011 ) 012508 .
http://dx.doi.org/10.1103/physreva.83.012508 [ ] .
y. kurashige , t. yanai , second - order perturbation theory with a density matrix renormalization group self - consistent field reference function : theory and application to the study of chromium dimer , j. chem .
135 ( 9 ) ( 2011 ) 094104 . http://dx.doi.org/10.1063/1.3629454 [ ] .
a. o. mitrushchenkov , g. fano , r. linguerri , p. palmieri , on the importance of orbital localization in qc - dmrg calculations , int .
j. quantum chem .
112 ( 6 ) ( 2012 ) 16061619 . http://dx.doi.org/10.1002/qua.23173 [ ] .
s. wouters , p. a. limacher , d. van neck , p.
w. ayers , longitudinal static optical properties of hydrogen chains : finite field extrapolations of matrix product state calculations , j. chem .
136 ( 13 ) ( 2012 ) 134110 . http://dx.doi.org/10.1063/1.3700087 [ ] .
t. yanai , y. kurashige , e. neuscamman , g. k .- l .
chan , extended implementation of canonical transformation theory : parallelization and a new level - shifted condition , phys . chem .
14 ( 2012 ) 78097820 . http://dx.doi.org/10.1039/c2cp23767a [ ] .
k. boguslawski , p. tecmer , o. legeza , m. reiher , entanglement measures for single- and multireference correlation effects , j. phys .
3 ( 21 ) ( 2012 ) 31293135 .
http://dx.doi.org/10.1021/jz301319v [ ] .
w. mizukami , y. kurashige , t. yanai , more @xmath162 electrons make a difference : emergence of many radicals on graphene nanoribbons studied by ab initio dmrg theory , j. chem .
theory comput . 9 ( 1 ) ( 2013 ) 401407 .
http://dx.doi.org/10.1021/ct3008974 [ ] .
y. ma , h. ma , assessment of various natural orbitals as the basis of large active space density - matrix renormalization group calculations , j. chem .
138 ( 22 ) ( 2013 ) 224105 . http://dx.doi.org/10.1063/1.4809682 [ ] .
m. saitow , y. kurashige , t. yanai , multireference configuration interaction theory using cumulant reconstruction with internal contraction of density matrix renormalization group wave function , j. chem .
139 ( 4 ) ( 2013 ) 044118 .
[ ] . f. liu , y. kurashige , t. yanai , k. morokuma , multireference ab initio density matrix renormalization group ( dmrg)-casscf and dmrg - caspt2 study on the photochromic ring opening of spiropyran , j. chem . theory comput . 9 ( 10 ) ( 2013 ) 44624469 . http://dx.doi.org/10.1021/ct400707k [ ] .
p. tecmer , k. boguslawski , o. legeza , m. reiher , unravelling the quantum - entanglement effect of noble gas coordination on the spin ground state of cuo , phys .
16 ( 2014 ) 719727
. http://dx.doi.org/10.1039/c3cp53975j [ ] .
k. hallberg , density matrix renormalization , in : d. snchal , a .- m .
tremblay , c. bourbonnais ( eds . ) , theoretical methods for strongly correlated electrons , crm series in mathematical physics , springer new york , 2004 , pp . 337 .
http://dx.doi.org/10.1007/0-387-21717-7_1 [ ] .
b. pirvu , j. haegeman , f. verstraete , matrix product state based algorithm for determining dispersion relations of quantum spin chains with periodic boundary conditions , phys .
b 85 ( 2012 ) 035130 . http://dx.doi.org/10.1103/physrevb.85.035130 [ ] .
j. haegeman , b. pirvu , d. j. weir , j. i. cirac , t. j. osborne , h. verschelde , f. verstraete , variational matrix product ansatz for dispersion relations , phys .
b 85 ( 2012 ) 100408 . http://dx.doi.org/10.1103/physrevb.85.100408 [ ] .
s. wouters , n. nakatani , d. van neck , g. k .- l .
chan , thouless theorem for matrix product states and subsequent post density matrix renormalization group methods , phys .
b 88 ( 2013 ) 075122 . http://dx.doi.org/10.1103/physrevb.88.075122 [ ] . n. nakatani , s. wouters , d. van neck , g .-
chan , linear response theory for the density matrix renormalization group : efficient algorithms for strongly correlated excited states , eprint arxiv:1311.1646 .
http://arxiv.org/abs/1311.1646 [ ] .
j. haegeman , j. i. cirac , t. j. osborne , i. pi orn , h. verschelde , f. verstraete , time - dependent variational principle for quantum lattices , phys .
( 2011 ) 070601 .
http://dx.doi.org/10.1103/physrevlett.107.070601 [ ] .
n. nakatani , g. k .- l .
chan , efficient tree tensor network states ( ttns ) for quantum chemistry : generalizations of the density matrix renormalization group algorithm , j. chem .
138 ( 13 ) ( 2013 ) 134113 . http://dx.doi.org/10.1063/1.4798639 [ ] .
k. h. marti , b. bauer , m. reiher , m. troyer , f. verstraete , complete - graph tensor network states : a new fermionic wave function ansatz for molecules , new j. phys . 12 ( 10 ) ( 2010 ) 103008 . http://dx.doi.org/10.1088/1367-2630/12/10/103008 [ ] .
e. r. davidson , the iterative calculation of a few of the lowest eigenvalues and corresponding eigenvectors of large real - symmetric matrices , j. comput .
17 ( 1 ) ( 1975 ) 8794 . http://dx.doi.org/10.1016/0021-9991(75)90065-0 [ ] .
g. k .- l .
chan , p. w. ayers , e. s. i. croot , on the distribution of eigenvalues of grand canonical density matrices , j. stat . phys .
109 ( 1 - 2 ) ( 2002 ) 289299 . http://dx.doi.org/10.1023/a:1019999930923 [ ] .
b. verstichel , w. poelmans , s. de baerdemacker , s. wouters , d. van neck , v2 dm study of the 2d hubbard model : benchmark results with three - index conditions and extended cluster constraints , eprint arxiv:1307.1002 .
http://arxiv.org/abs/1307.1002 [ ] . j. m. turney , a. c. simmonett , r. m. parrish , e. g. hohenstein , f. a. evangelista , j. t. fermann , b. j. mintz , l. a. burns , j. j. wilke , m. l. abrams , n. j. russ , m. l. leininger , c. l. janssen , e. t. seidl , w. d. allen , h. f. schaefer , r. a. king , e. f. valeev , c. d. sherrill , t. d. crawford , psi4 : an open - source ab initio electronic structure program , wires comput .
sci . 2 ( 4 ) ( 2012 ) 556565 .
http://dx.doi.org/10.1002/wcms.93 [ ] .
e. m. siegbahn , j. almlf , a. heiberg , b. o. roos , the complete active space scf ( casscf ) method in a newton - raphson formulation with application to the hno molecule , j. chem .
74 ( 4 ) ( 1981 ) 23842396 .
http://dx.doi.org/10.1063/1.441359 [ ] .
s. shaik , d. danovich , w. wu , p. su , h. s. rzepa , p. c. hiberty , quadruple bonding in c2 and analogous eight - valence electron species , nat .
4 ( 2012 ) 195200 .
http://dx.doi.org/10.1038/nchem.1263 [ ] .
k. a. peterson , a. k. wilson , d. e. woon , t. h. dunning jr .
, benchmark calculations with correlated molecular wave functions xii . core correlation effects on the homonuclear diatomic molecules b2-f2 , theor .
97 ( 1 - 4 ) ( 1997 ) 251259 .
http://dx.doi.org/10.1007/s002140050259 [ ] .
d. shi , x. zhang , j. sun , z. zhu , mrci study on spectroscopic and molecular properties of @xmath163 , @xmath164 , @xmath137 , @xmath165 , @xmath166 and @xmath167 electronic states of the c2 radical , mol .
109 ( 11 ) ( 2011 ) 14531465 . http://dx.doi.org/10.1080/00268976.2011.564593 [ ] .
m. boggio - pasqua , a. voronin , p. halvick , j .- c .
rayez , analytical representations of high level ab initio potential energy curves of the c2 molecule , j. mol .
theochem 531 ( 1 - 3 ) ( 2000 ) 159 167 .
http://dx.doi.org/10.1016/s0166-1280(00)00442-5 [ ] .
m. l. abrams , c. d. sherrill , full configuration interaction potential energy curves for the @xmath168 , @xmath163 , and @xmath132 states of c2 : a challenge for approximate methods , j. chem .
121 ( 19 ) ( 2004 ) 92119219 . http://dx.doi.org/10.1063/1.1804498 [ ] .
a. j. c. varandas , extrapolation to the complete - basis - set limit and the implications of avoided crossings : the @xmath168 , @xmath163 , and @xmath132 states of c2 , j. chem .
129 ( 23 ) ( 2008 ) 234103 . http://dx.doi.org/10.1063/1.3036115 [ ] .
d. l. kokkin , g. b. bacskay , t. w. schmidt , oscillator strengths and radiative lifetimes for c2 : swan , ballik - ramsay , phillips , and @xmath169 systems , j. chem .
126 ( 8) ( 2007 ) 084302 .
http://dx.doi.org/10.1063/1.2436879 [ ] .
w. jiang , a. k. wilson , multireference composite approaches for the accurate study of ground and excited electronic states : c2 , n2 , and o2 , j. chem .
134 ( 3 ) ( 2011 ) 034101 .
http://dx.doi.org/10.1063/1.3514031 [ ] .
c. d. sherrill , p. piecuch , the @xmath168 , @xmath163 and @xmath132 states of c2 : a comparison of renormalized coupled - cluster and multireference methods with full configuration interaction benchmarks , j. chem .
122 ( 12 ) ( 2005 ) 124104 . http://dx.doi.org/10.1063/1.1867379 [ ] .
u. s. mahapatra , s. chattopadhyay , r. k. chaudhuri , molecular applications of state - specific multireference perturbation theory to hf , h2o , h2s , c2 , and n2 molecules , j. chem .
129 ( 2 ) ( 2008 ) 024108 .
http://dx.doi.org/10.1063/1.2952666 [ ] .
w. purwanto , s. zhang , h. krakauer , excited state calculations using phaseless auxiliary - field quantum monte carlo : potential energy curves of low - lying c2 singlet states , j. chem .
130 ( 9 ) ( 2009 ) 094107 . http://dx.doi.org/10.1063/1.3077920 [ ]
. g. h. booth , d. cleland , a. j. w. thom , a. alavi , breaking the carbon dimer : the challenges of multiple bond dissociation with full configuration interaction quantum monte carlo methods , j. chem .
135 ( 8) ( 2011 )
084104 . http://dx.doi.org/10.1063/1.3624383 [ ] .
d. e. woon , t. h. dunning , gaussian basis sets for use in correlated molecular calculations . v. core - valence basis sets for boron through neon , j. chem .
103 ( 11 ) ( 1995 ) 4572 4585 . http://dx.doi.org/10.1063/1.470645 [ ] .
j. ivanic , j. r. collins , s. k. burt , theoretical study of the low lying electronic states of oxox(salen ) ( x = mn , mn- , fe , and cr- ) complexes , j. phys .
a 108 ( 12 ) ( 2004 ) 23142323 .
http://dx.doi.org/10.1021/jp031214g [ ] .
t. bogaerts , a. van yperen - de deyne , y .- y .
liu , f. lynen , v. van speybroeck , p. van der voort , mn - salen@mil101(al ) : a heterogeneous , enantioselective catalyst synthesized using a bottle around the ship approach , chem .
49 ( 2013 ) 80218023 . http://dx.doi.org/10.1039/c3cc44473b [ ] . | the density matrix renormalization group ( dmrg ) has become an indispensable numerical tool to find exact eigenstates of finite - size quantum systems with strong correlation . in the fields of condensed matter , nuclear structure and molecular electronic structure
, it has significantly extended the system sizes that can be handled compared to full configuration interaction , without losing numerical accuracy . for quantum chemistry ( qc ) ,
the most efficient implementations of dmrg require the incorporation of particle number , spin and point group symmetries in the underlying matrix product state ( mps ) ansatz , as well as the use of so - called complementary operators .
the symmetries introduce a sparse block structure in the mps ansatz and in the intermediary contracted tensors .
if a symmetry is non - abelian , the wigner - eckart theorem allows to factorize a tensor into a clebsch - gordan coefficient and a reduced tensor .
in addition , the fermion signs have to be carefully tracked .
because of these challenges , implementing dmrg efficiently for qc is not straightforward .
efficient and freely available implementations are therefore highly desired . in this work we present chemps2 , our free open - source spin - adapted implementation of dmrg for ab initio qc . around chemps2
, we have implemented the augmented hessian newton - raphson complete active space self - consistent field method , with exact hessian .
the bond dissociation curves of the 12 lowest states of the carbon dimer were obtained at the dmrg(28 orbitals , 12 electrons , d@xmath0=2500)/cc - pvdz level of theory .
the contribution of @xmath1 core correlation to the @xmath2 bond dissociation curve of the carbon dimer was estimated by comparing energies at the dmrg(36o , 12e , d@xmath3=2500)/cc - pcvdz and dmrg - scf(34o , 8e , d@xmath3=2500)/cc - pcvdz levels of theory .
density matrix renormalization group , matrix product state , su(2 ) spin symmetry , abelian point group symmetry , ab initio quantum chemistry * program summary * _ manuscript title : _ chemps2 : a free open - source spin - adapted implementation of the density matrix renormalization group for ab initio quantum chemistry + _ authors : _ sebastian wouters , ward poelmans , paul w. ayers , and dimitri van neck + _ program title : _ chemps2 + _ journal reference : _ + _ catalogue identifier : _
+ _ licensing provisions : _ gnu
general public license version 2 + _ programming language : _
c++ + _ computer : _ x86 - 64 + _ operating system : _ scientific linux 6.0 + _ ram : _ 10 mb - 64 gb + _ number of processors used : _ 1 - 16 ( single node ) + _ supplementary material : _ doxygen documentation can be generated + _ keywords : _ density matrix renormalization group , matrix product state , su(2 ) spin symmetry , abelian point group symmetry , ab initio quantum chemistry + _ classification : _
16.1 molecular physics and physical chemistry : structure and properties + _ external routines / libraries : _ basic linear algebra subprograms ( blas ) , linear algebra package ( lapack ) , gnu scientific library ( gsl ) , and hierarchical data format release 5 ( hdf5 ) + _ nature of problem : _ + the many - body hilbert space grows exponentially with the number of single - particle states .
exact diagonalization solvers can therefore only handle small systems , of up to 18 electrons in 18 orbitals .
interesting active spaces are often significantly larger .
+ _ solution method : _ + the density matrix renormalization group allows to extend the size of active spaces , for which numerically exact solutions can be found , to about 40 electrons in 40 orbitals .
in addition , it provides a rigorous variational upper bound to energies , as it has an underlying wavefunction ansatz , the matrix product state .
+ _ restrictions : _
+ our implementation of the density matrix renormalization group is spin - adapted .
this means that targeted eigenstates in the active space are exact eigenstates of the total electronic spin operator .
hamiltonians which break this symmetry ( a magnetic field term for example ) can not be handled by our code .
as electron repulsion integrals in gaussian basis sets have eightfold permutation symmetry , we have used this property in our code .
+ _ unusual features : _ + the nature of the matrix product state ansatz allows for exact spin coupling . in chemps2 , the _ total _ electronic spin is imposed ( not just the spin projection ) , in addition to the particle - number and abelian point - group symmetries . + _ running time : _ + the running time depends on the size of the targeted active space , the number of desired eigenstates , their symmetry , the density of states , the individual orbital symmetries , the orbital ordering , the desired level of convergence , and the chosen convergence scheme . to converge a single point of one of the dissociation curves of the carbon dimer ( @xmath4 symmetry ) in the cc - pvdz basis ( 28 orbitals ; their ordering is described in section [ secirreporder ] ) with 2500 reduced renormalized basis states ( see the convergence scheme in section [ howtoextrapolateenergies ] ; the variational energy then lies 0.1 @xmath5 above the fully converged result ) takes about 8 hours on a single node with a dual - socket octa - core intel xeon sandy bridge ( e5 - 2670 ) ( 16 cores at 2.6 ghz ) , and requires 6 gb of ram . |
gamma ray burst(grbs ) are extremely high energy releasing phenomena in the universe and are usually divided into two classes @xcite : short grbs ( @xmath6s ) and long grbs ( @xmath7s ) .
numerous models have been proposed to explore the central engines of grbs , and one of the mostly discussed model is the neutrino dominated accretion flows ( ndafs ) with a hyper accreting stellar massive black hole with accretion rate @xmath8/s . due to the high density and high temperature in the inner part of ndafs ,
the optical depth of photons is very large and photons are completely trapped , then neutrinos and antineutrinos become the most promising candidates that carry away thermal energy and cool the disk .
the annihilation of neutrino pairs above the disk is believed to be the energy source of grbs .
@xcite first proposed that neutrino pairs annihilation into electron pairs during the merger of compact objects binaries may power grbs .
after that , @xcite investigated ndafs under the assumption that the disk is transparent to neutrinos , but they also pointed out that the assumption fails when the accretion rate is higher than @xmath5/s .
@xcite improved the model by using a simplified neutrino transport model which was believed to bridge the neutrino optically thin limit and optically thick limit .
@xcite improved the model further by dealing with the neutrino emission and chemical composition in the optically thin regime , optically thick regime and intermediate regime separately .
though many works on ndafs have confirmed the validity of the ndafs model as the central engine of grbs @xcite , there are still some uncertainties including : 1 , the distribution of electron fraction . in many previous works , which is assumed to be a constant value , for example , 0.5 , throughout the disk . while the electron fraction has a great effect on the emission of neutrinos and antineutrinos , as shown by @xcite , so need more cautious disposal .
2 , neutrino transport and neutrino spectra .
the most commonly used approximation was the simplified neutrino transport model introduced in @xcite
. in their model , the difference between the neutrino and antineutrino transport in the disk was neglected , but as shown in @xcite , the precise spectra of neutrino and antineutrino sensitively determines the annihilation luminosity of neutrino pairs .
3 , the annihilation of neutrino pairs .
the most common method for calculating the annihilation luminosity was originally introduced by to calculate the annihilation luminosity during the merger of neutron star binaries .
this method was applied in the calculation of the annihilation luminosity above ndafs under the assumption that the emission of neutrinos and antineutrinos are isotropic and symmetric @xcite .
it is evident that the most strict approach to determine the neutrino / antineutrino luminosity and their annihilation luminosity above ndafs is to build the two - dimensional disk model in which neutrino transport , vertical structure , chemical evolution , thermal evolution and the distribution of electron fraction , mass density , and temperature are self consistently considered .
first investigated the vertical structure of ndafs by using the eddington approximation to deal with the neutrino transport in the vertical direction , neglecting the contribution of advection term to the disk cooling , and dealing with neutrino emission and chemical composition following the similar method of @xcite and @xcite .
recently , @xcite also investigated the vertical structure of ndafs with many simplifications on neutrino transport , equation of state and the annihilation efficiency .
the first one is that the neutrino emission is directly integrated to calculate the neutrino luminosity by neglecting the absorption of neutrinos , which is viable in the neutrino optically thin limit @xcite ; the second one , that a simplified equation of state @xmath9 is used in the vertical direction , which is viable when relativistic degenerate electrons dominate the pressure of the disk ; the third one that a toy annihilation efficiency of neutrino pairs @xmath10 is applied , where @xmath11 is the annihilation luminosity and neutrino luminosity before annihilation respectively , and @xmath12 is the so called annihilation volume . with all the above simplifications , the annihilation efficiency @xmath13 can even reach @xmath14 !
obviously , the unrealistic result is caused by too many unrealistic assumptions . in this work ,
we investigate the vertical structure of ndafs , neutrino / antineutrino luminosity and their annihilation luminosity by self - consistently considering the neutrino / antineutrino transport , the vertical hydrostatic equilibrium , the precise equation of state , the chemical equilibrium and the thermal balance between neutrino cooling , advection cooling and the viscosity heating under the self similar assumption of the distribution of mass density and internal energy density in the radial direction .
especially , we strictly solve the bolzmann equation to deal with the neutrino transport , instead of taking the assumption of the gray body spectra @xcite or the eddington approximation .
correspondingly , we can precisely obtain the energy spectra of neutrino pairs . combining the conservation of the number of lepton , the distribution of chemical compositions are self - consistently and accurately determined .
this paper is organized as follows . in
ii , we introduce the basic equations in our calculation , including the boltzmann equation of neutrino / antineutrino transport , angular momentum equation , hydrostatic equilibrium equation , equation of state , lepton number conservation equation and thermal evolution equation . in iii , we briefly introduce our numerical methods to find the steady solution of the structure of the disk . in
iv , we list our numerical results of the disk structure , neutrino / antineutrino luminosity and their annihilation luminosity . conclusions and discussions are summarized in v.
we assume a steady accretion disk with accretion rate @xmath15/s around a central black hole with mass @xmath16 and adopt the standard @xmath17 viscosity prescription of with @xmath3 for the viscous stress of the disk .
we discuss the structure of the disk and neutrino transport in cylindrical coordinate @xmath18 and we assume the inner boundary of the disk to be @xmath19 and outer boundary to be @xmath20 .
we solve the one dimensional boltzmann equation of neutrino and antineutrino transport in the vertical direction of the disk and obtain the energy dependent and direction dependent neutrino spectrum .
we define @xmath21 and @xmath22 to be the distribution function for up moving neutrinos / antineutrinos and down moving ones respectively , where @xmath23 is the vertical coordinate of the disk , @xmath24 is the energy of neutrinos / antineutrinos , and @xmath25 for up moving neutrinos / antineutrinos and @xmath26 for down moving ones , where @xmath27 is the angle of neutrinos / antineutrinos moving direction to the vertical direction of the disk . for the up - moving neutrinos / antineutrinos , their distribution function is determined by @xcite : @xmath28+\lambda_s\left[-f_+(z , p,\mu)+\frac{1}{2 } \int_0 ^ 1 d\mu f_-(z , p,\mu)+f_+(z , p,\mu)\right],\ ] ] and for the down - moving ones , their distribution function is determined by @xmath29-\lambda_s\left[-f_-(z , p,\mu)+\frac{1}{2 } \int_0 ^ 1 d\mu f_-(z , p,\mu)+f_+(z , p,\mu)\right].\ ] ] where @xmath30 is the absorption coefficient and @xmath31 is the scattering coefficient of neutrinos / antineutrinos , and here @xmath32 for neutrinos , @xmath33 for anti - neutrinos , where @xmath34 , and @xmath35 is the chemical potential of electron , proton and neutron , respectively . because urca process @xmath36 and @xmath37 dominate the creation and the absorption of neutrinos and antineutrinos , and the neutrino / antineutrino scattering by neutrons and protons @xmath38 and @xmath39 dominates the scattering opacity @xcite , so we include no other neutrino / antineutrino processes .
thus , for simplicity , in this paper we use @xmath40 and @xmath41 , @xmath42 and @xmath43 interchangeably and it will not cause any confusion . as for the explicit expression of the absorption coefficient @xmath30 and scattering coefficient @xmath31 of neutrinos / antineutrinos , please refer to @xcite .
considering that the disk is symmetric about the equator plane , the boundary conditions for the distribution function @xmath44 of neutrinos / antineutrinos can be written as @xmath45 and @xmath46 , where @xmath47 is the upper boundary of the disk . adopting @xmath17 prescription ,
the tangential stress and angular momentum equation can be written as , @xmath48 and @xmath49 where @xmath50 is the viscous stress , @xmath51 is the mass density of the disk , @xmath52 is the acoustic speed , @xmath53 is the radial drift velocity and @xmath54 is kepler angular velocity at radius @xmath55 . integrating eq.(4 ) over radius @xmath55 and height @xmath23 , we obtain @xmath56 and by taking into consideration the steady accretion condition @xmath57 and using torsion condition in the inner boundary @xmath58 , the angular momentum equation eq.(5 ) is transformed to be @xmath59 the vertical hydrostatic equilibrium equation is simply as follows , @xmath60 in our calculation , the eos of accreted gas including protons , neutrons , and electron pairs are determined by the exact fermi - dirac integral , the eos of neutrinos pairs are determined by the numerical integration of their distribution functions @xmath61_{\nu,\bar\nu}$ ] , and the eos of photons is simply according to eq.(15 ) and we do not include other kinds of particles in this work , especially , helium , which was included in the most of the previous works ( we will check the validity of neglecting the contribution of helium in
iv ) : @xmath62 where @xmath24 and @xmath63 is the total pressure and total internal energy density respectively , @xmath64 is the electron fraction @xmath65 , and @xmath66 is the local temperature of the disk . specifically , the eos of gas are expressed as @xcite , @xmath67,\\ u_i=\frac{2\sqrt 2}{3\pi^2}\frac{(m_ic^2)^4}{(\hbar c)^3}\beta_i^{5/2}\left[f_{3/2}(\eta_i,\beta_i)+\beta_i f_{5/2}(\eta_i,\beta_i ) \right],\\
n_i=\frac{\sqrt 2}{\pi^2}\left(\frac{m_ic^2}{\hbar c}\right)^3\beta_i^{3/2}\left[f_{1/2}(\eta_i,\beta_i)+\beta_i f_{3/2}(\eta_i,\beta_i ) \right],\end{aligned}\ ] ] where @xmath68 is the pressure , internal energy density and number density of particle @xmath69 respectively ( @xmath70 ) , @xmath71 is the fermi - dirac integral of order @xmath72 , @xmath73 is the degeneracy parameter of particle @xmath69 ( @xmath74 , where @xmath75 is the chemical potential of particle @xmath69 not including the rest mass ) and @xmath76 is the relativity parameter of particle @xmath69 ( @xmath77 ) . and the eos of neutrino pairs and radiation are expressed as @xmath78 where @xmath24 is the energy of neutrinos and antineutrinos and @xmath79 is the pressure of particle @xmath80 ( @xmath81 ) .
it is easy to write down the equation of lepton number conservation of fluid using the eulerian description : @xmath82 where @xmath83 is the number density of baryon , @xmath84 is the fraction of lepton number @xmath85 and @xmath86 is the lepton number flux in the vertical direction of the disk ( here @xmath86 is contributed by neutrinos and antineutrinos @xmath87 ) @xmath88 due to the fact that the time scale of accretion is much longer than that of chemical evolution under the condition of the inner part of ndafs : @xmath89 g/@xmath90 , or @xmath91 @xmath92 , and @xmath93k , the time scale for the typical urca process @xmath94 is about @xmath95 ms ( see the fig.2 of @xcite ) which is much shorter than the time scale of accretion , so it is reasonable to neglect the advection term in the equation of lepton number conservation , hence eq.(18 ) is simplified to be @xmath96 the energy equation of fluid is written as @xmath97 or in the equivalent form @xmath98 where @xmath99 is the advection cooling term and @xmath100 is the alpha - viscosity heating rate @xmath101 and @xmath102 is neutrino cooling rate contributed by the energy flux of neutrinos and antineutrinos : @xmath103 @xmath104 we assume the velocity distribution to be @xmath105 , @xmath106 , and @xmath107 , which is similar to the self similar radial distribution of the mass density of the gas pressure dominated thin disk , we also assume the distribution of mass density and energy density to be self similar @xmath108 and @xmath109 in the radial direction . according to the mass conservation equation of steady flows @xmath110 , we obtain @xmath111 , so the advection term @xmath112 is simplified to be @xmath113 and we will check the validity of the self similar assumption in iv . for simplicity , we adopt the value of acoustic speed at @xmath114 when calculating viscosity heating rate and radial velocity at any location , i.e. more numerically economic expressions @xmath115 $ ] for the viscous stress and @xmath116/r\omega$ ] for the radial velocity , and @xmath117 is the square of adiabatic sound speed , which is given by @xcite @xmath118_{\rm ad}=\frac{1}{\rho+p+u}\left[\rho\left(\frac{\partial{p}}{\partial{\rho}}\right)_u+(p+u)\left(\frac{\partial{p}}{\partial{u}}\right)_{\rho}\right].\ ] ]
the two - stream approximation is a simplification to the full boltzmann equation that replaces the full direction dependent distribution @xmath21 and @xmath22 by two streams with angle @xmath119 to the vertical direction . under the two stream approximation ,
the full boltzmann equation is simplified to be @xmath120+\frac{1}{2}\lambda_s[f_-(z , p)-f_+(z , p ) ] , \\
\frac{1}{\sqrt3}\frac{\partial f_-(z , p)}{\partial z}=-\lambda_a[f^{\rm{eq}}(t(z),\mu_{\rm{eq}},p)-f_-(z , p)]+\frac{1}{2}\lambda_s[f_-(z , p)-f_+(z , p)].\end{aligned}\ ] ] two stream approximation has been confirmed to be valid by @xcite and is much easier to deal with compared with the full boltzmann equation , so it is a prime choice for the calculation of the energy density @xmath121 , the cooling rate @xmath122 and the lepton number flux @xmath123 of neutrinos and antineutrinos : @xmath124 now , we start to seek for a steady solution to the disk in which hydrostatic equilibrium , thermal balance and chemical balance are satisfied . first , we assume an initial distribution of temperature @xmath125 mev and election fraction @xmath126 ( in fact , our numerical calculation has shown that the final convergent solution does not depend on the specific choice of the initial conditions which only affects the speed of convergence of numerical calculation .
such independence proves the validity of our numerical methods in turn ) , then we solve the vertical hydrostatic eq.(8 ) subjected the boundary condition ( 7 ) , and it is not hard to obtain the zero order mass density distribution @xmath127 and corresponding the thickness of the disk @xmath128 . then , we solve the two stream approximation eq.(29 ) and ( 30 ) , and obtain the neutrino / antineutrino spectra @xmath129 . with the neutrino / antineutrino spectra
, we can solve the chemical evolution eq.(22 ) and thermal evolution eq.(24 ) with the zero order mass density distribution @xmath127 or the baryon number density @xmath130 fixed , until a steady state @xmath131 and @xmath132 . with the first order distribution @xmath133 and @xmath134 , we solve the vertical hydrostatic eq.(8 ) subject to the boundary condition ( 7 ) once more to get the corresponding first order mass density distribution @xmath135 . iterating this process until the final overall convergent solution of mass density @xmath136 , electron fraction @xmath137 and temperature @xmath138 is obtained .
after that , we switch to the full boltzmann equation eq.(1 ) and ( 2 ) to solve the full energy dependent and direction dependent distribution function of neutrinos and antineutrinos : @xmath61_{,\nu}$ ] and @xmath61_{,\bar\nu}$ ] . with
the full energy dependent and direction dependent spectra of neutrinos and antineutrinos , we can calculate their annihilation luminosity precisely .
the annihilation rate of neutrino pairs is expressed as follows : @xmath139.\end{aligned}\ ] ] where the typical cross section of neutrino interaction is @xmath140 @xmath141 , the weak interaction constants are @xmath142 , @xmath143 , @xmath24 and @xmath144 is the energy of neutrinos and antineutrinos respectively , @xmath54 and @xmath145 is the solid angle of the incident direction of neutrinos and antineutrinos respectively , @xmath146 is the angle between neutrino beams and antineutrino beams ( see fig.1 ) .
in this section , taking the case @xmath147 , @xmath3 as an example , we show the numerical results of the vertical structure of the accretion flow , its radial structure , the spectral energy distrition of the neutrino / anti - neutrino from the disk and the final annihilation luminosity of neutrinos and anti - neutrinos .
( solid lines ) and @xmath148 ( dashed lines ) , where @xmath47 is the thickness of the disk ( see fig.6),title="fig:",scaledwidth=50.0% ] ( solid lines ) and @xmath148 ( dashed lines ) , where @xmath47 is the thickness of the disk ( see fig.6),title="fig:",scaledwidth=50.0% ] ( solid lines ) and @xmath148 ( dashed lines ) , where @xmath47 is the thickness of the disk ( see fig.6),title="fig:",scaledwidth=50.0% ] ( solid lines ) and @xmath148 ( dashed lines ) , where @xmath47 is the thickness of the disk ( see fig.6),title="fig:",scaledwidth=50.0% ] figure 2 shows the distribution of the mass density @xmath51 , temperature @xmath66 , electron fraction @xmath64 and the ratio of internal energy density to pressure @xmath149 in the vertical direction at two different radius @xmath150 and @xmath151 , respectively .
the mass density is about @xmath152 g/@xmath90 and the temperature is about @xmath153 mev in the inner part ( @xmath154 ) of the disk .
the precise value of the temperature in the inner part of ndafs is vitally important , which sensitively determines the annihilation luminosity of neutrino pairs @xmath155 according to eq.(34 ) .
when the disk is in chemical equilibrium , the electron fraction can not be describe by only a constant value throughout the disk , while it varies a few times from the bottom to the surface of the disk .
it is the specific distribution of the electron fraction that guarantees the chemical equilibrium , which was not included in the most previous works , as a result , the electron fraction has to be an artificial assumption there .
it is noticeable that the vertical distribution of @xmath149 and electron fraction @xmath156 are positively correlated ( fig.2c and 2d ) .
the positive correlation implies that relativistic electrons dominate the pressure at the surface of the disk where @xmath157 , and non - relativistic protons and neutrons dominated the pressure at the bottom of the disk where @xmath158 . to justify the conclusion , we plot the vertical distribution of the ratio @xmath159 at radius @xmath150 and @xmath151 in the fig.3 , where @xmath160 , @xmath161 and @xmath162 .
it is indeed so , baryons ( neutrons and protons ) dominate the pressure at the bottom of disk and leptons ( electron pairs ) dominated the pressure at the surface of the disk .
so it is not reasonable to assume a polytropic equation of state @xmath163 @xcite which is the eos of relativistic and strongly degenerate electron gas whose ratio of internal energy density to pressure is @xmath164 . if a polytropic equation of state is needed to do some approximation and estimation , actually @xmath165 is a better choice . of different components : @xmath160 , @xmath161 and @xmath162 in the vertical direction at different radius : @xmath154 ( left panel ) and @xmath148 ( right panel).,title="fig:",scaledwidth=50.0% ] of different components : @xmath160 , @xmath161 and @xmath162 in the vertical direction at different radius : @xmath154 ( left panel ) and @xmath148 ( right panel).,title="fig:",scaledwidth=50.0% ] two main assumptions are used in the above discussion :
we apply the self similar assumption of the distribution of mass density @xmath51 and internal energy density @xmath63 in the radial direction when calculating the advection cooling term and we neglect the contribution of helium to the eos and the contribution of helium disintegration to the cooling term .
we now check their validity .
fig.4 shows the radial distribution of mass density @xmath51 , internal energy density @xmath63 , pressure @xmath24 and temperature @xmath66 on the equator plane @xmath114 and their corresponding fitting lines : @xmath166,@xmath167 and @xmath168 . according to fig.4
, the self similar assumption of mass density @xmath51 and internal energy density @xmath63 in the radial direction is rather a self consistent and accurate description .
now we give a more physical explanation about the self - similar behavior in the radial direction : according to fig.2 and fig.3 , it is evident that non - relativistic protons and neutrons dominate total pressure at the bottom of the disk and determine the surface density of the disk .
hence , we use the polytropic equation of state @xmath169 to estimate vertical structure of the disk . combining eq.(7 ) and
( 8) , it is easy to get the scale thickness of the disk @xmath170 , and the mass density @xmath108 . in order to estimate the radial behavior of temperature and pressure , we must consider the more general eos of non - relativistic gas @xmath171 , and the thermal balance between viscosity heating and neutrino cooling @xmath172 or @xmath173 , so @xmath174 , consequently @xmath168 and @xmath175 ( see fig4 ) . in order to gain an insight to the nature of the disk we investigate
, we also calculate the advection factor @xmath176 where @xmath112 and @xmath100 are the advection cooling rate and heating rate defined in the ii , and the result is shown in fig.5a .
so it is evident that the disk is indeed a neutrino cooling dominated accretion disk as its name suggests : neutrino radiation dominates the cooling process and advection is always a minor role .
the self similar assumption is only applied in the calculation of the advection term , so even if there is some small deviation between the realistic distribution and the self - similar assumption as shown in fig.4 , it has no much influence on the final results .
thus the self similar assumption is rather a reasonable simplification .
the number density of helium @xmath177 is expressed as @xmath178 where the fraction of free nucleons @xmath179 ( protons and neutrons ) is given by @xcite @xmath180 where @xmath181 is the mass density in unit of @xmath182 g/@xmath90 and @xmath183 is the temperature in unit of @xmath184 k and if @xmath185 , then @xmath186 .
we plot the free nucleons fraction of the gas on the equator plane @xmath187 calculated from eq.(37 ) versus radius @xmath188 in fig.5b . according to fig.5b , it is easy to know that @xmath189 , i.e. , the assumption @xmath186 and @xmath190 perfectly hold here .
so it is sound to neglect the contribution of helium to eos and the contribution of the helium disintegration to disk cooling .
it should be noted that the radius where helium dissociation becomes important depends on the viscous parameter @xmath17 @xcite .
so in the case of smaller @xmath17 , the contribution of helium may not be negligible .
g@xmath191@xmath192 , internal energy density @xmath193g@xmath191@xmath192 and pressure @xmath194g@xmath191@xmath192 versus @xmath188 .
the straight lines are the corresponding linear fitting lines : @xmath166 ( solid line ) , @xmath109(dashed line ) and @xmath195 ( dot - dashed line ) .
_ right panel _ : the radial behavior of the temperature @xmath196 versus @xmath197 and the solid straight line is the corresponding fitting line @xmath168.,title="fig:",scaledwidth=50.0%]g@xmath191@xmath192 , internal energy density @xmath193g@xmath191@xmath192 and pressure @xmath194g@xmath191@xmath192 versus @xmath188 . the straight lines are the corresponding linear fitting lines : @xmath166 ( solid line ) , @xmath109(dashed line ) and @xmath195 ( dot - dashed line ) . _ right panel _ : the radial behavior of the temperature @xmath196 versus @xmath197 and the solid straight line is the corresponding fitting line @xmath168.,title="fig:",scaledwidth=50.0% ] ; ( b)_right panel _ : the radial distribution of free nucleons fraction @xmath187 calculated from eq.(37).,title="fig:",scaledwidth=50.0% ] ; ( b)_right panel _ : the radial distribution of free nucleons fraction @xmath187 calculated from eq.(37).,title="fig:",scaledwidth=50.0% ] in addition , we plot the radial distribution of the surface density @xmath198 in fig.6a and the profile of the disk thickness @xmath199 in fig.6b . here , the thickness @xmath199 is defined according to the mass density contrast @xmath200 .
@xmath199 serves as the upper boundary of boltzman equation ( see 2.1 ) , and it is different from definition of the usual scale thickness of the disk which is mostly used in the one - dimensional disk model . according to fig.6 , the disk we consider is a geometrically thick disk with thickness @xmath201 , which will contribute to higher annihilation efficiency as we will discuss in the next section . in unit of @xmath202 g/@xmath141 ; ( b)_right panel _ :
the profile of the disk thickness @xmath199 versus radius @xmath55.,title="fig:",scaledwidth=50.0% ] in unit of @xmath202 g/@xmath141 ; ( b)_right panel _ : the profile of the disk thickness @xmath199 versus radius @xmath55.,title="fig:",scaledwidth=50.0% ] ( left panel ) and @xmath148 ( right panel).,title="fig:",scaledwidth=50.0% ] ( left panel ) and @xmath148 ( right panel).,title="fig:",scaledwidth=50.0% ] when the disk is in chemical equilibrium , the flux of lepton number @xmath203 vanishes , i.e. : @xmath204 figure.7 shows the direction averaged neutrino / antineutrino spectrum @xmath205 at the surface of the disk at different radius , where @xmath206 , @xmath207 mev for @xmath208 and @xmath209 mev for @xmath210 ( see fig.2b ) .
then , at the surface of the disk the chemical equilibrium condition eq.(38 ) is transformed to be @xmath211
so , chemical equilibrium only requires that the total amount of neutrino number flux and antineutrino number flux are equal , i.e. integration of @xmath212 of neutrinos and neutrinos over energy @xmath24 are equal , while it is interesting that the form @xmath212 of neutrinos and antineutrinos almost coincide with each other in fact .
in addition , it is necessary to explain that there is a cusp on the antineutrino spectrum in the lower energy end : it is easy to know that there is a lower energy limit @xmath213 for antineutrinos from the urca process @xmath37 which gives rise to the cusp on the energy spectrum , while there is no such an energy limit for neutrinos from the process @xmath214 , so the neutrino energy spectrum is smooth as expected . in order to calculate the neutrino / antineutrino luminosity and their corresponding luminosity
, we have to make a discount on the vertical structure of the disk : we assume the disk to be lying on the equator plane , and then modify the resulting annihilation luminosity by taking the thickness of the disk into consideration . with the thin disk simplification ,
all the elements in the eq.(34 ) for the annihilation rate are available ( see fig.1 ) , so it is easy to numerically do the integration of eq.(34 ) over the entire surface of the disk and the whole energy span of neutrinos and antineutrinos .
in addition , the neutrino / antineutrino energy flux @xmath215 , luminosity at the surface of the disk @xmath216 and their annihilation luminosity @xmath217 are expressed as @xmath218 where @xmath19 , @xmath20 in our calculation and @xmath199 is the thickness of the disk at radius @xmath55 ( see fig.6b ) .
the results are as follows : @xmath219 ergs / s , @xmath220 ergs / s , the annihilation efficiency @xmath221 .
while it is not the final result , we must take into consideration of the thickness of disk @xmath201 ( see fig.6b ) to correct the thin disk simplification .
fig.8 shows the concrete distribution of the annihilation rate @xmath222 calculated with the thin disk simplification in the @xmath223 plane .
the distribution of the annihilation rate is nearly isotropic aside the directions occupied by the disk and the half open angle of empty funnel along the central axis of the disk is about @xmath224 degrees which is determined by the ratio of thickness to radius of the disk @xmath225 .
the major contribution to the total annihilation luminosity is concentrated at the zone near the central black hole of the disk .
the solid angle that the assumed thin disk lying on the equator plane subtends to the zone is about @xmath226 , while solid angle that the realistic thick disk subtends to that zone is about @xmath227 , and so final annihilation luminosity should be multiplied by a modification factor @xmath228 .
hence , the final results are as follows : neutrino / antineutrino luminosity is about @xmath229 ergs / s , annihilation luminosity is about @xmath230 ergs / s and annihilation efficiency is about @xmath231 . /(ergs@xmath191
@xmath232s@xmath233 ) in @xmath223 plane , and the blank zones are occupied by the disk.,scaledwidth=50.0% ]
temperature of the inner part of ndafs sensitively determines the final luminosity of neutrino annihilation as @xmath234 . in the extreme model of this paper ( @xmath235 , @xmath236 and @xmath3 )
, it is found that temperature in the inner part of the disk is about @xmath237 mev , and the corresponding annihilation luminosity is about @xmath238 ergs / s which roughly satisfies the energy demand of the most energetic grbs .
if the temperature of disks changes from 8 mev to 4mev , the annihilation luminosity reduces about 500 times , which is not enough to power the fireball of the most energetic grbs with the isotropic luminosity about @xmath239 ergs / s . in the standard model of grbs ,
the duration of the energy injection is the duration of the grb prompt emission , which could be about 2s for short grbs and about 100 - 1000s for long bursts .
assuming an extreme constant accretion rate of 10 msun / s means therefore an enormous total mass consumption for powering grbs , especially for powering long bursts , which might be realistic .
therefore , the simple ndaf model which is investigated in this work can not produce sufficient energy to power grbs , the effects of the spin of black hole or / and the magnetic field in the accretion flow might be introduced to power the central engine of grbs .
it is very easy to understand that the temperature of the inner disk mainly depends on the accretion rate of the flow . besides the extreme model
, we also investigate the models with the lower accretion rate , such as @xmath240/s and @xmath241/s , and the results including the resulting temperature @xmath66 in the inner part of disk @xmath242 , neutrino and antineutrino luminosity @xmath243 , annihilation luminosity @xmath217 and the corresponding annihilation efficiency @xmath244 are listed in table 1 .
for all the three different accretion rate @xmath245/s , the thickness @xmath199 we defined in 4.2.3 is roughly equal to radius @xmath55 .
it is obvious that , accretion rate sensitively determines the annihilation luminosity of neutrino pairs and the annihilation luminosity will not exceed @xmath4 ergs / s when the corresponding accretion rate is lower than @xmath246/s .
so ndafs with accretion rate lower than @xmath246/s are unlikely to serve as central engines of grbs .
neutrino and antineutrino spectra is the second major factor determining the annihilation luminosity .
the resulting neutrinos and antineutrino spectra obtained by solving the boltzmann equation shows that , when the disk is in chemical equilibrium , the emission of neutrinos and antineutrinos are almost symmetric with nearly identical energy spectra , but the spectra is neither in the form of black body nor in the form of the gray body . and as shown in @xcite , the black body spectra of neutrino and the neutrino spectra based on the most commonly used simplified model of neutrino transport @xcite or equivalently the gray body spectra @xcite can overestimate the annihilation luminosity by nearly one order of magnitude . in the following , we will also check the validity of the previous assumption on the neutrino transport .
it is easy to estimate the mean optical depth of neutrinos in the inner part of the disk . from fig .
6a and fig .
2b , at @xmath154 , the surface density is @xmath247 g/@xmath141 and the temperature is @xmath207mev .
according to @xcite , the neutrino opacity of absorption @xmath248 and scattering @xmath249 are expressed as @xmath250 @xmath251 where @xmath183 is the temperature in unit of @xmath184k and @xmath252 is the surface density in unit of @xmath202 g/@xmath141 .
therefore , the optical depth of neutrinos we get is @xmath253 , so the disk is optically thick for neutrinos at @xmath208 .
however , as @xcite has shown in quasi - optically opaque case ( @xmath254 ) , the neutrino spectra are neither the black body spectra @xmath255 @xcite nor the gray body spectra @xmath256 @xcite : @xmath257 and the block factor @xmath258 is given by @xmath259 where @xmath260 , so here @xmath261 . for comparison , we plot the direction - averaged spectra @xmath212 of the black body spectra , the gray body spectra and the more realistic spectra obtained by solving boltzmann equation ( see fig.9 ) . according to fig.9 , it is obvious that the first assumption on neutrino transport overestimates the neutrino luminosity about @xmath262 and the corresponding annihilation luminosity about @xmath263 , while the second assumption underestimates the neutrino luminosity about 5 times and the corresponding annihilation luminosity about 25 times .
it may explain why @xcite drew a conclusion that the annihilation luminosity of ndafs is no more than @xmath4 ergs / s , so ndafs can not serve as the central engine of grbs , on the contrary , @xcite claimed an opposite conclusion .
the thickness of the disk is the third factor that affects the final annihilation luminosity : a larger ratio of thickness to radius means a larger solid angle the disk subtends to the annihilation zone . in the case
we consider , the thick disk ( @xmath201 ) will enhance the annihilation luminosity about 3 times than that of a thin disk .
the significance of the distribution of electron fraction in ndafs has been discussed by many previous works @xcite . in this work ,
the distribution of electron fraction is a natural result of chemical equilibrium , thermal balance and hydrostatic equilibrium , rather than an artificial assumption in most previous works , so which should be most reliable . as shown in fig .
5 , helium is completely absent within @xmath264 of ndafs , so it is not needed to consider the contribution of helium to the total pressure and internal energy or the contribution of helium disintegration to the cooling of the disk .
it should be emphasized that our calculation also has its own limitations .
first , in this work , the disk is thick and has two dimensional structure , while the neutrino transport is treated in one dimension .
second , the effects of the motion of accretion flow and the curved spacetime on the vertical transport of neutrinos are neglected .
these effects are significant in the inner part of the disk , for example , at @xmath208 , the special relativity correction to the neutrino energy can be @xmath265 , and the general relativity correction to the energy can be @xmath266 .
third , in the zone near to the event horizon of the central black hole where the annihilation concentrates , the trajectories of neutrinos are severely bent by the central black hole and a considerable amount of neutrinos will undoubtedly be captured by the hole .
in addition , the gravitational instability of the disk was proposed by @xcite to explain the energetic x - ray flares after the prompt emission in grbs .
the outer region of ndafs with extremely high accretion rate is gravitational unstable @xcite , will fragment and cause a variable accretion rate in the inner region .
we would like to thank the annonymous referee for her / his constructive suggestions and comments which are helpful for the improvement of this paper .
this work is partially supported by national basic research program of china ( 2009cb824800 , 2012cb821800 ) , the national natural science foundation ( 11073020 , 11133005 ) , and the fundamental research funds for the central universities ( wk2030220004 ) . | we investigate the vertical structure of neutrino dominated accretion disks by self - consistently considering the detailed microphysics , such as the neutrino transport , vertical hydrostatic equilibrium , the conservation of lepton number , as well as the balance between neutrino cooling , advection cooling and viscosity heating . after obtaining the emitting spectra of neutrinos and antineutrinos by solving the one dimensional boltzmann equation of neutrino and antineutrino transport in the disk
, we calculate the neutrino / antineutrino luminosity and their annihilation luminosity .
we find that the total neutrino and antineutrino luminosity is about @xmath0 ergs / s and their annihilation luminosity is about @xmath1 ergs / s with an extreme accretion rate @xmath2/s and an alpha viscosity @xmath3 .
in addition , we find that the annihilation luminosity is sensitive to the accretion rate and will not exceed @xmath4 ergs / s which is not sufficient to power the most fireball of grbs , if the accretion rate is lower than @xmath5/s .
therefore , the effects of the spin of black hole or / and the magnetic field in the accretion flow might be introduced to power the central engine of grbs . |
understanding the origin of lenticular galaxies and how closely related they are to spiral galaxies is still a challenge for contemporary astrophysics . a wealth of evolutionary evidence links s0 galaxies to spirals .
the morphology - density relation ( @xcite ) shows that while the number of spiral galaxies decreases towards the centre of a cluster , the number of s0 galaxies increases .
it has also been shown that s0 galaxies are more common at the present epoch than at higher redshifts , while the inverse is true for spiral galaxies ( @xcite ) .
the mechanism responsible for transforming a spiral into an s0 galaxy has to stop star formation in the disk and enhance the spheroidal component ( @xcite ) .
interestingly , s0 galaxies are found in all environments , from high density clusters to the field , allowing for a variety of evolutionary paths ( gas stripping , strangulation , harassment , minor mergers , pestering , secular evolution ) and raising the question of whether s0 galaxies are a unique class , or a collection of objects whose formation mechanisms are environment dependent ( @xcite ) .
recent work , moreover , suggests the possibility of creating s0s through major mergers under specific initial conditions ( @xcite ) .
the past history of a galaxy is imprinted in its faint outer regions , where we can find signs of minor mergers , traces of accretion , or of a gentle passive fading . in general , it is difficult to study the kinematics and the stellar populations of early - type galaxies at large radii due to the absence of an undisturbed hi disk and the low surface brightness of the stellar component at large radii .
recently , this last issue has been solved by using discrete kinematic tracers , which are detectable at large radii .
large datasets are becoming available from globular clusters ( @xcite , @xcite ) , planetary nebulae ( pne ) ( see e.g. @xcite , @xcite ) and integrated stellar light ( see eg .
@xcite , @xcite , @xcite ) .
moreover , gcs provide us with a unique tool for understanding the chemodynamical properties of their host galaxies .
they tend to be old ( @xmath110gyrs ) and are the closest examples we have to simple stellar populations ( @xcite ) .
gcs are ubiquitous across a range of galaxy types and luminosities , and can be studied up to distances of @xmath2100mpc ( see @xcite ) and beyond ( @xcite ) .
they form a bimodal distribution in optical - colours that is often interpreted as evidence for metallicity bimodality ( @xcite ; for an alternative view see @xcite , @xcite )
. in general , blue gcs tend to reside in the galaxy halo , while red globular clusters are associated with the metal rich stellar population , i.e. the overall stellar population in elliptical galaxies ( @xcite ) and the bulge or the thick disk of spiral galaxies ( @xcite , @xcite ) .
studying the properties of the gc sub - populations in lenticular galaxies can help constrain s0 formation histories .
if s0 galaxies are gently quenched spiral galaxies , we expect their gc sub - populations to have the same properties as in spiral galaxies , only their specific frequency would be higher , since the total number of gcs would be the same in s0s and spirals , while spiral galaxies would be brighter ( @xcite ) .
in contrast , violent events , such as mergers or the accretion of small companions , would alter the initial distribution of gc in the galaxies ( @xcite ) .
@xcite studied the red gc system in the lenticular galaxy ngc2768 , finding that the red gcs share the same properties as pne belonging to the spheroidal component of the galaxy .
this suggests that , in this s0 , the red gcs behave as in spiral galaxies .
however this work is based on only a colour selection of the gc sub - populations . in an attempt to study the gc sub - populations in s0 galaxies to investigate their formation histories , we present a new method which uses the galaxy photometry and kinematics , as recovered from pne and integrated stellar light , to obtain the likelihood for each gc of belonging to the spheroid or the disk of a galaxy .
this method ensures a physically meaningful separation , based on the kinematics rather than the colour , and allows us to trace streams and substructures .
subsequently , we construct a chromo - dynamical model of the galaxy , where we correlate the gc probability of belonging to one of the galaxy components with its colour . our final goal is to understand the build up of the blue and the red gc sub - populations in light of the hierarchical merging framework of galaxy formation and shed light on the possibility of using gcs as probes of galaxy formation .
in particular in this paper we study the lenticular galaxy ngc1023 .
a number of studies have been carried out on this galaxy ( @xcite , @xcite , @xcite , @xcite ; and many others ) and it is an ideal case for testing this method . a new catalog of gc positions and velocities has been obtained for this work by the sluggs team ( see below ) and is published in this paper . the paper is structured as follows . in section 2
we present the data we employ for our analysis .
the method is presented in section 3 and the results are presented in section 4 . in section 5
we discuss our findings in the context of galaxy evolution .
summary and conclusion are presented in section 6 .
in this paper we study three different types of kinematics tracers : pne , gcs and integrated stellar light . in particular , we test a new method where we use the galaxy kinematics as recovered using pne and stars to define a likelihood for a given gc to belong to the disk or the spheroid of the galaxy .
all the data have been published previously with the exception of the gcs that are published here . in the present section
we briefly describe the detection techniques and data reduction analysis of the three different datasets .
fig.[fig : figure1 ] shows the spatial distribution of pne ( small , green , open circles ) , gcs ( large , magenta circles ) and integrated stellar light ( filled , orange squares ) over a dss image of ngc1023 , with the hst field - of - view outlined in black .
note the large radial extent that such tracers cover , especially pne and gcs that reach out to @xmath3 disk scale length , @xmath4 , with @xmath5 arcsecs ( @xcite ) .
east of the galaxy it is possible to see an excess of light at the location of the companion galaxy ngc1023a .
we have adopted a distance for the galaxy of 11.1 mpc @xcite .
gc data were acquired as part of the sages legacy unifying globulars and galaxies ( sluggs ) survey @xcite .
gc candidates were identified in hst / acs imaging in @xmath6 and @xmath7 filters , as described in @xcite .
the latter was supplemented with archival cfht data in @xmath6 and @xmath8 from @xcite to target gc candidates outside the hst footprint . for both datasets ,
gc candidates were selected based on their location in the colour - magnitude diagram and , for hst imaging only , based on their partially resolved half - light radius @xmath9 .
we adopt the classification of @xcite , for which gcs are fainter than @xmath10 mag in g - band and ultra compact dwarfs ( ucds ) are brighter than @xmath10 mag .
the radius definitions to divide gcs from faint fuzzies ( ffs , @xcite ) are somewhat arbitrary , in this work we follow the definition of @xcite , for which ffs have @xmath11 pc and gcs have @xmath12 pc .
the total number of gcs recovered using only photometry ( _ photometric gcs _ ) is 360 .
spectroscopic follow - up of a subsample of gc candidates was performed with the multi - object spectrograph keck / deimos .
a total of four deimos masks were observed , see table [ tab : masks ] .
two masks were aligned along the major axis of the galaxy , and two along the minor axis .
priority was given to bright and partially - resolved gcs in hst , whereas remaining slits were assigned to un - resolved objects outside the hst field - of - view .
the reader should see @xcite for a detailed description of the data reduction , while we summarise here the main points of the data analysis .
we measure the redshift of the calcium triplet ( cat ) lines at 8498 , 8542 and 8662 .
this was performed by cross - correlating eleven template spectra of galactic stars spanning a wide range of metallicities , with the actual gc spectrum .
figure [ fig : figure1 ] shows that most of the confirmed gcs turned out to be inside the hst field - of - view .
this is because the gc system of ngc 1023 is @xmath13 arcmin in diameter and it is elongated along the photometric major axis @xcite as confirmed by the paucity of gcs along the galaxy minor axis .
c c c c c c mask & date & exposure time + @xmath14 & @xmath15 & @xmath16 + @xmath17 & @xmath15 & @xmath18 + @xmath19 & @xmath20 & @xmath21 + @xmath22 & @xmath23 & @xmath24 + @xmath19 & @xmath25 & @xmath26 + we adopt the hst catalog as our primary photometric catalog .
we convert the cfht @xmath27 colours into hst @xmath28 colours by matching the objects in common between the two datasets .
we find @xmath29 and we apply this conversion to the objects with ground - based photometry only .
the @xmath28 gc colour distribution is bimodal , see figure [ fig : figure1tris ] , with a colour separation between blue and red gcs set at @xmath30 mag @xcite .
the spectra for 115 candidates belonging to the ngc 1023 systems have been obtained and the velocities have been confirmed by two independent observers . for 96 of the 115 detections , which fall in the fields observed with hst
, we have measured the effective radius , @xmath9 .
following @xcite , we identify as ffs objects with @xmath31 pc and as ucds objects with @xmath32 mag , see figure [ fig : figure1bis ] . in figure
[ fig : figure1bis ] , we also show the location of the ffs observed in this galaxy by @xcite , which will be used in the following sections to complete our ffs catalog .
the remaining 19 objects with cfht photometry are also classified as gcs , since they have a magnitude @xmath33 mag in g band and lie in the outskirts of the galaxy , while ffs are generally found close to the galaxy centre . in total
we have 104 gcs , 63 blue and 41 red ( _ spectroscopic gcs _ ) .
following @xcite we have assigned 6 gcs to the companion galaxy using the luminosity ratio between the two galaxies and requiring that the radial velocity of the object has a maximum difference with respect to the systemic velocity of ngc1023a of @xmath34 km / s .
the companion galaxy gcs are 3 red and 3 blue .
the integrated stellar light individual data points were presented in @xcite and the integrated stellar light data are taken from the kinemetry fits presented in foster et al .
( 2015 , submitted ) , they were extracted using the skims technique ( @xcite ) .
the method used to extract integrated kinematic information of the underlying galaxy starlight from a multi - slit spectrograph was first developed by @xcite and @xcite .
we use the individual data points to assign a probability of belonging to the companion galaxy , following the same procedure as for pne and gcs .
we find that 90 % of the observed stellar light is associated with the main galaxy .
we use the binned kinemetry fits to compare the stellar kinematics with the kinematics of pne and gcs , see figure [ fig : figure3 ] .
the pne catalog was obtained using the planetary nebulae spectrograph ( pns .
it was published in @xcite and information on the data reduction and completeness correction can be found in @xcite : where 203 pne were detected , 20 of which are associated with the companion galaxy n1023a .
the pne trace the galaxy light profile well , once corrected for incompleteness in the centre . when fitting with a tilted disk model
the pne present a strange velocity profile , with decreasing rotation velocity and increasing values of the tangential components of the random motions ( @xcite ) .
this was explained as a contamination of the disk kinematics by spheroid pne . when this effect is properly accounted for , by assigning a probability to each pn of belonging to the spheroid or the disk ,
the derived kinematic presents a spiral - like behaviour , with a flat rotation curve and decreasing dispersion velocity in the disk .
the colours of gcs are believed to be connected to the galaxy formation history .
one scenario ( @xcite ) has blue , metal - poor gcs formed during the collapse of the protogalactic cloud and associated to the galaxy halo .
red , metal - rich , gcs on the other hand might have been formed during a starburst that created the galaxy spheroid .
alternatively , @xcite suggest that red gcs might have formed in super star - forming clumps at @xmath0 during the gas rich phase of galaxy evolution ( see also @xcite ) .
these clumps would have migrated toward the centre of the host galaxy , transporting gcs toward the proto - bulge .
some of the gcs would have been stripped from the migrating clumps and become part of the newly born thick disk . both those sub - populations would be old ( @xmath35 gyrs ) . in this simple scenario
we expect to find a correlation between colour and kinematics of gcs : i.e. red gcs should share the same kinematics as spheroid or thick disk stars ( @xcite ) and blue gcs would trace the halo kinematics . traditionally , this analysis has been performed dividing gcs according to their colour and then studying their kinematics ( @xcite ) . in section [ sec : analysis1 ] we reproduce such an analysis . in section [ sec : model ]
we describe the galaxy model obtained from pne and integrated stellar light . in section [ sec : analysis2 ] we present a new method that , instead , uses the gc kinematics to divide them into different sub - populations and then study their properties .
this method assumes that the galaxy halo is an extension of the galaxy bulge .
we call the bulge + halo components the galaxy spheroid , i.e. the spherically symmetric component dominated by random motion .
this approximation could be relaxed when more data for the galaxy outskirts become available .
gcs follow the total light profile very well , the disk dominates the galaxy light .
there is no visible difference between the two gc sub - populations .
gcs might have been destroyed toward the galaxy centre or not been observed due to the bright galaxy background , therefore their distribution might be incomplete in the centre . ] as a first analysis , we compare the one - dimensional radial distribution of gcs with the galaxy starlight : if gcs follow the underlying stellar population , then , to within a normalisation factor , the two might be expected to show the same profile .
figure [ fig : figure2 ] shows the comparison of the total , blue and red _ photometric gc _ number densities ( magenta , blue and red circles respectively ) against the total , spheroid , and disk surface brightness ( upper panel , middle panel , low panel , respectively ) , as modelled running galfit on an r - band image of the galaxy ( @xcite ) .
the recovered parameters are : disk total absolute magnitude @xmath36 mag , disk scale length @xmath37 arcsec , disk axis ratio @xmath38 and disk position angle @xmath39 degrees ; for the spheroid we estimated the total absolute magnitude to be @xmath40 mag , effective radius @xmath41 arcsec , sersic index @xmath42 , axis ratio @xmath43 , position angle @xmath44 .
the spheroid over total light fraction is @xmath45 .
the gc number density has been obtained by separating the objects belonging to the main galaxy , ngc1023 , from the objects associated to the companion galaxy , ngc1023a , using a light weighted probability ( @xcite ) .
first , we bin the gcs of ngc1023 into elliptical annuli , going outward in radius , using the ellipticity obtained from a non parametric isophotal analysis of the r - band image of the galaxy .
the total number of gcs in each bin is then divided by the bin area . in figure
[ fig : figure2 ] , the gc total , red and blue number densities have been shifted in the y - axis to match the modelled surface brightness profile .
the total model , black continuous line in the upper panel of figure [ fig : figure2 ] , describes well the light profile as obtained with an ellipse fitting ( black diamonds ) of the same r - band image .
the total gc number density follows the total light profile , supporting the idea that they are associated with the galaxy .
this galaxy light profile is dominated by the disk component and the gcs seem to follow the disk light profile . @xcite
found that the distribution of red and blue gcs in this galaxy is elongated along the major axis , resembling the disk isophotes , supporting this result .
nevertheless the spheroid profile would provide a good fit for the outer gcs , suggesting that red and blue gcs belong to both the galaxy components .
a second clue comes from the study of the kinematics of the gc system as a whole , and of the blue and red sub - populations .
figure [ fig : figure3 ] shows the gc two - dimensional smoothed velocity fields , using an adaptive gaussian kernel smoothing ( @xcite ) , for the total , red and blue sample ( left , middle , right panel , respectively ) .
rotation with a maximum velocity of @xmath46 km / s , is clear in the total gc population as a whole , and for the red and the blue sub - populations separately .
a similar result was found by @xcite , studying the gc system of the spiral galaxy ngc 2683 . the milky way red gcs also show large systemic rotation and were first associated with the old thick disk ( @xcite ) and then with the galactic bulge , due to the similarity of the net rotation of the innermost red gcs with that of the old field giant stars in the galactic bulge ( @xcite ) . to compare the gc kinematics with those of the other tracers ,
we obtain the rotation velocity by fitting an inclined disk model using a maximum likelihood method ( @xcite ) for pne and gcs , while the skims kinematics are obtained with kinemetry ( foster et al . , 2015 , _
submitted _ ) .
the results are shown in figure [ fig : figure4 ] .
all the tracers ( even when dividing gcs into the blue and red sub - populations ) have the unusual kinematic profile found by @xcite for pne , characterised by a decreasing rotation velocity and increasing random motions , in the outermost points ( i.e. beyond 140 arcsec ) .
the agreement found in the spatial distribution and the kinematics of the three tracers suggests that we can use the stellar kinematics as recovered by integrated stellar light and pne to model the gc velocity field . in section [ sec : analysis1 ] we showed that , when recovering the kinematics of n1023 using pne , gcs and integrated stellar light as tracers , the rotation velocity decreases at large radii , while the velocity dispersion increases . in a work based only on pne data , @xcite explained this behaviour as a result of contamination of the disk kinematics by pne belonging to the spheroidal component .
@xcite also show that assuming a spheroid + disk kinematic model , the unusual behaviour of the kinematics vanishes and the galaxy presents a spiral - like velocity profile , with a flat rotation curve , decreasing random motion in the disk and a pressure supported spheroidal component . as a first test , we want to investigate whether , when a spheroid + disk model is applied , the skims data also present disk - like kinematics , as found for the pne . in the present section ,
we give a brief summary of the method used to obtain the spheroid and disk kinematics for pne and integrated stellar light data , but we refer to @xcite for a more detailed explanation .
a galaxy r - band image ( @xcite ) is decomposed into its spheroid and disk component using galfit ( @xcite ) .
the pne and skims data are super - imposed on the image obtained by dividing the spheroid model by the total model , called f - map , and a probability @xmath47 ( from 0 to 1 ) is assigned to each tracer ; i.e. objects with @xmath48 would definitely be in the spheroid and object with @xmath49 would definitely be in the disk .
we use a maximum likelihood method to estimate the best kinematics parameters ( i.e. rotation velocity and velocity dispersion in the disk , dispersion velocity in the spheroid ) , assuming a velocity distribution function which is the sum of the spheroid distribution function ( with zero mean rotation velocity ) and the disk distribution function .
the assumption that the mean spheroid rotation is zero will be relaxed below .
both distribution functions are assumed to be gaussian and the contribution of each one of them to the total distribution is given by the parameter @xmath47 .
the fit is carried out in 4 radial bins for the pne and 3 radial bins for the skims data , with the same number of objects in each bin , always higher than 30 in order to assure a reliable fit .
the fit is iterated several times , each time rejecting three - sigma outliers , i.e. objects with a likelihood lower than the _ likelihood clipping probability threshold _ , which is the value at which we cut the velocity distribution function ( @xcite ) .
the fit rapidly converges rendering this process robust against a small amount of contamination . since the skims data are not point sources but slitlets
, we have assumed that every slit is approximately described by a gaussian velocity distribution function and we have extracted 10 random realisations of every slit , in order to have 10 different mock catalogs , and carried out the analysis aforementioned on every catalog independently . in figure
[ fig : figure5 ] , left panel , the orange dots represent the median of the recovered parameters and the error is given by their standard deviation .
the pne kinematics are shown by green dots .
error bars on the @xmath50-axis are 1 @xmath51 error bars for the recovered parameters .
error bars on the @xmath52-axis show the size of the bin where the fit was performed .
the kinematics of stars and pne is clearly similar .
we can therefore unite the two data sets to obtain a catalog of more than 300 tracers and recover the galaxy kinematics with more precision .
moreover , we estimate the rotation of the spheroid , see black circles with continuous error bars in the bottom , right panel of figure [ fig : figure5 ] .
this estimate assumes that @xmath53 , where @xmath54 is the rotation velocity of the galaxy spheroid and @xmath55 the spheroid ellipticity ( see figure 4.14 in @xcite ) .
the galfit modelling shows that the spheroid in ngc1023 has an ellipticity of @xmath56 , which translates into a predicted value of @xmath57 .
we also recover two components of the velocity dispersion @xmath58 , see second row of figure [ fig : figure5 ] , right panel ( we assume @xmath59 ) .
the average ratio between rotation velocity and velocity dispersion in the azimuthal direction , in the two outermost bins , is @xmath60 .
this value is lower than the one expected for a spiral galaxy ( @xcite ) suggesting that this s0 galaxy is not simply a faded spiral .
following @xcite , this value of @xmath61 is consistent with this galaxy being involved in only very minor mergers in its past , see the discussion section for more details .
gcs give us , therefore , a unique opportunity for verifying the various possible formation histories of s0 galaxies , in particular whether s0s are faded spirals or an independent class of galaxies .
in fact , if the spiral galaxy has gone through a series of minor mergers that transformed it into an s0 galaxy , gcs may have been stripped from the accreted dwarf galaxies , which will pollute the chromo - dynamical relation found for spiral galaxies ( @xcite ) .
nevertheless , we would expect to find the majority of red gcs to be associated with the spheroid or thick disk and blue gcs with the galaxy halo , as found in spiral galaxies .
[ cols= " < , > " , ]
having computed the likelihood for every gc of belonging to the model of the galaxy obtained combining photometric information from an r - band galaxy image and pne and skims kinematics , in this section we study the properties of the gc population and sub - populations .
we have defined four different sub - populations : red and blue gcs , on one hand , and disk and spheroid gcs , on the other .
it should be noted that in this preliminary study the spheroid may incorporate bulge and halo sub - populations with different properties . in section [ sec : kinecolor ] we investigate the relation between these four sub - populations and the galaxy as a whole . in section [ sec : rej ] we study all the kinematic tracers unlikely to belong to ngc1023 system and their connection with the companion galaxy , ngc1023a .
to obtain a clearer picture of the correlation between gc colour and kinematics , we have to take into consideration the shape of the bi - modal colour distribution .
in fact , the red and blue sub - populations overlap at intermediate colours , and our artificial colour cut at @xmath30 is an approximation .
we fitted the two sub - populations with a gaussian function and estimated that the area in which the two sub - populations significantly overlap is @xmath62 and we exclude gcs with these colours from our analysis .
as a first step , we calculate the total number of disk and spheroid gcs , simply summing up the probability for each one of them of being in the disk or in the spheroid .
we find that out of the 32 red gcs , 19 are associated with the disk and 13 with the spheroid .
while 18.6 blue gcs are associated with the disk and 25.4 with the spheroid . finally , 15 gcs are found to have a probability lower than 2.1 @xmath51 of belonging to the system .
the majority of the rejected gcs have blue colours .
looking at the trend of @xmath63 vs the ( g - z ) colour of the gcs , we realise that an extra parameter is needed to make this plot meaningful .
figure [ fig : figure7 ] shows the distribution of the gcs ( associated with the system ) in a colour vs galactocentric distance plot , colour coded according to the probability of being in the spheroid @xmath63 .
it is clear that the reddest gcs are concentrated in the centre of the galaxy and that they are associated with the galaxy disk .
this fact is of particular interest since in this galaxy a population of ffs has been found to have rotational velocity similar to the stellar one ( @xcite ) .
ffs are also shown in figure [ fig : figure7 ] as red open circles .
they share the same position as the disk red gcs . for red gcs
it seems that the reddest the gc the higher is its probability of belonging to the disk . for blue gcs ,
instead , this probability seems to correlate with its distance from the galaxy centre , with gc at low radii having a higher probability of belonging to the disk .
the method used in this paper to recover the galaxy kinematics allows us to identify outliers , i.e. objects that are not in equilibrium with ngc1023 disk + spheroid potential ( see @xcite for details ) .
some of the rejected gcs fall in an area of the galaxy where there are no data available to recover the kinematics .
therefore we have extrapolated the galaxy kinematics in these outer regions from the outermost bin , for which we have pne and stellar kinematics .
the outer parts of this galaxy are generally dominated by the spheroid component that we assume to represent the bulge + halo components of the galaxy .
if this assumption is not correct these outliers could be halo gcs .
on the other hand , they could be objects that are in the process of being accreted to the system .
@xcite , in a study of the kinematics of the milky way satellites , shows that it is possible to quantify the time since they have started falling onto the mw using a 1-d phase space diagram .
in such a diagram the infall time of an object is proportional to its distance from the galaxy systemic velocity .
figure [ fig : figure8 ] , left panel , shows the 1-d phase space diagram for ngc1023 .
rejected objects are shown as grey symbols ( gcs ; crosses , pne and stars ; squares and companion galaxy objects ; asterisks ) .
objects included in our model are shown as coloured symbols .
the pne and integrated stellar light individual data points are shown as tiny green squares , gcs as small circles and ffs as large circles .
the galaxy systemic velocity is shown as a dashed line .
pne , integrated stellar light data points , and gcs form a half - diamond structured centred around the galaxy systemic velocity . the companion galaxy pne and gcs and
some of the rejected objects lie at the borders of this half - diamond , as if they are just starting in - falling .
figure [ fig : figure8 ] , right panel , shows the velocity vs distance along the major axis for all the objects , following the same symbols convention as in the left panel .
the pne and stars follow a radial gradient in velocity , typical of rotating system .
gcs and ffs follow the same distribution as pne .
in particular , the gcs belonging to the disk ( open small circle ) lie in the external part of the distribution , i.e. they have a high velocity respect to the galaxy systemic velocity .
the majority of the gcs associated with the galaxy spheroid ( filled small circle ) , instead , are located near the galaxy centre or along the minor axes .
the rejected objects seem to form a coherent structure in the phase - space , connected with the companion galaxy , suggesting they may have been stripped from the companion while orbiting the main galaxy .
three of the ucds are in the same locus as the companion galaxy and one belongs to the stream , suggesting a possible origin of the ucd connected with the interaction of the two galaxies ( @xcite ) .
this result needs confirmation with more data .
figure [ fig : figure8 ] support a scenario where the rejected gcs are part of a stream connecting ngc1023a with the main galaxy .
this possibility can be fully confirmed when data to properly describe the galaxy halo will be available .
in this paper we present a new catalog of spectroscopically confirmed gcs associated with the lenticular galaxy ngc1023 and its companion galaxy ngc1023a .
we find that in this galaxy the gcs follow the total galaxy light profile and are therefore good tracers of the overall stellar population .
their kinematics are also consistent with the stellar kinematics .
we divide the gcs according to their colour , at @xmath30 , and find that the light profile and kinematics of both the red and blue sub - populations also resemble the stellar light profile and kinematics . however , for both red and blue gcs
, it is not clear from the comparison of the number density profiles with the decomposed disk and spheroid profiles whether the gcs better match either the disk or spheroid light profiles .
we , therefore , develop a new method to associate gcs to the galaxy disk and spheroid on the basis of their position and velocities , using pne , integrated stellar light and an r - band image of the galaxy , to model the galaxy kinematics .
we find that the majority of red gcs ( 19 ) have a high likelihood of belonging to the galaxy disk .
the other ( 13 ) red gcs belong to the spheroid .
interestingly ngc1023 has a sub - population of red ffs ( @xcite ) whose kinematics are indistinguishable from the stellar kinematics of the disk ( @xcite ) and whose azimuthal distribution is elongated along the galaxy major axis ( @xcite ) .
the majority of blue gcs are spheroid - like objects ( 25.4 ) , but , surprisingly @xmath64 are found to be disk - like objects .
the ratio between rotation velocity and velocity dispersion in the azimuthal direction in ngc1023 is @xmath60 .
this value is lower than the one expected for spiral galaxies ( @xcite ) suggesting that this s0 galaxy is not simply a faded spiral .
following @xcite , this value of @xmath61 is consistent with this galaxy having been involved in only very minor mergers in its past .
an alternative explanation for our finding that @xmath65 is rather low in the disk of n1023 is provided by the possibility of forming the galaxy disk at high redshift through the merger of super star - forming clumps during the gas rich phase of galaxy evolution .
@xcite , in fact , suggested that red gcs are born within ` turbulent and clumpy ' disks at @xmath66 , i.e. disks forming in a violent way . at @xmath67 , these disks ( when they survive )
may look much hotter than the thin disks in spiral galaxies .
clumpy disk formation , moreover , would leave imprints in the halo , as suggested by simulations from @xcite .
halo objects could interact gravitationally with the giant clumps and acquire rotation in a dynamical friction timescale of 0.5gyr .
the influence of the clumps would be limited to a region around the disk .
this formation scenario explain the finding that red and blue gcs in the inner part of the galaxy have a disk - like kinematics .
nevertheless , the galaxy is going through an interaction ( and maybe a minor merger ) at the present epoch with a companion galaxy , ngc1023a , visible inside the galaxy disk optical isophotes .
this little companion may have been stripped of its gas , producing an extended and irregular hi distribution , displaced from the optical centre ( @xcite ) .
we find that some of the gcs that are unlikely to be in equilibrium with the potential of ngc1023 are part of a stream , probably connected to the companion galaxy .
blue ffs , moreover , are associated with the companion galaxy and the hi clumps ( @xcite ) .
some of the gcs unlikely to belong to the model of the galaxy could be halo gcs that are not well represented by our simple model and are , therefore , rejected during the fitting .
ngc1023 seems to have a population of gcs associated with its disk .
in contrast , in the s0 galaxy ngc2768 red gcs share the same kinematics and density profile as spheroid stars ( @xcite ) , as found for the mw and the sombrero galaxy .
this framework provides an interesting possibility for testing how a given galaxy disk formed : by seeing if there are associated red gcs . as more gc data in spiral galaxies become available , together with the ages of s0 galaxy disk and spheroid , and pne kinematics of galaxy halos
are obtained , the formation scenario for s0 galaxies will become clearer . for the moment , we find that gcs are powerful probes of galaxy evolution and that the red gcs in the two lenticular galaxies ngc1023 and ngc2768 are associated with different galaxy components , suggesting that s0 galaxies form through different pathways ( or that their progenitors formed through different pathways ) .
in this paper we have published a new catalog of gcs and introduced a new method to associate gcs with different galactic components , given a kinematic model .
this method also allows us to identify streams and can be applied to any discrete kinematic tracers .
in particular we have used pne , stellar light and an r - band image to constrain the kinematic model of ngc1023 , decomposing it into its disk and spheroid components . following that we have studied its gc population .
the main findings are : * the gcs follow the stellar density profile and kinematics .
* both red and blue gc sub - populations show rotation . *
when decomposing the galaxy into its spheroid and disk component , we find that the majority of the red gcs show disk - like kinematics , the others are consistent with being in the spheroid .
most of the blue gcs have a high likelihood to belong to the spheroid , but some of them show disk - like kinematics . *
the gcs that are unlikely to belong to the galaxy model appear to be associated with a stellar stream , connected to the companion galaxy ngc 1023a , previously identified by @xcite using pne .
we note that they could also be halo gcs that are not well represented by our simple model and are , therefore , rejected during the fitting .
all these findings are consistent with this galaxy having been involved in at least one minor mergers , while its disk was created at high redshift from super - star - forming clumps . moreover , in the only other galaxy for which similar data are available , ngc2768 , the red gcs share the same kinematics and density profile as spheroid stars .
this suggests that s0 galaxies are an heterogeneous class and that gcs are powerful probes of galaxy evolution .
the newly developed method will be applied on a sample of lenticular galaxies to place stronger constraints in the galaxy - globular clusters assembly history .
the authors wish to thank the anonymous referee for useful comments .
the authors would like to thank dr .
d. kruijssen and dr .
y. jaffe for interesting discussions and suggestions .
this work was supported by nsf grant ast-1211995 .
ac would like to thank fapesp for the fellowship 2013/04582 - 4 .
acs acknowledge receipt of a royal society travel grant and the hospitality at uc santa cruzacs and funding from a cnpq , bjt - a fellowship ( 400857/2014 - 6 ) .
daf thanks the arc for financial support via dp130100388 .
rp acknowledges the support of cnpq through grant 157761/2014 - 2 .
j. p. , romanowsky a. j. , strader j. , forbes d. a. , foster c. , jennings z. g. , pastorello n. , pota v. , usher c. , blom c. , kader j. , roediger j. c. , spitler l. r. , villaume a. , arnold j. a. , kartha s. s. , woodley k. a. , 2014 , , 796 , 52 | globular clusters ( gcs ) can be considered discrete , long - lived , dynamical tracers that retain crucial information about the assembly history of their parent galaxy . in this paper , we present a new catalogue of gc velocities and colours for the lenticular galaxy ngc1023 , we study their kinematics and spatial distribution , in comparison with the underlying stellar kinematics and surface brightness profile , and we test a new method for studying gc properties . specifically , we decompose the galaxy light into its spheroid ( assumed to represent the bulge + halo components ) and disk components and use it to assign to each gc a probability of belonging to one of the two components .
then we model the galaxy kinematics , assuming a disk and spheroidal component , using planetary nebulae ( pne ) and integrated stellar light .
we use this kinematic model and the probability previously obtained from the photometry to recalculate for each gc its likelihood of being associated with the disk , the spheroid , or neither . we find that the reddest gcs are likely to be associated with the disk , as found for faint fuzzies in this same galaxy , suggesting that the disk of this s0 galaxy originated at @xmath0 .
the majority of blue gcs are found likely to be associated with the spheroidal ( hot ) component .
the method also allows us to identify objects that are unlikely to be in equilibrium with the system . in ngc1023 some of the rejected gcs form a substructure in phase space that is connected with ngc1023 companion galaxy .
[ firstpage ] galaxies : elliptical and lenticular , cd , galaxies : individual : ngc 1023 , galaxies : kinematics and dynamics . |
the general term ` pattern formation ' refers to the spontaneous appearance of structure in a dissipative dynamical system in a continuum or spatially discrete medium , when a spatially homogeneous state undergoes an instability of some kind @xcite . in many physical situations
it is appropriate to use mathematical models for which the spatial dynamics are modelled discretely , rather than as a continuum .
such lattice models arise in many physical applications : in particular we mention recent work in nonlinear optics @xcite and mathematical modelling of processes in developmental biology @xcite . in many applications
the existence of domain - filling ( almost ) regular solutions exhibiting spatial periodicity is of interest .
mathematical work on periodic patterns in regular lattices has also been the subject of considerable recent interest @xcite .
in addition , there has been a renewed interest in the study of localized states in nonlinear dissipative systems in recent years , motivated by a wealth of experimental and numerical results both in continuum systems @xcite and spatially discrete models @xcite .
such localized states have been referred to by a number of different names , including ` light bullets ' and ` dissipative solitons ' . in continuum systems one recurring mechanism for the generation of localized states
is the existence of a subcritical turing ( pattern - forming ) instability . in the simplest case ,
two branches of small - amplitude localized states emerge at the turing instability , also in this context referred to as a hamiltonian
hopf bifurcation @xcite .
away from the initial instability they develop into a complicated structure of fully nonlinear localized states which exist on a pair of intertwining curves @xcite ; a complete asymptotic description of the bifurcation structure involves consideration of exponentially - small effects @xcite . in an infinite domain
this ` homoclinic snaking ' behaviour continues , and the width of the localized states increases without bound .
in a finite domain the branches of localized states terminate on the branches of spatially periodic patterns which also originate at the initial turing instability @xcite .
there is now a pretty complete picture of the existence and stability of localized states in canonical 1d model equations , for example the swift hohenberg model @xcite with quadratic - cubic or cubic - quintic nonlinearities , but in contrast , comparatively little is known about the detailed structure of localized states in spatially discrete problems .
motivation for the detailed investigation of discrete models comes from a number of sources , but in particular discrete models arise as natural descriptions of arrays of waveguides in nonlinear optics @xcite .
for example , the discrete bistable nonlinear schrdinger equation i_n + c_n + s_n |_n|^2 - _ n|_n|^4 = 0 , [ eqn : dnls ] for an array of complex fields
@xmath0 naturally describes properties of discrete optical ` dissipative solitons ' in a @xmath1-dimensional square lattice whose sites are indexed by the ( multiple ) suffix @xmath2 .
the operator @xmath3 is an appropriate sum of nearest neighbour ( and possibly next - nearest neighbour ) differences depending on the lattice dimension coefficient @xmath1 . substituting the ansatz @xmath4 where @xmath5 is a real stationary field
, we see that @xmath5 must satisfy the real difference equation u_n + c u_n + su_n^3 - u_n^5 = 0 .
[ eqn : ac ] hence the possible steady states correspond to those of a discrete allen cahn equation @xcite , albeit with a cubic - quintic nonlinearity rather than the well - studied purely cubic case . in this letter
we investigate the existence of localized states in the canonical nonlinear lattice model ( [ eqn : ac ] ) in one and two dimensions .
such models are in many senses simpler than the corresponding spatially continuous equations and we discuss the similarities and differences that arise . to aid this discussion we add a time derivative term to ( [ eqn : ac ] ) which makes stability correspond to that found in the spatially - continuous swift hohenberg model .
in contrast to previous work on this canonical problem by other authors , our central focus in this paper is on the appearance of isolas .
in particular we show that , firstly , in 1d and on a finite lattice , isolas are an immediate consequence of changing the boundary conditions from periodic to neumann ( or , in fact , dirichlet ) . secondly we show that in 2d isolas exist when periodic boundary conditions are used , and reconnect and then disconnect again as the cell to cell coupling parameter is increased .
we also comment on ( i ) the overall bifurcation structure in 1d , ( ii ) the similarities between the 1d bifurcation structure and that of the continuum 1d swift hohenberg model and ( iii ) the connections between 1d lattice structures and 2d lattice structures , using the former to explain features of the bifurcation diagram we obtain for the latter .
the structure of the letter is as follows . in section [ sec:1d ]
we introduce the 1d lattice case and explore the bifurcation structure of localized states .
we discuss the effects of different boundary conditions when the problem is restricted to a finite lattice . in section [ sec:2d ]
we consider a 2d lattice , highlight the new features that arise , and use the 1d results to explain aspects of the snaking curve .
we summarise the main results of the paper in section [ sec : conclusions ] .
motivated by , after applying the rescaling @xmath6 , rescaling @xmath7 and @xmath8 appropriately , dropping the @xmath9 and adding a time derivative term appropriate for strongly dissipative rather than conservative dynamics , we consider the following equation posed on the integer lattice @xmath10 : [ eq:1d ] _
n = c(u_n+1+u_n-1 - 2u_n ) + u_n + 2u_n^3 - u_n^5 , where @xmath5 is a real scalar variable defined at each lattice site , @xmath8 is the strength of linear coupling between adjacent sites , and @xmath11 is the primary ( real ) bifurcation parameter .
the addition of a time derivative term is , of course , not necessary for the investigation of the existence of equilibrium states .
however , it is a useful addition to indicate stability and to provide a contextual similarity to the studies of the swift hohenberg equation in spatially continuous systems referred to above .
if @xmath12 then we apply the parameter symmetry given by applying the ` staggering ' transformation [ eq : stagger ] u_n ( -1)^n u_n , c -c , - 4c , which leaves the equation unchanged , but switches the sign of @xmath8 .
thus in this paper we will only consider positive @xmath8 , noting that the existence ( and stability ) of states @xmath13 with @xmath14 is equivalent to the existence ( and stability ) of states @xmath15 when @xmath12 and @xmath7 is shifted as in ( [ eq : stagger ] ) .
importantly , and in common with the swift hohenberg equation , equation is a gradient flow , that is , we can write @xmath16 where the potential @xmath17 is given by f(u ) = _ n c ( u_n+1-u_n)^2 - u_n^2 - u_n^4 + u_n^6 .
thus @xmath18 , so that every solution of flows down gradients of the potential toward an equilibrium solution , and every stable equilibrium state is a local minimum of the potential .
no periodic or complex dynamics are therefore possible .
there are five homogeneous equilibria : the trivial state @xmath19 , and four non - trivial equilibria @xmath20 , where u_^2 = 1 .
the state @xmath19 is linearly stable for @xmath21 and unstable for @xmath22 .
taking the positive square roots , the lower uniform state @xmath23 exists in @xmath11 and is always unstable , and the upper uniform state @xmath24 exists in @xmath25 and is always stable ; there is a saddle - node bifurcation at @xmath26 .
the per - cell potential for a homogeneous equilibrium @xmath27 is given by @xmath28 .
since the system acts to minimize the total potential , the relative values of potential at the homogeneous equilibria are important .
the trivial state has zero potential , and the potential of the upper state @xmath29 depends only on @xmath7 . by looking for a double root of @xmath30
we find that the upper state has zero potential when @xmath31 ; this is the so - called _ maxwell point _ at which the upper uniform state and the trivial state are energetically equal .
when @xmath32 the upper state is energetically more favourable ; when @xmath33 the trivial state is energetically more favourable . in many physical applications
it is appropriate to consider posed on a finite grid @xmath34 with boundary conditions specified with the aid of ` ghost cells ' @xmath35 and @xmath36 .
there are three obvious choices : periodic boundary conditions ( pbc ) for which @xmath37 and @xmath38 , neumann boundary conditions ( nbc ) for which @xmath39 and @xmath40 , and dirichlet boundary conditions ( dbc ) for which @xmath41 . with pbc or nbc
the homogeneous state is an equilibrium configuration . with dbc
this is not the case and this obvious difference leads us to focus in this discussion only on pbc and nbc : we find there are substantial differences between these two cases that require some discussion .
equation posed on the infinite integer lattice has a symmetry group generated by the independent operations of translation @xmath42 , reflection @xmath43 and an up - down symmetry @xmath44 ( since the nonlinearity is an odd function of @xmath5 ) : : u_nu_n+1 , : u_nu_-n,:u_n - u_n .
thus the full symmetry group of the infinite lattice problem is @xmath45 . posing the equation on a finite lattice implies changes in the symmetry group .
consider the following actions of the elements @xmath42 , @xmath43 and @xmath44 : : u_nu_n+1(n ) , : u_nu_n+1-n,:u_n - u_n .
with pbc the translation subgroup @xmath10 , generated by @xmath42 , is simply replaced with the cyclic group @xmath46 so that the full symmetry group is @xmath47 . with nbc or dbc
we lose translation symmetry entirely , so the symmetry group becomes @xmath48 .
we will be particularly interested in localized solutions to which lie in @xmath49 or @xmath50 , which are symmetric under reflection either about a lattice point , or about a point halfway between two lattice points .
we refer to such solutions _ site - centred _ and _ bond - centred_. when @xmath51 is even , site - centred solutions lie in @xmath50 and bond - centred solutions lie in @xmath49 .
when @xmath51 is odd , site - centred solutions lie in @xmath49 and bond - centred solutions lie in @xmath50 .
the eigenfunctions @xmath52 and eigenvalues @xmath53 of the discrete laplacian , obtained by solving @xmath54 on an @xmath51-cell lattice with pbc are easily computed to be @xmath55 and @xmath56 where @xmath57 for @xmath58 .
thus the zero state undergoes an instability to a mode with ` wavenumber ' parameter @xmath59 when @xmath60 .
since @xmath14 the first instability encountered when increasing @xmath7 is a bifurcation to the lower uniform state @xmath61 at @xmath62 .
this branch bifurcates subcritically , turning around in a saddle - node bifurcation and restabilising at @xmath26 to form the upper uniform state @xmath29 as shown in figure [ fig:1dsnake ] . )
solved on a 1d lattice with @xmath63 using pbc and taking @xmath64 .
profiles correspond to the right - hand saddle - node bifurcation points of figure [ fig:1dsnake](a ) .
the upper three solutions are site - centred and the lower three solutions are bond - centred.,title="fig:",scaledwidth=32.0% ] ) solved on a 1d lattice with @xmath63 using pbc and taking @xmath64 .
profiles correspond to the right - hand saddle - node bifurcation points of figure [ fig:1dsnake](a ) .
the upper three solutions are site - centred and the lower three solutions are bond - centred.,title="fig:",scaledwidth=32.0% ] ) solved on a 1d lattice with @xmath63 using pbc and taking @xmath64 .
profiles correspond to the right - hand saddle - node bifurcation points of figure [ fig:1dsnake](a ) .
the upper three solutions are site - centred and the lower three solutions are bond - centred.,title="fig:",scaledwidth=32.0% ] + ) solved on a 1d lattice with @xmath63 using pbc and taking @xmath64 .
profiles correspond to the right - hand saddle - node bifurcation points of figure [ fig:1dsnake](a ) .
the upper three solutions are site - centred and the lower three solutions are bond - centred.,title="fig:",scaledwidth=32.0% ] ) solved on a 1d lattice with @xmath63 using pbc and taking @xmath64 .
profiles correspond to the right - hand saddle - node bifurcation points of figure [ fig:1dsnake](a ) .
the upper three solutions are site - centred and the lower three solutions are bond - centred.,title="fig:",scaledwidth=32.0% ] ) solved on a 1d lattice with @xmath63 using pbc and taking @xmath64 .
profiles correspond to the right - hand saddle - node bifurcation points of figure [ fig:1dsnake](a ) .
the upper three solutions are site - centred and the lower three solutions are bond - centred.,title="fig:",scaledwidth=32.0% ] the upper uniform state @xmath29 undergoes no instabilities as @xmath7 is increased .
in contrast , the lower state @xmath61 has many secondary bifurcation points , occuring at = _ k^-- c^2(_k/2 ) , [ eq : instability ] for values of @xmath1 such that the expression under the square root is positive
. the bifurcations are supercritical ` double pitchforks ' , each creating two new bifurcating branches of ` odd ' and ` even ' modulated states which develop into single- and multi - pulse localized states , each of which terminates on @xmath61 at both ends . in figure [ fig:1dsnake ]
only the single pulse branches are shown for clarity .
this behaviour of secondary bifurcation creating branches of odd and even - symmetric modulated and localized states is extremely similar to that given for the continuum swift
hohenberg equation in @xcite without the additional complexity that arises from continuous changes of wavenumber along the branch which is prohibited in the present case by the spatial discreteness .
noting that the instability creating single - pulse localized states is the @xmath65 branch , inspection of ( [ eq : instability ] ) shows that the @xmath65 case is the last instability to remain as @xmath8 increases at fixed @xmath51 .
hence the maximum coupling strength above which there are no modulational instabilities of @xmath61 is @xmath66 $ ] .
@xmath67 therefore also provides an upper bound on the lower limit of the range of @xmath8 over which localized states asymptotic to zero may exist .
we use the numerical continuation program to explore the bifurcation structure of localized solutions to using both pbc and nbc .
stability results come directly from , since is simply a collection of odes . with pbc
the bifurcation diagram figure [ fig:1dsnake ] closely resembles the homoclinic snaking picture for the swift hohenberg equation in a finite domain , with branches of localized solutions emerging from the uniform branch at @xmath68 and turning around in a succession of fold bifurcations to form two intertwined snakes of localized structures , one of site - centred and one of bond - centred solutions , before reconnecting to the uniform branch at @xmath69 . for @xmath51 even , as is the case in figure [ fig:1dsnake ] , these solutions lie in @xmath50 and @xmath49 respectively ; this is shown clearly in figure [ fig:1dprofiles ] .
again , just as in the continuum swift
hohenberg equation , the two snaking branches are connected by ` ladders ' of asymmetric structures : the ladders comprise solutions with no symmetry at all that are always unstable .
these branches bifurcate subcritically from the main snake , at locations near to the fold bifurcations making up the main body of the snake ( not shown in figure [ fig:1dsnake ] ) .
solution profiles at the lowest ( in terms of @xmath70 norm ) six right - hand saddle nodes are shown in figure [ fig:1dprofiles ] .
plane : @xmath71-cell localized state ( red ) , @xmath72-cell ( green ) and @xmath73-cell ( magenta ) .
the front solution is indicated by the blue curves : on an infinite lattice the edges of the snaking region asymptote to these curves as the width of the localized state increases .
the vertical dashed line is the maxwell point @xmath74 for reference .
results obtained for @xmath75.,scaledwidth=80.0% ] with pbc the bifurcation structure of the discrete model exactly parallels that of the continuous bistable swift - hohenberg equation . when we switch to nbc one of the snaking branches persists but the other fragments into a series of isolas ( closed curves of equilibria that no longer connect up into a snake ) and ` bubbles ' ( curves of equilibria that begin and end in bifurcations from the persistent snaking curve ) . in figure [ fig:1dsnake ]
the persistent snaking branch is indicated in black , isolas are indicated by the red curves and the bubbles are shown in blue . for @xmath51 even , the persistent snaking curve consists of bond - centred states with symmetry @xmath49 .
the two halves of the pitchfork bifurcation where the bond - centred states bifurcate from the uniform state are no longer related by symmetry : in figure [ fig:1dsnake](b ) we show the branch corresponding to states that are localized in the centre of the lattice rather than at the edges of the lattice .
the snaking curve of site - centred states in figure [ fig:1dsnake](a ) breaks up in figure [ fig:1dsnake](b ) into isolas and bubbles which do not directly interact . in figure
[ fig:1dsnake](b ) the lowest isola approaches the bifurcation point where the bond - centred branch connects to the uniform branch but does not bifurcate from either . as the @xmath76 norm increases , isolas alternate with bubbles which do bifurcate from the bond - centred snaking branch .
the bubbles bifurcate from points close to where the ` ladders ' bifurcate in figure [ fig:1dsnake](a ) and lack symmetry : in fact they comprise states which evolve smoothly from the exactly bond - centred state on the snaking curve and become very close to the subspace @xmath50 before moving back towards @xmath49 and reconnecting with the snaking branch further up .
moreover there are , properly speaking , four copies of each isola and bubble branch of site - centred asymmetric localized states , related by the symmetries @xmath44 and @xmath43 .
this behaviour mimics what has been observed for the swift hohenberg equation with non - periodic boundary conditions @xcite .
it is important to note however , that in contrast with these investigations , the domain size can not be used as a continuously varying parameter in the present context .
in fact , we anticipate that increasing @xmath51 by unity reverses the roles of the bond - centred and site - centred solutions when the boundary conditions are left unchanged .
for example , with nbc as we consider here , when @xmath51 is odd we expect the site - centred snaking curve to persist and the bond - centred one to fragment into isolas and bubbles in a manner similar to that shown in figure [ fig:1dsnake ] .
in two dimensions our interpretation of ( [ eqn : ac ] ) must be modified since there are additional possibilities for the coupling terms . on the square lattice @xmath77
there are two natural nearest - neighbour difference operators : @xmath78 this leads to the following 2d generalization of ( [ eq:1d ] ) [ eq:2d ] _ nm = c^+u_nm + c^u_nm + u_nm + 2u_nm^3 - u_nm^5 , where there are now two coupling parameters @xmath79 and @xmath80 , the coefficients of nearest - neighbor ( nn ) and next - nearest - neighbor ( nnn ) coupling . as in the 1d case , ( [ eq:2d ] ) is variational .
for this equation our primary interest is in establishing that even with pbc there are isolas containing stable localized states .
therefore , we confine our presentation to localized solutions that are far from the lower and upper ends of the snake where reconnection to the uniform state occurs in a finite domain . to further simplify the discussion we set @xmath81 for the majority of this letter .
clearly it is of interest to investigate how the snaking structure evolves with @xmath80 nonzero , and we will return to this and other issues in subsequent work . ,
@xmath81 and @xmath82 .
the three solutions are respectively site - centred , bond - centred and hybrid.,title="fig:",scaledwidth=32.0% ] , @xmath81 and @xmath82 .
the three solutions are respectively site - centred , bond - centred and hybrid.,title="fig:",scaledwidth=32.0% ] , @xmath81 and @xmath82 .
the three solutions are respectively site - centred , bond - centred and hybrid.,title="fig:",scaledwidth=32.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] + , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] + , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] + , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] + , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] , moving along the snaking curve with @xmath83 increasing , for fixed @xmath84 .
labels ( a ) - ( f ) correspond to the labelled saddle - node bifurcations on the snake indicated in figure [ fig:2dsnake](b).,title="fig:",scaledwidth=16.0% ] the bifurcation structure of localized solutions in 2d is made complicated by the existence of at least three snaking curves : a localized state can be site - centred , bond - centred or a hybrid of the two , being site - centred in one lattice direction and bond - centred in the other ( for a more complete discussion see @xcite ) .
examples of each of these types of solutions are shown in figure [ fig : local2d ] .
for simplicity we will now focus on bond - centred states and follow the evolution of snaking curves of such states .
figure [ fig : bond - centred ] shows the evolution of bond - centred localized states along the continuous snaking curve shown in figure [ fig:2dsnake](b ) .
it is not at all clear theoretically how the well - established theory for homoclinic snaking in 1d extends to 2d since there does not appear to be an analogous paradigm to the spatial dynamics approach .
however , numerical computations show that the bifurcation diagram for localized states contains qualitative similarities ( see figure [ fig:2dsnake ] ) .
we remark that to generate this figure our computations were carried out using on a @xmath85 grid . to accelerate the computation we simulated only a quarter of the domain , using appropriate ` reflecting ' boundary conditions to ensure that the new states generated were bond centred .
this imposed symmetry implies that bifurcations to asymmetric ` ladder ' states could not be detected .
other computations indicate that such bifurcations exist , and we will discuss the existence of asymmetric states elsewhere .
the states shown in figure [ fig : bond - centred ] lie on the snaking curve shown in figure [ fig:2dsnake](b ) in which the bifurcation parameter @xmath7 is plotted on the vertical axis and the solution measure @xmath86 on the horizontal axis .
@xmath83 is a scaled version of the @xmath76 norm defined as : @xmath87 use of the solution measure @xmath83 confers the advantage that at low coupling strengths it measures the ( near - integer ) number of cells that are ` active ' in the localized state .
figure [ fig:2dsnake](a ) indicates that ` high peaks ' ( e.g. the saddle - nodes labelled ( a ) , ( b ) , ( c ) , ( e ) and ( f ) at which the value of @xmath7 is markedly less negative than at most saddle - nodes ) are closely aligned with even integer values of @xmath83 .
these labels correspond to the labelled localized states in figure [ fig : bond - centred ] that are in the form of complete @xmath88 square arrays of active cells .
the evolution of bond - centred localized states as we move up the snaking curve is shown in figure [ fig : bond - centred ] .
starting from [ fig : bond - centred](a ) in which four cells are close to @xmath29 and the rest are near zero , new @xmath29 cells are added to the solution profile starting in the centre of the faces , and then progressing outwards towards the corners until the profile is square , at which point the sequence repeats itself .
numerical continuation suggests that it is possible to proceed arbitrarily far up the snaking curve in this manner as long as the coupling coefficient is sufficiently small .
there are similar snaking curves of site - centred and hybrid localized states which for brevity , and because they are not central to our discussion of isolas , we will not discuss here .
the horizontal lines in each part of figure [ fig:2dsnake ] indicate the locations of saddle - node bifurcations far up the 1d snake for two values of the coupling coefficient : @xmath89 ( red ) and @xmath90 ( blue ) .
these values are determined from the part of the blue curve in figure [ fig : muc ] which lies in @xmath91 . at small coupling strengths it is clear numerically in figure [ fig:2dsnake](a ) that , as @xmath86 increases and we move far up the snaking curve , the width of the snaking region is given overall by @xmath92 since the existence of square @xmath88 arrays of localized states is determined by the existence of stationary fronts aligned with a lattice direction .
this is easy to see from the lattice equations [ eq:2d ] ) by removing the dependence on the second coordinate @xmath93 , i.e. setting @xmath94 we recover ( [ eq:1d ] ) with @xmath95 .
all other saddle - node bifurcations involve localized states which are not complete squares and therefore involve stationary diagonal fronts between @xmath29 and zero . substituting the ansatz @xmath96 where @xmath97 into ( [ eq:2d ] ) we obtain [ eq:2dreduced ] _
= 2c^+(u_+1+u_-1 - 2u _ ) + c^(u_+2+u_-2 - 2u _ ) + u_+ 2u_^3 - u_^5 .
hence , in the case @xmath81 this reduces exactly to ( [ eq:1d ] ) with @xmath98 . when @xmath99 and the coupling coefficients are small ( and hence the front is sharp compared to the lattice spacing ) we anticipate that @xmath100 on one side of the front and @xmath101 on the other and so ( [ eq:2dreduced ] ) can be well approximated by ( [ eq:1d ] ) setting @xmath102 .
figure [ fig:2dsnake](d ) shows the bond - centred snake with @xmath103 and @xmath104 . qualitatively there is little difference between this and parts ( a ) and ( b ) .
we observe two distinct scales in the bifurcation diagram , with the widths of turns on the snake corresponding roughly to those in the 1d model with @xmath89 or @xmath90 .
we anticipate that for @xmath80 small compared with @xmath79 there will be no significant differences from the @xmath81 case , but numerical continuation suggests that larger @xmath80 can alter the snaking curve significantly , especially when @xmath105 .
we should also note that with two independent coupling parameters there is no equivalent to the staggering transformation presented for the 1d model . as the coupling strength @xmath79 increases the snaking curves qualitatively
depart from the 1d case due to the existence of regions in the bifurcation diagram where the snaking curve turns back on itself and the @xmath76 norm ( or , equivalently , @xmath83 ) of the localized state descreases for a number of twists before turning back again and continuing to snake upwards .
we refer to such an episode as a ` switchback ' .
examples of switchbacks are shown in figure [ fig:2dsnakecloseups ] . in part
( a ) of figure [ fig:2dsnake ] switchbacks are observed to begin only at @xmath106 , @xmath107 and @xmath108 . in
( b ) an additional switchback appears at @xmath109 . in ( c )
almost all the twists in the switchbacks have disappeared , although the bifurcation curves near @xmath110 , @xmath107 and @xmath108 show clear similarities .
we find that the appearance and disappearance of switchbacks as the coupling strength @xmath79 is varied can be explained by collisions between solution curves in the bifurcation diagram .
when @xmath79 is small it is important to note that there are many localized solutions to possessing square symmetry which do not appear on the snaking curve , e.g. as shown in figure [ fig : switchback](a ) .
these additional localized states exist on isolas in the bifurcation diagram . as @xmath79
is increased these isolas collide with the snaking branch and form a switchback .
assuming , as is reasonable at low coupling strength , that the isola contained a stable solution , the snaking curve will now contain ( at least ) one extra stable solution , by which we mean a localized state now present , but not present on the snake at lower @xmath79 .
as @xmath79 is increased still further the snake narrows .
intriguingly , and unexpectedly , the switchback may detach itself from the snake as @xmath79 is increased further , re - forming a disconnected isola .
figure [ fig : switchback ] illustrates these events with a sequence of bifurcation diagrams in the @xmath111 plane .
the processes of attachment and detachment are facilitated by cusp singularities in which pairs of saddle - nodes on the snake collide and disappear as the coupling strength increases .
figure [ fig : detachment ] shows the cusp bifurcations in the @xmath112 plane corresponding to the attachment and detachment of the isola near @xmath113 shown in figure [ fig:2dsnakecloseups](a ) and figure [ fig : switchback ] . in summary
, these isolas provide a mechanism by which additional turns can be added to the snake , and they explain the unintuitive switchback episodes of the snaking curve as the localized states increase in size .
this process of isola attachment and detachment typically occurs over a remarkably small range of @xmath79 , indicated by the vertical scales in figure [ fig : detachment ] . at larger @xmath76 norms the complete bifurcation diagram in the @xmath112 plane is complicated since many isolas attach and detach from the snake at nearby values of @xmath79 .
plane by following limit points : on the left an isola attaches itself to the snake as @xmath79 is increased , briefly creating a situation with three limit points , two of which disappear in a cusp singularity [ corresponding to fig .
[ fig : switchback](a)(b ) ] , and on the right the same process in reverse , as an isola detaches at a higher value of @xmath79 [ corresponding to fig .
[ fig : switchback](c)(d)].,title="fig:",scaledwidth=49.0% ] plane by following limit points : on the left an isola attaches itself to the snake as @xmath79 is increased , briefly creating a situation with three limit points , two of which disappear in a cusp singularity [ corresponding to fig .
[ fig : switchback](a)(b ) ] , and on the right the same process in reverse , as an isola detaches at a higher value of @xmath79 [ corresponding to fig .
[ fig : switchback](c)(d)].,title="fig:",scaledwidth=49.0% ]
in this letter we have discussed the bifurcation structure of a model lattice equation , the discrete allen - cahn equation with cubic and quintic nonlinearities , in one and two dimensions .
this equation is of interest in its own right as a model for spatially discrete pattern formation as well as being closely related to coupled nonlinear schrdinger equations which have received substantial recent attention in nonlinear optics .
a subcritical instability of the trivial state , and the resulting bistability , leads to the creation of localized states which are stable over an open interval of the parameter @xmath7 , stabilized by being pinned to the lattice . as remarked on by previous authors ,
these localized states arrange themselves into homoclinic snaking curves . in a finite domain we demonstrated the existence of a maximum coupling strength
above which there is no modulational secondary instability which would allow creation of localized states : in systems with a sufficiently large coupling strength localized states do not appear . in 1d we showed that breaking discrete translational symmetry , for example by using neumann boundary conditions , allows one of the snakes to persist , but cuases the other to fragment into a sequence of isolas and bubbles ( ` s'shaped curves bifurcating from the persistent snake ) .
this contrasts with the situation in 2d where isolas exist even when discrete translational symmetry is not broken .
these isolas attach and detach themselves from the primary snaking curve as the linear coupling strength is varied , creating kinks in the snaking curve , where the solution norm decreases along a part of the curve before beginning to increase again , that we refer to as ` switchbacks ' .
numerical investigations suggest quantitative relations between the widths of the snaking curves in 1d and 2d . for wide localized states in 2d
the values of @xmath7 that limit the region over which localized states exist correspond to the limiting values for fronts aligned with lattice directions or on the diagonal .
this is the natural interpretation and agrees with the observations and computations by lloyd et al .
@xcite for localized patches of hexagons in the 2d swift hohenberg equation .
this correspondance between 1d and 2d is asymptotically exact in the limit of small coupling coefficients : higher - order corrections arise at non - zero coupling and from the corners of the wide localized state where the fronts meet .
our results are closely related to the work of carretero - gonzles and chong et al .
@xcite on the discrete nonlinear schrdinger equation with cubic and quintic nonlinearities .
they discussed the time - independent allen - cahn equation from a spatial dynamics viewpoint , explaining the one - dimensional snaking curve as arising from intersections of the stable and unstable manifolds of a hyperbolic fixed point in a two - dimensional map .
they also explored a variational approach , minimizing an effective lagrangian to give approximate localized states in one and two dimensions .
our work considers additional mechanisms that complicate the usual snaking picture by causing isolas to appear : these include the effect of boundary conditions in 1d and exploring larger localized states , higher up the snaking curve in 2d where new effects , including the switchbacks , arise .
collier , n.a.m .
monk , p.k .
maini & j.h .
lewis 1996 pattern formation by lateral inhibition with feedback : a mathematical model of delta notch intercellular signalling
. _ j. theor .
biol . _ * 183 * , 429446 e. j. doedel , a.r champneys , t. fairgrieve , y. kuznetsov , b. oldeman , r. paffenroth , b. sandstede , x. wang & c. zhang 2007 auto-07p : continuation and bifurcation software for ordinary differential equations .
available to download from ` http://indy.cs.concordia.ca/auto/ ` . | we consider localized states in a discrete bistable allen - cahn equation .
this model equation combines bistability and local cell - to - cell coupling in the simplest possible way .
the existence of stable localized states is made possible by pinning to the underlying lattice ; they do not exist in the equivalent continuum equation .
in particular we address the existence of ` isolas ' : closed curves of solutions in the bifurcation diagram .
isolas appear for some non - periodic boundary conditions in one spatial dimension but seem to appear generically in two dimensions .
we point out how features of the bifurcation diagram in 1d help to explain some ( unintuitive ) features of the bifurcation diagram in 2d .
homoclinic snaking , dissipative soliton , coupled cells 05.45.-a , 42.65.sf |
the physical size and shape of the materials strongly effect the nature , the dynamics of the electronic excitations , the lattice vibrations , and the dynamics of carriers .
for example , in the mesoscopic systems , the dimension of system is comparable with the coherence length of carriers and this leads to some new phenomena that they do not appear in a bulk semiconductor , such as quantum interference between carrier s motion @xcite . in these physical systems different particles
are confined in a small space and interact with each other . as usual , we use quantum field theory ( qft ) and second quantization procedure for considering interacting many particles physical systems .
standard qft is based on quantum mechanics on an infinite line without any boundaries .
however , the presence of infinite walls in standard qft can detect vacuum effect of electromagnetic field and gives rise to casimir effect @xcite .
hence , in a system with small dimensions we expect some new phenomena appear , and barriers effects show themselves . + recent progress in growth techniques and development of micromachinig technology in designing mesoscopic systems and nanostructures , have led to intensive theoretical @xcite and experimental investigations @xcite on electronic and optical properties of those systems .
the most important point about the nanoscale structures is that the quantum confinement effects play the center - stone role .
one can even say in general that recent success in nanofabrication technique has resulted in great interest in various artificial physical systems with usual phenomena driven by the quantum confinement ( quantum dots , quantum wires and quantum wells ) .
a number of recent experiments have demonstrated that isolated semiconductor quantum dots are capable of emitting light @xcite .
it becomes possible to combine high - q optical microcavities with quantum dot emitters as the active medium @xcite .
furthermore , there are many theoretical attempts for understanding the optical and electronic properties of nanostructures especially semiconductor quantum dots @xcite . because of intensive researches in this area , it is reasonable to consider the finite size effects on the em field including the quantization of the em field in confined regions that their sizes are of order of electromagnetic wavelength , such as microcavities . on the other hand ,
a nanostructure such as quantum dot , is a system that carrier s motion is confined inside a small region , and during the interaction with other systems , the generated excitations such as phonons , excitons , plasmons are confined in small region .
hence we want to answer this question : what are the spatial confinement effects on excitation states in quantum field theoretical description of nanostructures ?
it seems that to answer this question we need to know the confinement and boundary conditions effects in qft .
first , we consider spatial confinement effect on a simple quantum harmonic oscillator and then we shall use this oscillator in quantizing the fields .
+ as mentioned before , the standard qft is based on the quantum mechanics on an infinite line . in the canonical qft
the main tool is quantum oscillator .
energy eigenvalues of quantum harmonic oscillator ( qho ) is given by @xmath0 , and these successive energy levels were interpreted as being obtained by creation of a quantum particle of energy @xmath1 .
this interpretation of the energy spectrum of qho was successfully used in the second quantization formalism @xcite .
plank s hypothesis is realized in the second quantization formalism by using creation and annihilation operators of the qho .
this realization is obtained for qho defined on an infinite line .
+ it is reasonable to claim that , in considering qft in a finite region one can use energy levels of a qho confined in that finite space and therefore analyze the consequences of this assumption in construction of such qft on a compact manifold .
as we shall see in subsequent sections , the spatial confinement of the qho leads to a deformed heisenberg algebra for the ordinary harmonic oscillator .
a deformed algebra is a nontrivial generalization of a given algebra through the introduction of one or more complex parameters , such that , in a certain limit of parameters the non - deformed algebra is recovered ; these parameters are called deformation parameters .
there have been several attempts to generalize heisenberg algebra , and a particular deformation of heisenberg algebra has led to the notion f - oscillator @xcite .
an f - oscillator is a non - harmonic system , that from mathematical point of view its dynamical variables ( creation and annihilation operators ) constructed from a non canonical transformation through @xmath2 where @xmath3 and @xmath4 are corresponding harmonic oscillator operators and @xmath5 .
the function @xmath6 is called deformation function that depends on the number of quanta and some physical parameters .
the presence of operator - valued deformation function causes the heisenberg algebra of the standard qho to transform into a deformed heisenberg algebra .
the nonlinearity in f - oscillators means dependence of the frequency on the intensity @xcite . on the other hand , in contrast to the standard qho , f
- oscillators have not equal spaced energy spectrum .
if we confine a simple qho inside an infinite well , due to the spatial confinement , the energy levels constitute a spectrum that is not equal spaced .
therefore , in this case it is reasonable to expect to find a corresponding f - oscillator .
one of the most interesting features of the qho is the construction of coherent states , as the eigenfunction of annihilation operator .
as is well known @xcite one can introduce nonlinear coherent states or f - coherent states as the right - hand eigenstates of deformed annihilation operator @xmath7 .
it has been shown that these families of generalized coherent states exhibit various non - classical properties @xcite . due to
these properties and their applications , generation of these states is a very important issue in the context of quantum optics .
the f - coherent states may appear as stationary states of the center - of - mass motion of a trapped ion @xcite .
furthermore , a theoretical scheme for generation of these states in micromaser in the frame work of intensity - dependent jaynes - cummings model has been proposed @xcite .
+ it has also been shown @xcite that there is a close connection between the deformation function appeared in the nonlinear coherent states algebraic structure and the non - commutative geometry of the configuration space .
furthermore , it has been shown recently @xcite , that if a two - mode qho confined on the surface of a sphere , can be interpreted as a single mode deformed oscillator , whose and its quantum statistics depends on the curvature of sphere .
+ motivated by the above - mentioned results , in the present contribution we are intended to investigate the spatial confinement effects on physical properties of a standard qho .
it will be seen that the confinement leads to deformation of standard qho
. then we use this confined oscillator to considering boundary effects in qft . in a recent work @xcite
the authors have considered boundary effects in qft and for this purpose they have used a qho defined on a circle and its associated algebra , which is a realization of a deformed heisenberg algebra has been introduced in ref.@xcite . to construct qft
they have used this special deformed algebra and the calculus on a lattice without any definite commutation relation between field operators . in this paper
, we consider a qho confined in a one - dimensional infinite well without periodic boundary conditions , and we find its energy levels , as well as associated ladder operators .
we show that the ladder operators can be interpreted as a special kind of the so - called f - deformed creation and annihilation operators @xcite .
then , we use this oscillator as a basis for the canonical quantization of the electromagnetic ( em ) field in a confined space . in ref .
@xcite the quantization of the electromagnetic field is performed by making use of the q - deformed oscillator without any quantization postulate . in our quantization scheme
we use the quantization postulate and impose canonical commutation relation on hamiltonian of the system under consideration . in order to keep commutation relation between field and its conjugate momentum we deform hilbert space of the system .
+ this paper is organized as follow : in section 2 , we review some physical properties of f - oscillator and its coherent states . in section 3
we consider the spatially confined qho in a one - dimensional infinite well and construct its associated coherent states .
we shall also examine some of their quantum statistical properties , including sub - poissonian statistics and quadrature squeezing .
in section 4 we use the confined oscillator under consideration and its algebra to construct a quantum theory of fields , and as an example we quantize the electromagnetic field . in section 5 we propose a dynamical scheme for generating the nonlinear coherent state associated with the em field in a confined region . finally we summarize our conclusions in section 6 .
in this section , we review the basics of the f - deformed quantum oscillator and the associated coherent states known in the literature as nonlinear coherent states . for this purpose , we consider an eigenvalue problem for a given quantum physical system and we focus our attention on the properties of creation and annihilation operators , that allows to make transition between the states of discrete spectrum of the system hamiltonian . as usual , we expand the hamiltonian in its eigenvectors @xmath8 where we choose @xmath9 .
we introduce the creation ( raising ) and annihilation ( lowering ) operators as follows @xmath10 so that @xmath11 .
these ladder operators satisfy the following commutation relation @xmath12=\sum_{i=1}^{n}(e_{i+1}-e_{i})|i\rangle\langle i|\:.\ ] ] obviously if the energy spectrum is equally spaced , because of this condition , energy spectrum must be linear in quantum numbers , ( as in the case of ordinary qho ) , then @xmath13 , where @xmath14 is a constant and the commutator of @xmath7 and @xmath15 becomes a constant ( a rescaled weyl - heisenberg algebra ) .
on the other hand , if the energy spectrum is not equally spaced , the ladder operators of the system satisfy a deformed heisenberg algebra , i.e. their commutator depends on quantum numbers that appear in energy spectrum .
this is one of the most important properties of the quantum f - oscillators @xcite .
+ an f - oscillator is a non - harmonic system characterized by a hamiltonian of the harmonic oscillator form @xmath16 with a specific frequency @xmath17 and deformed boson creation and annihilation operators defined in ( [ defo ] ) .
the deformed operators obey the commutation relation @xmath18=(\hat{n}+1)f^2(\hat{n}+1)-\hat{n}f^2(\hat{n})\:.\ ] ] the f - deformed hamiltonian @xmath19 is diagonal on the eingenstates @xmath20 in the fock space and its eigenvalues are @xmath21.\ ] ] in the limit @xmath22 , the ordinary expression @xmath23 and the usual ( non - deformed ) commutation relation @xmath24=1 $ ] are recovered .
+ furthermore , by using the heisenberg equation of motion with hamiltonian ( [ hami ] ) we have @xmath25\hspace{1cm}(\hbar=1).\ ] ] we obtain the following solution to the heisenberg equation of motion for f - deformed operators @xmath7 and @xmath15 defined in equation ( [ defo ] ) @xmath26 where @xmath27 in this sense , the f - deformed oscillator can be interpreted as a nonlinear oscillator whose frequency of vibrations depends explicitly on its number of excitation quanta @xcite .
it is interesting to point out that recent studies @xcite have revealed strictly physical relationship between the nonlinearity concept resulting from f - deformation and some nonlinear optical effects , e.g. , kerr nonlinearity , in the context of atom - field interaction .
+ the nonlinear transformation of the creation and annihilation operators leads naturally to the notion of nonlinear coherent states or f - coherent states .
the nonlinear coherent states @xmath28 are defined as the right - hand eigenstates of the deformed operator @xmath29 @xmath30 from eq.([coh ] ) one can obtain an explicit form of the nonlinear coherent states in a number state representation @xmath31 where the coefficients @xmath32 s and normalization constant @xmath33 are respectively given by @xmath34 in recent years nonlinear coherent states have been paid much attentions because they exhibit nonclassical features @xcite and many quantum optical states , such as squeezed states , phase states , negative binomial states and photon - added coherent states can be viewed as a sort of nonlinear coherent states @xcite .
in this section we consider a quantum harmonic oscillator confined in a one dimensional infinite well .
many attempts have been done for solving this problem ( see @xcite-@xcite , and references therein ) . in most of those works , authors tried to solve the problem numerically . but
in our consideration we try to solve the problem analytically , to reveal the relationship between the confinement effect and given deformation function .
we start from the schrdinger equation ( we assume @xmath35 ) @xmath36\psi(x)=e\psi(x),\ ] ] where @xmath37 instead of solving the schrdinger equation for the qho confined between infinite rectangular walls in positions @xmath38 , we propose to solve the eigenvalue equation for the potential @xmath39 where @xmath40 , is a scaling factor depending on the width of the well .
this potential models a qho placed in the center of the rectangular infinite well @xcite .
the potential @xmath41 fulfills two asymptotic requirements : 1 ) @xmath42 when @xmath43 ( free harmonic oscillator limit ) .
2 ) @xmath41 at equilibrium position have the same curvature as a free qho , @xmath44_{x=0}=k$ ] .
+ now we consider the following equation @xmath45\psi(x)=0\;.\ ] ] to solve analytically this equation , we use the factorization method @xcite . by changing the variable and some mathematical manipulation , the corresponding energy eigenvalues are found as @xmath46 where @xmath47,and @xmath48 is the frequency of the qho .
the first term in the energy spectrum can be interpreted as the energy of a free particle in a well , the second term denotes the energy spectrum of the qho , and the last term shifts energy spectrum by a constant amount .
it is evident that if @xmath43 then @xmath49 and the energy spectrum ( [ eee ] ) reduces to the spectrum of the free qho .
as is clear from ( [ eee ] ) , different energy levels are not equally spaced , hence confining a free qho leads to deformation of its dynamical algebra , and we can interpret the parameter @xmath50 as the deformation parameter . in table ( [ malek ] ) the numerical results associated with the original potential are compared with the generated results from model potential . as is seen the results are in a good agreement when boundary size is of order of characteristic length of the harmonic oscillator . on the other hand , the numerical results given in ref .
@xcite are related to the original potential , confined qho in the one - dimensional infinite well .
this oscillator when approached to the boundaries of well suddenly becomes infinite , while the model potential is smooth and approach to infinity asymptotically .
therefore , the model potential ( [ pot ] ) is more appropriate for the physical systems will be considered later .
+ if we renormalize eq.([eee ] ) to energy quanta of the simple harmonic oscillator and introducing the new variables @xmath51 , @xmath52 , and @xmath53 then eq.([eee ] ) takes the following form @xmath54 by comparing this spectrum with the energy spectrum of an f - deformed oscillator ( [ energy ] ) , we find the corresponding deformation function as @xmath55 this function leads to spectrum eq.([eee ] ) .
furthermore , the ladder operators associated with the confined oscillator under consideration can be written in terms of the conventional ( non - deformed ) operators @xmath3 , @xmath4 as follows @xmath56 these two operators satisfy the following commutation relation @xmath12=\gamma(2\hat{n}+1)+\alpha.\ ] ] it is obvious that in the limiting case @xmath43 ( @xmath57,@xmath58 ) , the right hand side of the above commutation relation becomes independent of @xmath59 , and the deformed algebra reduces to a the conventional weyl - heisenberg algebra for a free qho . + classically , harmonic oscillator is a particle that attached to an ideal spring , and can oscillate with specific amplitude . when that particle be confined
, boundaries can affect particle s motion if the boundaries position be in a smaller distance in comparison with a characteristic length that particle oscillate in it .
this characteristic length for the qho is given by @xmath60 where @xmath61 , and if @xmath62 , then the presence of boundaries affect the behavior of qho , otherwise it behaves like a free qho .
therefore , one can interpret @xmath63 as a scale length where the deformation effects become relevant .
now , we focus our attention on the coherent states associated with the qho under consideration . as usual
, we define coherent states as the right - hand eigenstates of the deformed annihilation operator @xmath64 from ( [ nlcs ] ) we can obtain an explicit form of the state @xmath65 in a number state representation @xmath66!\sqrt{n!}}|n\rangle,\ ] ] where @xmath67 ^ 2n!}\right)^{-\frac{1}{2}}$ ] is the normalization factor , @xmath68 is a complex number , and the deformation function @xmath69 is given by eq.([f1f ] ) .
the ensemble of states @xmath65 labeled by the single complex number @xmath68 is called a set of coherent states if the following conditions are satisfied @xcite : * normalizability @xmath70 * continuity in the label @xmath68 @xmath71 * resolution of the identity @xmath72 where @xmath73 is a proper measure that ensures the completeness and the integration is restricted to the part of the complex plane where normalization converges . the first two conditions can be proved easily . for the third condition
, we choose the normalization constant as @xmath74 where @xmath75 is similar to the modified bessel function of the first kind of the order @xmath76 with the series expansion @xmath77 .
resolution of the identity of deformed coherent states can be written as @xmath78 now we introduce the new variable @xmath79 and the measure @xmath80 where @xmath81 is the modified bessel function of the second kind of the order @xmath82 , @xmath83 and @xmath84 . using the integral relation @xmath85 @xcite , we obtain @xmath86 we therefore conclude that the states @xmath65 qualify as coherent states in the sense described by the condition ( [ c1])-([mmm ] ) .
we now proceed to examine some nonclassical properties of the nonlinear coherent states @xmath65 . as an important quantity
, we consider the variance of the number operator @xmath59 .
since for the conventional ( non - deformed ) coherent states the variance of number operator is equal to its average , deviation from poissonian statistics can be measured with the mandel parameter @xcite @xmath87 this parameter vanishes for the poisson distribution , is positive for super - poissonian distribution ( photon bunching effect ) , and is negative for a sub - poissonian distribution ( photon antibunchig effect ) .
+ figure [ f1 ] shows the size dependence of the mandel parameter for different values of dimensionless parameter@xmath88 .
as is seen , the mandel parameter exhibit sub - poissonian statistics and with further increasing values of @xmath89 it is finally stabilized at an asymptotical zero value corresponding to the poissonian statistics . + as another important nonclassical property we examine the quadrature squeezing . for this purpose
we first consider the conventional quadrature operators @xmath90 and @xmath91 defined in terms of undeformed operators @xmath3 and @xmath92 as @xmath93 the commutation relation for @xmath3 and @xmath4 leads to the following uncertainty relation @xmath94\rangle|^2=\frac{1}{16}.\ ] ] for the vacuum state @xmath95 , we have @xmath96 and hence @xmath97 . a given quantum state of the qho is said to be squeezed when the variance of one of the quadrature components @xmath90 and @xmath91 satisfies the relation @xmath98 the degree of quadrature squeezing can be measured by the squeezing parameter @xmath99 defined by @xmath100 then , the condition for squeezing in the quadrature component can be simply written as @xmath101 . in figure [ f2 ]
we have plotted the parameter @xmath99 corresponding to the squeezing of @xmath90 with respect to the phase angle @xmath102 for three different values of @xmath89 .
this diagram shows that the state @xmath65 exhibit squeezing for different values of the confinement size , and maximum value of squeezing occurs when @xmath103 .
figure [ f3 ] shows the plot of @xmath104 versus the dimensionless parameter @xmath88 for different values of phase .
as is seen , with the increasing value of @xmath88 quadrature squeezing is is stabilized to zero , according to mandel parameter .
+ let us also consider the deformed quadrature operators @xmath105 and @xmath106 defined in terms of the deformed operator @xmath7 and @xmath15 @xmath107 by considering the commutation relation for the deformed operators @xmath7 and @xmath15 ( [ comut ] ) , the squeezing condition for the deformed quadrature operators @xmath108 can be written as @xmath109 where @xmath110 or @xmath106 . figure [ f4 ] shows the plots of @xmath111 versus dimensionless parameter @xmath88 for three different values of @xmath112 .
as is seen , the deformed quadrature operator always exhibits squeezing .
[ malek ] @l*15@l state & boundary size & model potential & numerical results + 0 & a=0.5 & 4.98495312 & 4.95112932 + 0 & 1 & 1.41089325 & 1.29845983 + 0 & 2 & 0.67745392 & 0.53746120 + 0 & 3 & 0.57321464 & 0.50039108 + 0 & 4 & 0.54003728 & 0.50000049 + 1 & a=0.5 & 19.88966157 & 19.77453417 + 1 & 1 & 5.46638033 & 5.07558201 + 1 & 2 & 2.34078691 & 1.76481643 + 1 & 3 & 1.85672176 & 1.50608152 + 1 & 4 & 1.69721813 & 1.50001461 + 2 & a=0.5 & 44.66397441 & 44.45207382 + 2 & 1 & 11.98926850 & 11.25882578 + 2 & 2 & 4.62097017 & 3.39978824 + 2 & 3 & 3.41438455 & 2.54112725 + 2 & 4 & 3.00861155 & 2.50020117 + 3 & a=0.5 & 79.30789166 & 78.99692115 + 3 & 1 & 20.97955777 & 19.89969649 + 3 & 2 & 7.51800371 & 5.58463907 + 3 & 3 & 5.24620303 & 3.66421964 + 3 & 4 & 4.47421754 & 3.50169153
+ 4 & a=0.5 & 123.82141330 & 123.41071050 + 4 & 1 & 32.43724814 & 31.00525450 + 4 & 2 & 11.03188752 & 8.36887442 + 4 & 3 & 7.35217718 & 4.95418047 + 4 & 4 & 6.09403610 & 4.50964099 +
in this section , at first we introduce a mathematical structure on hilbert space developed recently @xcite .
we consider an abstract hilbert space @xmath113 .
let @xmath114 be an operator on this space with the properties : * @xmath114 is densely defined and closed ; we denote its domain by @xmath115 . * @xmath116 exists and is densely defined , with domain @xmath117 . * the vectors @xmath118 for all @xmath119 and there exist non - empty open sets @xmath120 and @xmath121 in @xmath122 such that @xmath123 and @xmath124 .
note that the first condition implies that the operator @xmath125 is self - adjoint ( here @xmath126 shows adjoint of operators ) . due to action of the operator @xmath114 , the hilbert space is transformed and orthogonal basis @xmath127
is transformed to a nonorthogonal basis
. this new basis can be considered orthogonal due to a new scalar product . + we define the two new hilbert spaces : * @xmath128 , which is the completion of the set @xmath115 in the scalar product @xmath129 the set @xmath130 is orthonormal in @xmath131 and the map @xmath132 extends to a unitary map between @xmath113 and @xmath128
. if both @xmath114 and @xmath116 are bounded , @xmath133 coincides with @xmath113 as a set .
* @xmath128 , which is the completion of @xmath134 in the scalar product @xmath135 the set @xmath136 is orthonormal in @xmath133 and the map @xmath137 extends to a unitary map between @xmath128 and @xmath133 . if the spectrum of @xmath138 is bounded away from zero then @xmath139 is bounded and one has the inclusions @xmath140 we shall refer to the spaces @xmath128 and @xmath133 as a dual pair and when ( [ sub ] ) is satisfied , the three spaces @xmath128 , @xmath113 and @xmath133 will be called a gelfand triple @xcite .
+ let @xmath141 be a ( densely defined ) operator on @xmath113 and @xmath142 its adjoint on this hilbert space .
assume that @xmath143 .
then unless @xmath144=0 $ ] , the adjoint of @xmath141 , considered as an operator on @xmath128 and which we denote by @xmath145 , is different from @xmath142 .
indeed , @xmath146 thus @xmath147 then due to the action of @xmath114 on hilbert space @xmath113 , we obtain other space @xmath128 . now
if we consider the oscillator operators @xmath3 , @xmath4 and @xmath5 , we have the following operators on @xmath128 @xmath148 clearly , considered as operators on @xmath128 , @xmath149 and @xmath150 are adjoints of each other and indeed they are just the unitary transforms on @xmath128 of the operators @xmath3 and @xmath4 on @xmath113 . on the other hand ,
if we take the operator @xmath149 , let it act on @xmath113 and look for its adjoint on @xmath113 under this action , we obtain by ( [ rrt ] ) the operator @xmath151 which , in general , is different from @xmath150 and also @xmath152\neq i$ ] , in general .
in an analogous manner , we shall define the corresponding operators @xmath153 , @xmath154 , etc , on @xmath155 . at this point
we must mention , according to this mathematical structure , operators @xmath149 and @xmath156 are exactly equivalent to generalized operators defined in ( [ defo ] ) that were adjoint of each other on the same hilbert space @xmath113 .
+ we use this mathematical structure to find proper representation for the problem under consideration and by a constraint we will determine operator @xmath114 . in previous sections
, we presented a description of the quantum harmonic oscillator confined in a one dimensional infinite well and we found its associated heisenberg - type algebra .
this algebra is a deformed heisenberg algebra which reduces to standard heisenberg algebra when the width of the well goes to infinity .
+ now using the hypothesis that successive energy levels of the qho confined in an infinite well are obtained by creation or annihilation of quantum particles in a box , we are going to construct a quantum field theory in a confined region and using it to quantize em field .
we use canonical field quantization approach .
the lagrangian associated with a given field confined within a certain region can be written as @xmath157 where @xmath158 defines the lagrangian of the free field and @xmath159 . if we constrained the problem to the confined region @xmath160 , the @xmath161 and we have @xmath162 .
this means that in the confined region we can use the lagrangian of the free field .
now if we impose quantization postulate , this postulate will be the same as free space .
+ for example , we consider the em field in a confined region and in this region we have the following lagrangian for the field @xmath163 where @xmath164 ( @xmath165 ) .
as is customary in quantization of the em field we use the four - vector potential as the dynamical variable of the field .
we use the coulomb gauge in which @xmath166 and @xmath167 . in this gauge ,
the hamiltonian of the em field is expressed in terms of the vector potential @xmath168 as @xcite @xmath169.\ ] ] we consider the vector potential @xmath170 as the field operator , and the quantization postulate for this field is expressed by the following commutation relation ( between @xmath170 and its conjugate momentum , @xmath171 ) @xmath172=-i\delta^3_{\perp ij}(r - r'),\ ] ] where @xmath173 is the transverse delta function .
now , we expand the field operator @xmath170 in terms of the ladder operators of the confined qho ( from here we show creation and annihilation operators of the confined qho by @xmath141 and @xmath142 ) @xmath174,\ ] ] where @xmath175 is the polarization vector of the em field , @xmath176 shows two independent polarization direction , and @xmath177 is the volume of confinement .
we interpret @xmath178 and @xmath179 , respectively , as the annihilation and creation operators for a deformed photon ( quantum excitation of the confined em field under consideration ) in direction @xmath180 , polarization @xmath176 and frequency @xmath181 .
the electric field operator or the conjugate momentum associated with @xmath170 is given by @xmath182.\ ] ] it is easy to show that @xmath183=&&\frac{-i}{2v}\sum_{\vec k,\lambda}\varepsilon_i(\vec k,\lambda)\varepsilon_j(\vec k,\lambda)\times\\ \nonumber & & [ u_k(\vec r)u_k^{\ast}(\vec{r'})+u_k^{\ast}(\vec r)u_k(\vec{r'})]h(\hat{n}_{k,\lambda}),\end{aligned}\ ] ] where @xmath184=h(\hat{n}_{k\lambda})=\gamma(2\hat{n}_{k\lambda}+1)+\alpha$ ] .
as is seen , in contrast to the quantization postulate ( [ quan ] ) , the right hand side of the above commutator is an operator - valued function .
if we use the deformed operators @xmath178 , @xmath179 as amplitudes of the field expansion , the quantization postulate imposed on the canonically conjugate variables of the em field is not preserved . to preserve the commutation relation ( [ quan ] ) , we propose using another pair of deformed operators in the fourier decomposition of the field operator . for this purpose ,
we consider the following dual operator of @xmath141 @xcite @xmath185 which satisfy the commutation relation @xmath186=\delta_{kk'}\delta_{\lambda\lambda'}.\ ] ] we use these operators to expand the field operator @xmath187.\ ] ] as is clear , the operators @xmath178 and @xmath188 are not adjoint of each other with respect to the ordinary scalar product , so the field operator is not hermitian .
it has been shown @xcite , there is a representation in which the operator @xmath189 is adjoint of the f - deformed operator @xmath141 with respect to a new scalar product in the carrier hilbert space .
hence , in order to preserve the quantization postulate , we should deform the hilbert space .
we show the ordinary scalar product by @xmath190 and the deformed one by @xmath191 .
since both scalar products are defined on the same hilbert space , they correspond to the same metric .
the relation between these two scalar product according to ( [ rrt ] ) can be written as @xmath192 where @xmath193 defines the relationship between two scalar products and it can be determined from the condition that @xmath141 and @xmath189 be adjoint of each other : @xmath194 therefore one can readily verify that @xmath193 is given by @xmath195 from eqs.([rrt ] ) and ( [ nsa ] ) operator @xmath114 can be found as @xmath196 and according to eq.([aaa ] ) the operators @xmath178 and @xmath179 can be obtained by the action of @xmath114 . now except other meaning of @xmath197 we can interpret it as a transformation , that by its action ordinary system can be changed to a confined system with definite barriers s position . + now instead of expanding the field operator in plane wave basis we expand it in a basis that is orthogonal with respect to the new scalar product ( [ scal ] ) @xmath198,\ ] ] where @xmath199 , is a basis that is orthogonal in the new representation as mentioned in mathematical preliminary section . in this new representation the field operator defined in eq.([field1 ] ) becomes hermitian .
furthermore , the electric field operator reads as @xmath200,\ ] ] and the quantization postulate is recovered @xmath201=-i\delta^3_{\perp ij}(r - r').\ ] ] as mentioned before , in the confined region the hamiltonian of the em field is the same as in free space .
this hamiltonian in the coulomb gauge is given by @xmath202 where @xmath141 refer to the magnetic field . by substituting the field operator @xmath170 given by ( [ field1 ] ) in the above expression
we arrive at the following hamiltonian @xmath203 thus , the hamiltonian can be interpreted as a collection of f - oscillators for different modes of the em field .
the eigensates of @xmath204 which form a complete set and span the hilbert space of the system , are given by @xmath205 where @xmath95 is the vacuum state of the system i.e. @xmath206 . in this manner
, we interpret each particle as an excitation of qho confined in an infinite well .
this formulation can be used in confined systems and nanostructures for considering elementary excitations , such as ecxitons ( which is a composite excitation ) , phonons and plasmons .
+ in quantum theory of fields , there are two important concepts that are very useful in considering interacting fields .
one of them is feynman propagator which is defined for a general field operator @xmath207 as @xcite @xmath208 where @xmath209 is the time - ordered operator ( we show the time ordering operator by @xmath209 for making distinction between this operator and the operator @xmath114 defined in ( [ hehehe ] ) ) .
now , if we assume that the field under consideration is spatially confined , then according to the definition of the deformed scalar product given by ( [ scal ] ) the corresponding feynman propagator is defined as @xmath210 making use of this definition for the photon field in a confined region and applying field operator ( [ field1 ] ) result in : @xmath211 where @xmath212 .
eq.([fey ] ) shows that the feynman propagator has not any difference in confined field theory except a constant factor that depends on some physical parameters such as the size of the system , and reduces to the standard propagator when the boundaries tend to infinity .
another important concept is the scattering matrix ( s matrix ) , that describes the probability amplitude for a process in which the system makes a transition from an initial state to a final state under the influence of an interaction . according to the concept of s matrix , the probability amplitude for a transition from the initial state @xmath213 into the final state @xmath214
is defined as @xmath215 where operator @xmath216 is defined in terms of the interaction hamiltonian in the interaction picture as @xcite @xmath217.\ ] ] in our formalism , according to the new definition of scalar product we define the probability amplitude as @xmath218 and due to the concept of fock states we have @xmath219 this equation shows that the new s matrix is proportional to the standard s matrix with a constant of proportionality that is a function of number of quantum excitations .
furthermore , we can conclude that spatial confinement of an interacting system results in an intensity - dependent coupling constant . as an example , consider an em field that interacts with a fermionic system in a confined region .
we assume that in this system , fermions be expressed by undeformed dirac field operator denoted by @xmath220 , and photons are described by ( [ field1 ] ) .
the interaction hamiltonian can be written as @xmath221 therefore the @xmath222 matrix is given by @xmath223+\cdots\\ \nonumber\end{aligned}\ ] ] where the symbol @xmath224 denotes the normal ordering .
so one can conclude from ( [ smat ] ) that coupling constant in each term of the expansion has a same dependence on the intensity of the photon field .
dependence of coupling constant on intensity is a indication of nonlinear interaction .
in this section we consider an infinite well directed in the z - direction , in which we have a current density ( @xmath225 ) .
for example , an electron that moving in axial direction of the well generates the following classical current @xmath226 where @xmath227 is the velocity of electron . in the presence of current density , the equation of motion for
the vector potential @xmath168 ( according to the maxwell equations ) reads as @xmath228 where @xmath229 is the transverse part of the current density .
this equation can be derived from the hamiltonian @xmath230 \nonumber \\ & = & \sum_{k,\lambda}\left(\omega_k\hat{b}^{\dag}_{fk\lambda}\hat{b}_{k\lambda}\right)+\frac{1}{c}\int d^3r\vec{j'}(r , t)\centerdot\hat{\vec{a}}(r , t ) \nonumber\\ & = & \sum_{k,\lambda}[\omega_k\hat{b}^{\dag}_{fk\lambda}\hat{b}_{k\lambda } \nonumber \\ & + & \frac{1}{c}\frac{\vec{\varepsilon}(k,\lambda ) } { \sqrt{2v \omega_k}}\int d^3r\left(\hat{b}_{k\lambda}g(k , r)+\hat{b}^{\dag}_{fk\lambda}g^{\ast}(k , r)\right)\centerdot\vec{j'}(r , t)],\end{aligned}\ ] ] the transverse density current @xmath231 and the polarization vectors are in the same plane and are in the same direction .
the hamiltonian ( [ ham ] ) can be rewritten as @xmath232.\ ] ] where @xmath233 .
the equation of motion for @xmath234 that follows from the above hamitonian reads as @xmath235 if we define a new variable @xmath236 , the solution of eq.([equm ] ) is @xmath237 the time dependence of the operator @xmath238 can be regarded as a result of the following unitary transformation @xmath239.\end{aligned}\ ] ] where by definition @xmath240 .
the operator @xmath241 is a displacement - like operator @xcite .
if we choose the initial state of the em field to be the vacuum @xmath95 , then the state vector at time @xmath242 is @xmath243|0\rangle\nonumber\\ & = & e^{-\frac{|\beta(k , t)|^2}{2}}\sum_n \frac{\beta^n(k , t)}{[f(n)]!n!}(\hat{a}^{\dag})^n|0\rangle \nonumber \\ & = & e^{-\frac{|\beta(k , t)|^2}{2}}\sum_n\frac{\beta^n(k , t)}{\sqrt{n!}[f(n)]!}|n\rangle.\end{aligned}\ ] ] in the sense of eqs.([coh])-([coh1 ] ) it is evident that this state can be regarded as a nonlinear coherent state .
in this paper , we have considered the relation between the spatial confinement effects and special kind of f - deformed algebra .
we have found that the confined simple harmonic oscillator can be interpreted as an f - oscillator , and we have obtained the corresponding deformation function .
then we have searched the effects of boundary conditions in quantum field theory .
we have used f - deformed operators as the dynamical variables and found that for preserving commutation relation between the field operator and its conjugate momentum we should deform hilbert space of the system under consideration . as a result of new definition of scalar product ,
we have concluded that the coupling constant of interactions in confined systems become a function of number of excitation , for example in the case of em field coupling constant becomes a function of intensity of em field .
finally we have proposed a theoretical scheme for generating nonlinear coherent states of em field through the coupling of a classical current to the vector potential operator @xmath170 inside a confined region .
d. k. ferry , s. m. goodnick , _ transport in nanostructures _ , ( cambridge university press 1999 ) .
+ b. l. altshuler , p. a. lee , r. a. web , _ mesoscopic phenomena in solids _ , ( north - holland , amsterdam 1991 ) .
g. plunien , b. mller , w. greiner , phys . rep . * 134 * , 87 ( 1986 ) .
h. haug , s. w. koch , _ quantum theory of the optical and electronic properties of semiconductors _
( world scientific , singapore , 2004 ) .
p. michler , _ single quantum dots : fundamental , application , and new concepts _ , topics in applied physics ( springer , 2003 ) .
p. michler , a. imamolu , m. d. mason , p. j. carson , g. f. strouse , s. k. buratto , nature ( london)*406 * , 968 ( 2000 ) ; + p. michler , a. kiraz , c. becher , w. v. shoenfeld , p. m. petroff , l. zhang , e. hu , a. imamolu , science * 290 * , 2282 ( 2000 ) .
p. michler , a. kiraz , l. zhang , c. becher , e. hu , a. imamolu , appl . phys . lett . * 77 * , 184 ( 2000 )
. t. feldtmann , l. schneebeli , m. kira , s. w. koch , phys .
b * 73 * , 155319 ( 2006 ) ; + c. gies , j. wiersing , m. lorke , f. jahnke , phys . rev a * 75 * , 013803 ( 2007 ) ; + k. ahn , j. frstner , a. knorr , phys .
b * 71 * , 153309 ( 2005 ) .
w. greiner , j. reihardt , _ field quantization _
( springer - verlag 1996 ) .
+ steven weinberg , _ the quantum theory of fields vol.1 _ ( cambridge university press 1995 ) . v. i. manko , g. marmo , g. c. e. sudarshan , f. zaccaria , phys .
scr . * 55 * , 528 ( 1997 ) . v. i. manko , r. vilela mendes , j. phys . a : math .
gen . * 31 * , 6037 ( 1998 ) ; + v. i. manko , g. marmo , s. solimeno , f. zaccaria , int . j. mod .
phys . a * 8
* , 3577 ( 1993 ) ; + v. i. manko , g. marmo , s. solimeno , f. zaccaria , phys . lett .
a * 176 * , 173 ( 1993 ) . v.
i. manko , g. marmo , e. c. g. sudarshan , and f. zaccaria , in : n. m. atakishiev ( ed ) , proc .
iv wigner symp .
( guadalajara , mexico , july 1995 ) , world scientific , singapore , 1996 , p.421 ; s. mancini , phys . lett .
a * 233 * , 291 , 55 ( 1997 ) ; s. sivakumar , j. phys .
a : math . gen . * 33 * , 2289 ( 2000 ) ; b. roy , phys .
a * 249 * , 55 ( 1998 ) ; h. c. fu , r. sasaki , j. phys .
a : math . gen . * 29 * , 5637 ( 1996 ) ; r. roknizadeh , m. k. tavassoly , j. phys .
a : math . gen . * 37 * , 5649 ( 2004 ) ; m. h. naderi , m. soltanolkotabi , r. roknizadeh , j. phys .
a : math . gen .
* 37 * , 3225 ( 2004 ) ; a - s .
f. obada , g. m. abd al - kader , j. opt .
b : quantum semiclass . opt .
* 7 * , s635 ( 2005 ) .
r. l. dematos filho , w. vogel , phys .
a * 54 * , 4560 ( 1996 ) .
m. h. naderi , m. soltanolkotabi , r. roknizadeh , eur .
j. d * 32 * , 397 ( 2005 ) .
m. h. naderi , phd thesis , university of isfahan , iran ( unpublished 2004 ) .
a. mahdifar , r. roknizadeh , m. h. naderi , j. phys . a : math
. gen . * 39 * , 7003 ( 2006 ) .
v. b. bezzera , m. a. rego - monteiro , phys .
d * 70 * , 065018 ( 2004 ) .
e. m. f. curado , m. a. rego - monteiro , j. phys . a : math
. gen . * 34 * , 3253 ( 2001 ) .
narayana swamy , physica a * 353 * , 119 ( 2005 ) .
m. h. naderi , m. soltanolkotabi , r. roknizadeh , j. phys .
japan * 73 * , 2413 ( 2004 ) .
s. sivakumar , phys , lett .
a * 250 * , 257 ( 1998 ) ; v. i. manko , a. wnshe , quantum semiclass
* 9 * , 381 ( 1997 ) ; g. n. jones , j. haight , c. t. lee , quantum semiclass . opt . * 9 * , 411 ( 1997 ) ; v. v. dodonov , y. a. korennoy , v. i. manko , y. a. moukhin , quantum semiclass . opt .
* 8 * , 413 ( 1996 ) ; nai - li liu , zhi - hu sun , hong - yi fan , j. phys .
a : math . gen .
* 33 * , 1993 ( 2000);b .
roy , p. roy , j. opt .
b : quantum semiclass . opt .
* 1 * , 341 ( 1999 ) . g. campoy , n. aquino , v. d. granados , j. phys . a : math . gen . * 35 * , 4903 ( 2002 ) ; + v. c. aguilera - navarros , e. ley koo , a. h. zimerman , j. phys .
a : math . gen . * 13 * , 3585 ( 1980 ) . c. zicovich - wilson , j. h. planelles , w. jackolski , int .
* 50 * , 429 ( 1994 ) . l. infeld , t. e. hull , rev .
* 23 * , 21 ( 1951 ) . j. r. klauder , b. skagerstam , _ coherent states : application in physics and mathematical physics _ ( world scientific , singapore , 1985 ) .
+ s. t. ali , j - p .
antoine , j - p .
gazeau , _ coherent states , wavalets , and their generalization ( new york : springer 2000 ) _ g. n. watson , _ theory of bessel functions _ , second ed .
( cambridge university press 1966 ) p.388 .
l. mandel , e. wolf , _ optical coherence and quantum optics _
( cambridge university press 1995 ) .
s. twareque ali , r. roknizadeh , m. k. tavassoly , j. phys . a : math
37 * , 4407 ( 2004 ) .
m. i. gelfand , n. ya vilenkin , _ generalized functions _ vol .
4 ( new york : academic press 1964 ) . w. hitler , _ the quantum theory of radiation _ , third edition ( clarendo press 1954 ) .
b. roy , phys .
a * 249 * , 25 ( 1998 ) ; + b. roy , p. roy , phys .
a * 263 * , 48 ( 1999 ) .
p. aniello , v. manko , g. marmo , s. solimeno , f. zaccaria , j. opt .
b : quantum semiclass . opt .
* 2 * , 718 ( 2000 ) . | we study some basic quantum confinement effects through investigation a deformed harmonic oscillator algebra .
we show that spatial confinement effects on a quantum harmonic oscillator can be represented by a deformation function within the framework of nonlinear coherent states theory . using the deformed algebra
, we construct a quantum field theory in confined space .
in particular , we find that the confinement influences on some physical properties of the electromagnetic field and it gives rise to nonlinear interaction .
furthermore , we propose a physical scheme to generate the nonlinear coherent states associated with the electromagnetic field in a confined region . |
a small class of active galactic nuclei ( agn ) emit broad , double - peaked , low - ionization lines , which most likely originate directly from the agn accretion disk .
the frequency of double - peaked lines is high ( 20% ) among the radio - loud equivalents of seyfert galaxies , broad - line radio galaxies ( blrg ; * ? ? ?
* ; * ? ? ?
* ) , and much lower ( 3% ) among the general population of low - redshift , optically selected agn ( * ? ? ?
* hereafter s03 ) .
the class of double - peaked emitters is diverse it includes both radio - loud ( rl ) and radio - quiet ( rq ) agn with luminosities ranging from low - luminosity low - ionization nuclear emission - line region ( liner ) galaxies to low - luminosity quasars .
the optical spectra of double - peaked emitters are sometimes dominated by the featureless continuum of the agn , but in a significant fraction of cases starlight from the host galaxy makes a dominant contribution to the observed optical continuum .
the handful of double - peaked emission - line blrg with reliable black - hole mass measurements ( 10 black hole masses measured via the triplet velocity dispersion and the relation between stellar velocity dispersion and black - hole mass ; * ? ? ?
* ) have @xmath0 and range from low - to - moderate eddington - ratio agn ( @xmath1 ) .
a large fraction of the originally studied double - peaked emitters were associated with low optical luminosity agn . the sloan digital sky survey ( sdss , * ? ? ?
* ) sample , which increased the number of medium - luminosity agn ( see figure [ luvz ] ) , also showed that the optical luminosity distributions of double - peaked emitters and normal agn at @xmath2 are not substantially different .
this suggests that the apparent lack of double - peaked emitters among the most optically luminous agn is a selection effect related to the rarity of very luminous sources and the small volume of space considered to date .
the large range of eddington ratios observed in agn with double - peaked emission lines together with their large range of luminosities ( which probably translates into a large range of eddington ratios ) suggest that the existence of strong balmer - line emission from the accretion disks of agn is either independent of the eddington ratio of the source , or that there are different explanations for the presence of double - peaked balmer lines in agn with different eddington ratios .
in addition to a large range of eddington ratios , the known double - peaked emitters tend to have broader balmer lines ( @xmath3 the full width at half maximum of normal agn ) and stronger low - ionization narrow lines ( e.g. , [ ] 6300 ) than similar agn without double - peaked lines . as a class ,
the double - peaked emitters possess a diverse set of optical / uv properties but most do not appear remarkably different in their other properties from the general agn population .
based on the prototype double - peaked emitter , arp102b , and the initial sample of double - peaked blrg and liner galaxies , there were reasons to believe that disk emission in agn is associated with low - luminosity , low eddington - ratio accretion in which the inner parts of the accretion disk ( less than a few tens of gravitational radii , @xmath4 ) became geometrically thick and optically thin ( e.g. , * ? ? ?
* ) , emitting energetic x - rays which illuminate the outer ( outside a few hundred gravitational radii ) regions of the disk @xcite .
since most of the initially known double - peaked emitters had strong balmer lines , which could not have been produced by the release of gravitational energy locally without invoking an unrealistic radiative efficiency , the x - ray illumination from the thick inner disk was necessary to produce strong disk - emission signatures .
the theoretical model presented above remains the best explanation for low - ionization broad - line disk emission in low - luminosity , low eddington - ratio agn . in its original form ,
however , it is not applicable to higher eddington - ratio ( and luminosity ) double - peaked emitters . as the eddington ratio and luminosity of the agn increases , the structure of the inner disk and the illumination pattern of the outer disk change .
in addition , line - driven winds could become important , filling - in the low - velocity minima of the double - peaked emission lines until the profiles are effectively single peaked ( e.g. , * ? ? ? * ; * ? ? ? * ) .
if we can establish clearly the need for external illumination in higher luminosity ( and presumably eddington ratio ) sources , we can revise the theoretical picture of their central regions by exploring evidence for the presence of a different kind of disk - illuminating structure for example a `` flared '' disk , a scattering corona , vertically extended wind , or a radiation supported torus . in search of clues that might reveal the mechanism responsible for the existence of double - peaked balmer lines in agn
, we decided to study in more detail the 0.510kev x - ray properties of this class .
if additional illumination is necessary to produce significant disk line - emission signatures in agn , their high - energy emission and their overall spectral energy distributions might be expected to differ from those of single - peaked agn .
s03 and @xcite calculated the uv - to - x - ray slopes ( @xmath5@xmath6 ) of double - peaked emitters with _
rosat _ all sky survey ( rass , * ? ? ?
* ) detections and concluded that they do not differ strongly from those of other agn .
these rough comparisons , however , included only x - ray detected subsamples of double - peaked emitters , which are not representative of the full double - peaked sloan digital sky survey sample and the general double - peaked line population .
they also did not take into account the dependence of @xmath5@xmath6 on agn luminosity ( e.g. , * ? ? ?
* ; * ? ? ?
in this paper we present a detailed investigation of the x - ray properties ( soft x - ray luminosities , uv - to - x - ray slopes , x - ray power - law photon indices , and x - ray absorption whenever available ) of a sample of double - peaked emitters serendipitously included in pointed _ rosat _ , _ xmm - newton _ , and _ chandra _ observations .
the paper is organized as follows : the sample of double - peaked emitters with x - ray observations is presented in [ sample ] , followed by the study of their uv - to - x - ray slopes and x - ray spectral shapes in [ sec3 ] .
[ ebudget ] discusses the evidence for external illumination , followed by the summary and conclusions in [ sec4 ] .
in the appendix we comment on the variability of mrk926 , a double - peaked emitter from our sample with several optical and x - ray observations . throughout this work
we use the _ wilkinson microwave anisotropy probe _ cosmology parameters from @xcite to compute the luminosities of agn : @xmath7 , flat cosmology , with @xmath8=72kms@xmath9mpc@xmath9 .
our sample consists of 29 double - peaked emitters selected from the sdss and serendipitously observed or previously targeted by _
rosat _ , _ xmm - newton _ , or _ chandra _ ( hereafter the main sample ) .
we add to this sample 10 blrg , liner galaxies , and rl quasars with good uv and x - ray observations ( hereafter the auxiliary sample ) .
a total of 16 of the 39 double - peaked emitters are rl ( 6 main - sample objects and 10 auxiliary sample objects ) , and the remaining 23 ( main - sample objects ) are rq . in order to increase the number of sdss double - peaked emitters with sensitive x - ray observations , we start with a sample of sdss agn with @xmath10 ( which guarantees that the h@xmath5 line is observed ) selected from data release 3 ( dr3 , * ? ? ? * ) which fall in pointed _ rosat _ , _ xmm - newton _ , or _
observations that are publicly available .
the effective exposure times ( i.e. , after accounting for the off - axis angle , x - ray background flares , etc . ) of these observations are all @xmath11ks and are @xmath12ks for 80% of the objects ; the median exposure time is 4.2ks . using sdss - rass matches would have substantially increased our sample size but decreased our detection fraction and the total x - ray counts of detected objects ( limiting our ability to extract x - ray spectral information ) . using only the s03 double - peaked emitter sample , which was selected from sdss observations performed prior to march 2003 ,
would have decreased the number of sources , thus weakening the significance of our statistical inferences . since the double - peaked emitters are a subsample of the asymmetric broad - line agn ( see s03 for more details ) , we start by selecting a subsample of the sdss agn with _ rosat _ , _ xmm - newton _ , or _ chandra _ observations and asymmetric broad h@xmath5 emission lines .
we fit the h@xmath5 line region of each spectrum with a combination of gaussians .
these gaussian representations of the h@xmath5 line region allow us to isolate the broad h@xmath5 line component ( excluding the narrow h@xmath5 , [ ] , and [ ] lines ) and to select the agn with double - peaked broad lines .
in addition , obtaining a smooth representation of the broad h@xmath5 line component allows easy measurement of a set of model - independent parameters : ( 1 ) the rest - frame full width at half maximum , fwhm , ( 2 ) the rest - frame full width at quarter maximum , fwqm , ( 3,4 ) the centroids of the line at half and quarter maximum , fwhmc and fwqmc ( measured with respect to the rest - frame narrow h@xmath5 line ) , ( 5,6 ) the rest wavelengths of the red and blue peaks with respect to the narrow h@xmath5 line , @xmath13@xmath14 and @xmath13@xmath15 , and ( 7 ) the peak separation , @xmath16 .
more details on the broad - line parameter measurements and the double - peak selection criteria can be found in 3.2 of s03 ; the h@xmath5 line measurements for the selected objects are reported in table [ tab7 ] .
eight of the resulting 29 main - sample double - peaked emitters are part of the s03 sample . for consistency with the rest of the sample
, we used the newest sdss spectroscopic reductions currently available and obtained new h@xmath5-line parameter measurements for these eight objects .
the results are consistent with those reported in the s03 paper , within the errors quoted in table 4 of s03 .
the h@xmath5-line luminosities of the eight new reductions were found to be slightly higher on the average ( by 20% ) .
ccccrrrrrr [email protected] & 0.3081 & 17 & 180 & 14600 & 10 & 11900 & @xmath18210 & 3200 & @xmath183400 + [email protected] & 0.1722 & 35 & 240 & 12900 & 1230 & 9570 & 1010 & 2300 & @xmath181010 + [email protected] & 0.3115 & 12 & 84 & 9800 & 90 & 7160 & 270 & 1860 & @xmath18240 + [email protected] & 0.2737 & 10 & 90 & 11800 & @xmath18190 & 8610 & @xmath18530 & 1350 & @xmath181970 + [email protected] & 0.1854 & 6.2 & 8.2 & 6350 & @xmath1820 & 4820 & @xmath1870 & 560 & @xmath18940 + [email protected] & 0.1744 & 1.5 & 87 & 17120 & @xmath183520 & 9270 & @xmath181320 & 130 & @xmath183590 + [email protected] & 0.1874 & 4.3 & 18 & 13800 & 690 & 9660 & 920 & 3180 & 1200 + [email protected] & 0.1704 & 31 & 37 & 11900 & 230 & 8250 & @xmath1840 & 2820 & @xmath18580 + [email protected] & 0.2126 & 11 & 18 & 42700 & 4380 & 36900 & 3020 & 20100 & @xmath1812400 + [email protected] & 0.0858 & 5.2 & 3.1 & 15400 & @xmath18980 & 9190 & @xmath18240 & @xmath18770 & @xmath186630 + [email protected] & 0.2481 & 4.5 & 25 & 15100 & 380 & 8520 & @xmath18500 & 6150 & @xmath182460 + [email protected] & 0.2416 & 4.9 & 22 & 7800 & 90 & 5760 & @xmath18100 & 640 & @xmath181000 + [email protected] & 0.3029 & 9.6 & 120 & 10600 & 1000 & 7340 & 1180 & 1020 & @xmath181290 + [email protected] & 0.2521 & 4.2 & 22 & 12000 & 420 & 9260 & 420 & 1960 & @xmath18620 + [email protected] & 0.2809 & 10 & 31 & 23800 & 710 & 18800 & 260 & 6670 & @xmath183980 + [email protected] & 0.1765 & 1.6 & 2.6 & 9850 & @xmath18190 & 6530 & @xmath18280 & @xmath18110 & @xmath181960 + [email protected] & 0.2310 & 3.0 & 29 & 10900 & 970 & 8310 & 670 & 3470 & @xmath181000 + [email protected] & 0.2624 & 7.1 & @xmath1910 & 9170 & @xmath18160 & 6830 & @xmath18160 & 1500 & @xmath181380 + [email protected] & 0.1095 & 0.41 & 3.1 & 8950 & 270 & 7080 & @xmath1840 & 320 & @xmath182100 + [email protected] & 0.2195 & 3.1 & @xmath197.3 & 11600 & @xmath18170 & 8890 & @xmath1850 & 1920 & @xmath181670 + [email protected] & 0.1158 & 0.62 & 21 & 9450 & @xmath1810 & 7350 & 330 & 2410 & @xmath18150 + [email protected] & 0.3231 & 8.0 & 110 & 10200 & @xmath18120 & 7990 & @xmath18170 & 1950 & @xmath181950 + [email protected] & 0.2261 & 1.6 & 5.8 & 7610 & 100 & 5720 & 460 & 2910 & @xmath1840 + [email protected] & 0.1412 & 3.2 & 73 & 9170 & @xmath1870 & 5830 & @xmath18470 & 3280 & @xmath18520 + [email protected] & 0.1500 & 12 & 6.2 & 23300 & 1150 & 19600 & 840 & 7400 & @xmath185490 + [email protected] & 0.2796 & 3.1 & 56 & 6790 & @xmath18400 & 4970 & @xmath18800 & @xmath1810 & @xmath182030 + [email protected] & 0.0946 & 3.0 & 8.0 & 7910 & @xmath18260 & 5800 & @xmath18450 & @xmath1890 & @xmath181760 + [email protected] & 0.1257 & 1.4 & @xmath192.2 & 10400 & 280 & 7970 & 30 & 2270 & @xmath181980 + [email protected] & 0.0469 & 4.8 & 85 & 11400 & @xmath181210 & 8790 & @xmath18970 & 1560 & @xmath181620 + [ tab7 ] of the 29 sdss agn with double - peaked lines , 22 were observed by the _ rosat _
pspc only , five by _ xmm - newton _ , and two by _ chandra _ , as shown in table [ tab1 ] . whenever hard - band ( above 2kev ) observations were available for double - peaked emitters in _ rosat _ pspc fields ( 3 cases with both _ xmm - newton _ and _ rosat _ data ) , we used the _ xmm - newton _ data . if both _
xmm - newton _ and _ chandra _ data were available , we chose the observation with the larger number of photons , and compare the results from the two observatories .
four sdss double - peaked emitters have been observed repeatedly with various x - ray satellites ( see the top portion of table [ tab1 ] ) ; we use all the available x - ray data for the best example , sdssj2304@xmath180841 , to study variability in appendix [ sec : app ] .
five of the x - ray observed double - peaked emitters have sufficient counts ( in the 210kev band ) to allow fitting of their x - ray spectra .
three of the objects observed by _
rosat _ are not detected in the soft ( ) band ; for these three we derive ( @xmath20 ) upper limits to the soft - band fluxes .
cccccrc [email protected] & _ rosat _ & rp800016n00 & pspc & 18/01/92 & 13.6 & + [email protected] & _ rosat _ & rp700249n00 & pspc & 01/04/91 & 6.5 & + [email protected] & _ rosat _ & rp700539n00 & pspc & 03/05/92 & 4.3 & target + [email protected] & _ rosat _ & rp200466n00 & pspc & 12/05/92 & 3.7 & + [email protected] & _ rosat _ & rp800481n00 & pspc & 03/11/93 & 1.1 & + [email protected] & _ rosat _ & rp701587n00 & pspc & 04/11/93 & 11.0 & target + [email protected] & _ rosat _ & rp700046n00 & pspc & 13/04/91 & 3.5 & + [email protected] & _ rosat _ & rp700150a02 & pspc & 20/10/93 & 1.9 & + [email protected] & _ rosat _ & rp800194n00 & pspc & 04/06/92 & 7.2 & + [email protected] & _ rosat _ & rf201357n00 & pspc & 29/11/92 & 2.1 & + [email protected] & _ rosat _ & rp800106n00 & pspc & 19/11/91 & 4.9 & + [email protected] & _ rosat _ & rp201242n00 & pspc & 23/06/92 & 5.8 & + [email protected] & _ rosat _ & rp200813n00 & pspc & 02/06/92 & 7.7 & + [email protected] & _ rosat _ & rp201367n00 & pspc & 07/07/92 & 3.3 & + [email protected] & _ rosat _ & rp700079a01 & pspc & 03/06/92 & 3.0 & + [email protected] & _ rosat _ & rp300394n00 & pspc & 10/11/93 & 12.0 & + [email protected] & _ rosat _ & rp700803n00 & pspc & 30/11/92 & 8.4 & + [email protected] & _ rosat _ & rp701408n00 & pspc & 13/07/93 & 6.8 & target + [email protected] & _ rosat _ & rp701589n00 & pspc & 18/08/93 & 10.1 & 3c332 , target + [email protected] & _ rosat _ & rp800503n00 & pspc & 31/07/93 & 5.2 & + [email protected] & _ rosat _ & rp201079n00 & pspc & 02/09/92 & 6.4 & + [email protected] & _ rosat _ & rp701252n00 & pspc & 29/05/93 & 18.6 & + [email protected] & _ xmm - newton _ & 0090070201 & pn / mos & 04/01/02 & 16.3 & um269 , target + ... & _ asca _ & 75020000 & gis / sis & 13/07/97 & 31.8 & um269 , target + ... & _ rosat _ & rp700377 & pspc & 30/12/91 & 5.1 & um269 + [email protected] & _ chandra _ & 00865n002 & acis - s7 & 28/07/00 & 4.7 & target , pileed - up + [email protected] & _ xmm - newton _ &
0084230601 & pn / mos & 26/04/01 & 16.0 & + ... & _ chandra _ & 00533n001 & acis - i1 & 05/09/00 & 11.3 & + [email protected] & _ chandra _ & 04035n001 & asis - s2 & 02/01/03 & 1.4 & target + [email protected] & _ xmm - newton _ & 0104861001 & pn / mos & 01/06/02 & 27.4 & + ... & _ xmm - newton _ & 0059750401 & pn / mos & 26/04/02 & 29.2@xmath21 & flaring background + ... & _ rosat _ & rp700297n00 & pspc & 15/05/92 & 3.1 & + [email protected] & _ xmm - newton _ & 0149900201 & pn / mos & 24/11/03 & 17.9 & + [email protected] & _ xmm - newton _ & 0109130701 & pn / mos & 01/12/00 & 7.3 & mcg-2 - 58 - 22 , mrk926 , target + ... & _ asca _ & 75049010 & gis / sis & 15/12/97 & 102.7 & mcg-2 - 58 - 22 , mrk926 , target + ... & _ asca _ & 75049000 & gis / sis & 01/06/97 & 104.9 & mcg-2 - 58 - 22 , mrk926 , target + ... & _ asca _ & 70004000 & gis / sis & 25/05/93 & 85.6 & mcg-2 - 58 - 22 , mrk926 , target + ... & _ rosat _ & rp701364 & pspc & 01/12/93 & 2.16 & mcg-2 - 58 - 22 , mrk926 + ... & _ rosat _ & rp700998 & pspc & 24/05/93 & 10.2 & mcg-2 - 58 - 22 , mrk926 + ... & _ rosat _ & rp701250 & pspc & 21/05/93 & 18.1 & mcg-2 - 58 - 22 , mrk926 + ... & _ rosat _ & rp700107 & pspc & 21/11/91 & 3.37 & mcg-2 - 58 - 22 , mrk926 + ... & _ rosat _ & rs931862 & pspc & 21/11/90 & 0.314 & mcg-2 - 58 - 22 , mrk926 + ... & _ rosat _ & rs931962 & pspc & 18/11/90 & 0.314 & mcg-2 - 58 - 22 , mrk926 + [ tab1 ] four of the objects observed with _
rosat _ were the targets of their respective pspc observations ( see the top panel of table [ tab1 ] ) .
four additional objects , for which x - ray spectra are available , were also the targets of their respective observations ( see the bottom panel of table [ tab1 ] ) .
a large fraction of targeted observations could have biased the sample of x - ray observed double - peaked emitters ;
[ representative ] below demonstrates that this is not the case for our sample . in this section
we show the basic properties of the main sample of double - peaked emitters studied here .
we demonstrate that the main sample is representative of the sample of sdss double - peaked emitters by performing a series of one- and two - dimensional statistical comparisons with the sdss sample from s03 ( hereafter , the s03 sample s03 double - peaked emitters from the quantitative comparisons presented here , since they lie outside the standard h@xmath5 selection region and were included in the s03 sample after being discovered serendipitously .
see s03 for more details . ] ) .
we use kolmogorov - smirnov ( k - s ) tests which measure the maximum distance , @xmath22 , between two cumulative distributions , and compute the null hypothesis probability that the data sets are drawn from the same distribution , @xmath23 .
small values of @xmath22 and large values of @xmath23 indicate that the observed data are highly probable given the null hypothesis ; we will consider this sufficient evidence that the x - ray observed agn are representative of the s03 sample .
for more details on the `` two - dimensional k - s test '' , see @xcite . using the starlight - subtracted monochromatic flux measurement at a rest - frame wavelength of 3700
( the bluest wavelength observed for all sdss agn , irrespective of their redshift ) , we can estimate the monochromatic flux at rest - frame 2500 , which is a convenient measure of brightness commonly used , for example , to compute the uv - to - x - ray index .
we assumed that the optical / uv flux is a power - law function of wavelength in the 25003700 region ( @xmath24 ) with an index of @xcite .
figure [ luvz ] shows the redshift vs. @xmath25 of the 2500 monochromatic luminosity ( @xmath26 $ ] ) diagram for the main sdss and auxiliary double - peaked samples in comparison to the s03 sample .
the correlation between redshift and luminosity in flux - limited samples is obvious for both the s03 and the main samples ; in addition , the double - peaked emitters from the main sample follow the luminosity - redshift trend of the s03 sample . as shown in the inset histograms in figure [ luvz ] ,
the monochromatic luminosities of the main and s03 samples are indistinguishable , which is confirmed by a one - dimensional k - s test ( @xmath27 , @xmath28% ) .
the redshift distribution of the main sample of double - peaked emitters is also similar to that of the s03 sample ( @xmath29 , @xmath30% ) .
the two - dimensional ( luminosity - redshift ) k - s test gives a similar result , with @xmath29 and @xmath31% .
these results remain qualitatively the same if we repeat the test for the rq subsamples .
there are six rl agn out of a total of 29 agn ( 21% ) in the main sdss double - peaked sample , compared to 15 out of 113 ( 13% ) in the similar - redshift ( @xmath10 ) s03 sample .
a fisher exact test confirms that the difference is not statistically significant , with a chance probability of 38% .
a careful examination of figure [ luvz ] reveals that five of the six rl objects in our main sample occupy the lowest redshift and uv luminosity corner of the diagram .
performing a two - dimensional k - s test for the rl subsamples of the main and s03 samples returns @xmath32 and @xmath33% .
the distance between the rl distributions is substantial , but not statistically significant .
figure [ gfrac ] shows the starlight fractions ( measured at rest - frame 3700 ) of the main - sample double - peaked emitters in comparison with those of the s03 sample . as a class
, rl double - peaked emitters were found to have larger starlight contributions to the continuum around h@xmath5 than other rl agn . a one - dimensional k - s test returns a small distance ( @xmath27 ) which is not statistically significant ( @xmath34% ) .
a two - dimensional ( starlight fraction - luminosity ) k - s test returns a larger distance ( @xmath35 ) which is also not statistically significant ( @xmath36% ) .
in addition to luminosity , redshift , and starlight - fraction distribution comparisons , we also compare the h@xmath5 line profiles of the main - sample agn with those of s03 , using the broad - line parameter measurements described in [ mains ] .
figure [ psepfwqm ] shows the distribution of peak separations vs. fwqm measurements for the main and s03 samples .
as described in 6 of s03 , the peak separation is largely determined by the outer radius of the disk emission region for axisymmetric disks , while the fwqm is determined by the disk inclination and inner radius of emission .
quantitative comparison of the two samples using a two - dimensional k - s test shows no evidence that the two samples are significantly different ( with @xmath37 and @xmath38% ) .
figure [ paramhist ] shows binned distributions of the fwqm and the fwqm centroid for the main and s03 samples of double - peaked emitters ( as in fig .
10 of s03 , the histograms of the parent sample of all agn with @xmath2 are also shown ) . in both cases
a one - dimensional k - s test suggests that the ( unbinned ) distributions of the main and s03 samples of double - peaked emitters are indistinguishable ( @xmath39 , @xmath40% ) .
lccccccc [email protected] & 25 & 2 & 100 & 9150 & 495 & 3.0e-4 & 5.5 + [email protected] & 36 & 3 & 190 & 2200 & 2000 & 6.5 & 3.2 + [email protected] & 32 & 2 & 100 & 5530 & 620 & 1.1e-4 & 1.6 + [email protected] & 16 & 3 & 210 & 2190 & 980 & 3.3 & 1.5 + [email protected] & 26 & 2 & 230 & 3450 & 900 & 6.7e-3 & 1.9 + [email protected] & 27 & 2 & 330 & 4260 & 760 & 6.6e-3 & 1.7 + [ tab2 ] we attempted to fit all the x - ray observed sdss agn with double - peaked lines with an axisymmetric disk - emission model @xcite .
six of the 29 agn ( 17% ) allow acceptable axisymmetric disk fits ( i.e. we reject fits with correlated residuals ) , two of which are shown in figure [ circfits ] .
the emission - region parameters for these six objects are given in table [ tab2 ] .
two double - peaked emitters , sdssj0918 + 5139 and sdssj1134 + 4912 , with @xmath41 in table [ tab2 ] , have poorly constrained inner radii ; the only evidence for a small inner radius in these agn comes from an extended blueshifted tail of the relatively narrow h@xmath5 lines , whose existence is strongly dependent on the continuum subtraction .
overall , the parameters listed in table [ tab2 ] cover the same ranges as those found by s03 and strateva et al .
( 2006 , in preparation ) for the subsample of s03 agn which allow axisymmetric disk fits .
the fraction of double - peaked emitters with h@xmath5 line profiles consistent with emission from an axisymmetric disk is also similar to that found for the s03 sample strateva et al .
( 2006 , in preparation ) .
we conclude that the objects in our main sample are representative of the s03 sample in their luminosity , redshift , host - starlight contributions , and broad line - profile parameter distributions .
in addition to the x - ray observed double - peaked sample presented above , we include for comparison purposes 10 rl agn two rl quasars , four blrg , and four liner galaxies previously studied in the x - ray band ( see table [ aux ] ) .
six of these 10 agn are part of a recently concluded program to obtain _ hubble space telescope _ ( hst )
uv spectra of nearby rl double - peaked emitters with high - quality x - ray , optical , and radio observations and a large range of x - ray luminosities ( eracleous et al .
2006 , in preparation ) .
the remaining four are arp102b , the prototype double - peaked emitter studied in the x - ray band by @xcite , and three low - luminosity liner galaxies with archival x - ray observations m81 , ngc4203 , and ngc4579 ( see table [ aux ] and associated references ) . an additional blrg with double - peaked lines from the eracleous et al . sample , 3c332 , was independently found in the sdss-_rosat _ sample , and is considered a main - sample object in this paper .
eight of these 10 double - peaked emitters have x - ray spectral fits , including estimates of x - ray power - law photon indices and absorbing column densities , as well as uv - to - x - ray slopes .
when multiple x - ray observations of an object were available , we used the best observations available ( longest exposures , coverage up to 10kev , lack of pile - up ) .
thus for ngc4203 we used the photon index from the spectral fit to the @xmath42ks _ asca _ observation reported in @xcite ( @xmath43 ) , and the x - ray fluxes and flux ratios obtained from the @xmath44ks _ chandra _ observation ( @xmath45 in the region , assuming no intrinsic absorption ) .
as noted by @xcite , the _ asca _ observation blends the true nuclear flux of ngc4203 with that of another object with the same spectral shape . our more accurate _
flux measurement in the band agrees to within 20% with the flux estimate of @xcite for the nuclear source in ngc4203 , suggesting that the object was in a similar flux state .
for ngc4579 we used a 2001 _ xmm - newton _ observation ( see also * ? ? ? * ) and confirmed that the results did not differ substantially from the 1995 _ asca _ observation which was obtained only 6 months after the uv data @xcite .
@xcite fit the full _ asca _ range with an absorbed power - law plus a raymond - smith model , with a photon index of @xmath46 , absorption consistent with the galactic column , and a 210kev flux of @xmath47erg@xmath48s@xmath9 . using the 2001 _
xmm - newton _ data we obtain a power - law fit with @xmath49 , @xmath50erg@xmath48s@xmath9 , and no intrinsic absorption .
ngc4579 was also observed by _ asca _ in 1998 @xcite and _ chandra _ in 2000 @xcite in a higher flux state : @xmath51erg@xmath48s@xmath9 ( 1998 ) and @xmath52erg@xmath48s@xmath9 ( 2000 ) , as well as slightly different photon index @xmath53 ( 1998 ) and @xmath54 ( 2000 ) . for m81 we used a 2001 _ xmm - newton _ pn observation ( obsid 0111800101 , 54.3ks ) and checked that the earliest 1993 _ asca _ observation @xcite , which was closer in time to the uv observation , gave similar results
@xmath55 and @xmath56erg@xmath48s@xmath9 ( _ asca _ ) vs. @xmath57 and @xmath58erg@xmath48s@xmath9 ( _ xmm - newton _ ; see table [ aux ] ) .
lccccccclcc ngc 1097 & 0.0043 & [email protected] & 0.041 & 0.00668 & 0.150 & @xmath181.22 & 25.763 & _ chandra _ & 1 & liner + pictor a & 0.0350 & [email protected] & 96 & 36.0 & 0.340 & @xmath180.80 & 27.960 & _ asca _ & 1 & blrg + b2 0742 + 31 & 0.4610 & 1.80 fixed & 3200 & ... & 0.640 & @xmath181.30 & 30.688 & _ rosat _ & 1 & rl quasar + pks 0921@xmath18213 & 0.0531 & [email protected] & 54 & 42.5 & 0.370 & @xmath181.04 & 28.370 & _ xmm _ & 1 & blrg + m 81 & 0.00086 & [email protected] & 0.039 & 0.0175 & 0.310 & @xmath180.91 & 24.686 & _ xmm _ & 2,3 & liner + ngc 4203 & 0.00362 & [email protected] & 0.0190 & 0.0330 & 0.036 & @xmath181.02 & 24.993 & _ chandra_,_asca _ & 3,4,5 & liner + ngc 4579 & 0.00507 & [email protected] & 0.26 & 0.20 & 0.059 & @xmath180.82 & 25.502 & _ xmm _ & 3,6 & liner + arp 102b & 0.0244 & [email protected] & 2.8 & 15.7 & 0.100 & @xmath181.02 & 27.108 & _ asca _ & 1 & blrg + pks 1739 + 184 & 0.1859 & [email protected] & 550 & ... & 0.790 & @xmath181.25 & 29.867 & _ rosat _ & 1 & rl quasar + 3c 390.3 & 0.0555 & [email protected] & 150 & 139 . &
0.330 & @xmath180.87 & 28.361 & _ asca _ & 1 & blrg + [ aux ] the uv analyses for the auxiliary objects are presented in eracleous et al .
( 2006 , in preparation , including the computation of @xmath60 ) , @xcite , @xcite , and @xcite .
the 3c332 monochromatic flux , measured at 3700 , was 40% brighter during the sdss observation than the earlier hst observation . since eracleous et al .
have a direct 2500 measurement , we use it instead of the value inferred by extrapolating the sdss spectrum .
we note that the auxiliary sample of double - peaked emitters is not representative of the sdss sample of double - peaked emitters they are all rl , at low redshift , and predominantly low optical / uv luminosity sources .
the presence of x - ray limits ( three main - sample double - peaked emitters and 40 steffen et al .
2006 rq agn ) in optically selected samples requires the use of proper statistical tools when making sample comparisons . in this section
we use the astronomy survival analysis package , asurv @xcite , to calculate sample means using the kaplan - meier ( k - m ) estimator , as well as a set of nonparametric sample - comparison tests for censored data the logrank test , gehan generalized wilcoxon test , and peto & prentice generalized wilcoxon test ( e.g. , iii of feigelson & nelson 1985 and references therein ) .
all three tests give better results if the sample sizes are similar and the censoring distributions are equal .
the peto & prentice test is less vulnerable to different censoring distributions than the logrank or gehan tests ( see the discussion in 5b of * ? ? ?
* ; * ? ? ?
@xcite develop a modified smirnov test for censored datasets , as the logrank and gehan tests are insensitive in the case of `` crossing - hazards alternatives '' ( i.e. , if the k - m estimator distribution functions of the different samples are not close to parallel but cross over instead ) ; this is relevant when comparing the rl double - peaked emitters with the @xcite sample discussed below .
to estimate the uv - to - x - ray spectral index , @xmath61 $ ] , we need a measure of the monochromatic flux at rest - frame 2kev .
the majority ( 22 ) of the objects in our sample have been observed only in the soft x - ray band ( ) , and their spectra have insufficient counts to study their spectral shapes in detail .
we first obtain counts and fluxes , and 2kev monochromatic fluxes ( assuming a power - law model with @xmath62 , modified by galactic absorption for all sources ) following the procedure described in @xcite and used by @xcite ; we list the results in table [ tab3 ] .
the use of a uniform power - law model to determine the rest - frame 2kev monochromatic luminosity and @xmath60 is necessary for the purpose of comparison with the @xcite and @xcite sources .
it is also preferable because of the large uncertainties of the hardness - ratio - estimated photon indices ( which are also inadequately constrained for 9 of the 22 rosat double - peaked emitters and potentially affected by intrinsic absorption as discussed below ) .
cccccccccccccr [email protected] & 0.3120 & 0.02240 & 232 & 2.91 & 2.78 & 26.10 & 29.45 & @xmath181.29 & @xmath63 & @xmath190.7 + [email protected] & 0.2740 & 0.03080 & 133 & 4.58 & 4.03 & 26.13 & 29.29 & @xmath181.21 & @xmath64 & @xmath190.6 + [email protected] & 0.1740 & 0.09070 & 379 & 1.51 & 10.79 & 26.11 & 28.45 & @xmath180.90 & @xmath65 & 4.0 + [email protected] & 0.1870 & 0.01460 & 38 & 3.76 & 1.86 & 25.42 & 28.57 & @xmath181.21 & @xmath19@xmath66 & @xmath190.6 + [email protected] & 0.2130 & 0.01120 & 11 & 3.10 & 1.40 & 25.42 & 29.36 & @xmath181.51 & @xmath19@xmath67 & @xmath190.4 + [email protected] & 0.0860 & 0.01420 & 152 & 3.01 & 1.77 & 24.66 & 28.39 & @xmath181.43 & @xmath68 & 3.2 + [email protected] & 0.2480 & 0.01200 & 26 & 0.98 & 1.40 & 25.57 & 28.97 & @xmath181.30 & @xmath69 & @xmath190.9 + [email protected] & 0.2420 & 0.01580 & 20 & 0.94 & 1.33 & 25.67 & 29.04 & @xmath181.29 & @xmath70 & @xmath190.7 + [email protected] & 0.3030 & 0.03340 & 142 & 4.44 & 4.34 & 26.26 & 29.29 & @xmath181.16 & @xmath71 & @xmath190.6 + [email protected] & 0.2520 & 0.01010 & 12 & 1.09 & 1.19 & 25.52 & 29.17 & @xmath181.40 & @xmath19@xmath72 & @xmath190.9 + [email protected] & 0.2310 & 0.01610 & 47 & 1.10 & 1.89 & 25.63 & 28.79 & @xmath181.21 & @xmath73 & @xmath190.7 + [email protected] & 0.2620 & 0.00420 & @xmath1916 & 2.28 & @xmath190.51 & @xmath1925.19 & 29.16 & @[email protected] & ... & @xmath191.1 + [email protected] & 0.1090 & 0.00870 & 44 & 2.17 & 1.06 & 24.67 & 27.59 & @xmath181.12 & @xmath19@xmath74 & @xmath190.7 + [email protected] & 0.2200 & 0.00440 & @xmath1911 & 2.32 & @xmath190.54 & @xmath1925.03 & 29.06 & @[email protected] & ... & @xmath190.6 + [email protected] & 0.1160 & 0.05370 & 86 & 1.67 & 6.43 & 25.50 & 27.94 &
@xmath180.94 & @xmath75 & @xmath190.4 + [email protected] & 0.3230 & 0.02700 & 241 & 1.46 & 3.21 & 26.20 & 29.40 & @xmath181.23 & @xmath76 & @xmath191.0 + [email protected] & 0.2260 & 0.00330 & 16 & 2.01 & 0.39 & 24.93 & 28.65 & @xmath181.43 & @xmath19@xmath77 & @xmath191.1 + [email protected] & 0.1410 & 0.12040 & 787 & 1.63 & 14.38 & 26.04 & 28.44 & @xmath180.92 & @xmath78 & 3.6 + [email protected] & 0.1510 & 0.00870 & 84 & 2.02 & 1.05 & 24.96 & 28.96 & @xmath181.53 & @xmath19@xmath79 & 3.3 + [email protected] & 0.2800 & 0.02010 & 60 & 1.32 & 2.38 & 25.92 & 28.74 & @xmath181.08 & @xmath80 & @xmath191.1 + [email protected] & 0.0950 & 0.03100 & 140 & 2.06 & 3.75 & 25.08 & 28.24 & @xmath181.21 & @xmath81 & @xmath190.3 + [email protected] & 0.1260 & 0.00420 & @xmath1916 & 4.83 & @xmath190.55 & @xmath1924.51 & 28.36 & @[email protected] & ... & @xmath190.9 + [ tab3 ] in addition to the above analysis , we also estimate the spectral shape of the _
rosat_-observed main - sample agn using the standard hardness ratio , hr1 , defined as @xmath82 , where @xmath83 and @xmath84 represent the total numbers of photons in the and bands , respectively .
assuming no intrinsic absorption above the galactic value , and spectra well represented by a power - law models , we can derive a one - to - one relationship between the power - law photon index @xmath85 and hr1 .
we used pimms to create a grid of expected hr1 values assuming an absorbed power - law model with different input photon indices and absorbing galactic column densities spanning the observed range ( @xmath86 ) .
we found cubic polynomial relations between @xmath85 and hr1 for each absorbing column and used these to estimate the photon index associated with each observed hr1 , @xmath87 pair . to estimate the photon - index uncertainties , we propagated uncertainties of the count rates in the and _ rosat _ bands using the method described in 1.7.3 of @xcite .
the results are given in table [ tab3 ] .
the three double - peaked emitters , which are undetected in the band , have no hr1 photon index estimates .
six of the remaining 19 double - peaked emitters detected by _
rosat _ were detected in the band only ; for these we can estimate only an hr1 lower limit and a corresponding upper limit on the power - law slope @xmath85 . any intrinsic absorption in excess of galactic will translate into an underestimate of the true value of @xmath85 .
this fact is of particular significance for the six double - peaked emitters without detections , since the upper - limit values of @xmath85 reported in table [ tab3 ] could be underestimated , rendering them useless for comparison . in [ sgamma ]
we point out that this could indeed be the case for a fraction of the agn in the current sample .
for the seven main - sample double - peaked emitters with hard - band coverage we were able to measure the monochromatic flux at 2kev ( rest - frame ) directly from the spectrum and report the corresponding 2kev monochromatic luminosities , together with their uv - to - x - ray indices in table [ tab6 ] .
these 2kev measurements might be underestimates of the true values if intrinsic absorption above @xmath88 ( assuming @xmath89 ) is present . in the analysis of individual objects below ,
we show that this is probably the case for only two of the _ xmm - newton _ or _ chandra _ observed objects and we correct the 2kev measurements accordingly .
ccccccccccccr 0043@xmath170051 & 0.3081 & 29.76 & 26.53 & @xmath181.24 & 9.3 & [email protected] & 210 & pl & 2.31 & ... & 1704 & 0.8 + 0057@xmath171446 & 0.1722 & 30.03 & 26.55 & @xmath181.33 & 38.0 & [email protected] & 110 & pl & 4.37 & ... & 403 & @xmath190.0 + 0918@xmath175139 & 0.1854 & 29.18 & 25.33 & @xmath181.48 & 2.9 & [email protected] & 0.411 & papl & 1.45 & @xmath90 & 243 & @xmath190.3 + ... & 0.1854 & 29.18 & 25.05 & @xmath181.58 & 2.0 & [email protected] & 0.311 & apl & 1.45 & @xmath91 & 145 & @xmath190.3 + 0938@xmath170057 & 0.1704 & 28.95 & 25.95 & @xmath181.15 & 18.2 & 1.5 fixed & 0.311 & apl & 4.22 & @xmath92 & 132 & @[email protected] + 1111@xmath174820 & 0.2809 & 29.49 & 25.73 & @xmath181.44 & 2.6 & [email protected] & 210 & pl & 1.24 & ... & 295 & 1.7 + 1134@xmath174912 & 0.1765 & 28.57 & 24.46 & @xmath181.58 & @xmath930.25 & 1.9 fixed & 0.210 & pl & 1.55 & ... & 24 & 0.8 + 2304@xmath180841 & 0.0469 & 28.59 & 26.19 & @xmath180.92 & 302 . &
[email protected] & 210 & plg & 3.60 & ... & 27200 & 1.0 + [ tab6 ] five of the double - peaked emitters
have sufficient counts to obtain absorbed power - law fits using xspec . for the remaining two objects we attempt to constrain the intrinsic absorption assuming a standard power - law photon index . in all cases
we include the galactic absorption in the model fits .
the results of the analysis for all seven double - peaked emitters with hard - band x - ray detections are presented below .
sdssj0043@xmath170051 ( @xmath94 ) has been observed by _ xmm - newton _ , _ asca _ , and _ rosat _ ( see table [ tab1 ] ) .
the _ xmm - newton _ spectral fits were published by @xcite ; we confirm their results : the observed spectrum can be represented by a power law with photon index @xmath95 and galactic absorption ( @xmath96 for 79@xmath97 , epic pn fit ) .
the spectrum and fit are shown in figure [ xspectra ] . when this model is extended to lower energies ,
a soft - excess is detected .
there is no evidence of fe k@xmath5 emission ; for a line centered at 6.4kev with a width of 10ev , we obtain an upper limit of 180ev for the equivalent width ( ew ) at 90% confidence ( page et al .
2004 estimate ew@xmath98kev at 90% confidence ) . for comparison ,
power - law fit for a 31ks observation obtained five years earlier ( as reported in the tartarus database at heasarc ) has @xmath99 , a 60% higher flux in the band , @xmath100erg@xmath48s@xmath9 , and no evidence for fe k@xmath5 emission .
the observed _ chandra _ acis - s count rate per frame for sdssj0057 + 1446 ( @xmath101 ) is @xmath102 counts s@xmath9 ( with frame - time of 0.441s ) and is high enough for photon pile - up to be important even for an off - axis angle of @xmath103 .
pile - up is the incidence of two or more x - ray photons in one ( or more ) neighboring ccd pixels within one frame time .
the ccd electronics may falsely regard these events as a single event with an amplitude given by the sum of the electron charge , resulting in a decrease of the apparent count rate of the source , an artificial hardening of its spectrum , and the apparent distortion of the point - spread function ( psf ) of point - like objects .
it can also alter the photon event grades and lead to a loss of events when standard grade filtering is applied to the data . to account for the effects of pile - up , we initially analyzed spectra of sdssj0057 + 1446 from annuli of various inner radii 0@xmath104 , 4@xmath104 , 1.0@xmath104 , 1.5@xmath104 , and 2@xmath104 and an outer radius of 3.4@xmath104 , excluding events from the core of the psf in annuli # 2 through # 5 , which are most affected by pile - up .
the effective - area files used when modeling the annular spectra were corrected for the energy dependence of the _ chandra _ psf .
we fit the annular spectra with simple power - law models modified by galactic absorption .
the resulting spectral slopes are @xmath105 ( annulus # 1 ) , @xmath106 ( # 2 ) , @xmath107 ( # 3 ) , and @xmath108 ( # 4 ) , with no need for absorption above galactic .
the last annulus ( # 5 ) has very few photons resulting in a poorly constrained spectral index , @xmath109 .
based on the spectral fit in annulus # 4 ( 1.53.4@xmath104 ) which is shown in figure [ xspectra ] , we estimate the pile - up corrected @xmath85 = @xmath110 . to estimate the fraction of events lost due to pile - up we used the forward spectral - fitting tool lynx that simulates pile - up ( chartas et al .
lynx simulates the propagation of individual photons through the _ chandra _ mirrors and acis , and it takes into account the possible overlap of events within one frame time .
we find that the pile - up fraction is @xmath111 20% in the band ( the pile - up corrected flux is @xmath112erg@xmath48s@xmath9 ) and is negligible in the band ( @xmath113erg@xmath48s@xmath9 ) .
sdssj0918@xmath175139 ( ) was serendipitously observed by both _
chandra _ and _ xmm - newton _ ( see table [ tab6 ] ) as a result of its proximity ( 6 arcmin ) to a nearby cluster , abell773 ( ) . the _ xmm - newton _ pn spectrum has 243 counts in the band .
a power - law fit with no intrinsic absorption above the galactic value provides a marginally acceptable fit in the band
@xmath114 for 15@xmath97 with .
the fit can be improved by adding intrinsic absorption
@xmath115 for 14@xmath97 , with and @xmath116@xmath48 , but neither of these fits is acceptable over the full _ xmm - newton _ band ( the best fit has @xmath117 , 21@xmath97 ) . a model with a partial absorber in addition to the galactic one , @xmath118@xmath48 , and
a covering fraction of @xmath119 provides an acceptable fit in the band : @xmath120 for 20@xmath97 .
spectrum has only 102 counts in the region and can be represented by a power law fit with no intrinsic absorption above the galactic value ( @xmath115 , 17@xmath97 ) .
the full _ chandra _ band ( , 145counts ) requires an intrinsic absorber ( @xmath114 vs. @xmath115 , 24@xmath97 ) @xmath121@xmath48 and has a slightly flatter spectral slope , .
the data suggests that the intrinsic absorber has changed in the seven months separating the _ xmm - newton _ and _ chandra _ observations . in the following sections ( [ saox ] and [ sgamma ] ) we use the _ xmm - newton _ results , which were obtained 4.6 months after the optical spectrum , have better signal - to - noise , and more accurate fit parameters .
the absorption corrected fluxes are consistent between the two observations within the ( 90% confidence ) errorbars : @xmath122erg@xmath48s@xmath9 ( _ chandra _ ) , and @xmath123erg@xmath48s@xmath9 ( _ xmm - newton _ pn ) . in the soft band , the observed fluxes are @xmath124erg@xmath48s@xmath9 ( _
chandra _ ) and @xmath125erg@xmath48s@xmath9 ( _ xmm - newton _ pn ) ; the intrinsic absorption correction could increase those by a factor of 3 ( _ chandra _ ) to 6 ( _ xmm - newton _ ) .
sdssj0938@xmath170057 ( @xmath126 ) has a 132 count detection ( ) in a 1.3ks _ chandra _ observation .
the spectrum is hard , implying high intrinsic obscuration and/or a flat power - law . if we fix the absorbing column to the galactic value , @xmath127@xmath48 , a simple power - law fit in the band ( 87counts ) returns @xmath128 . alternatively , assuming a standard power - law slope , @xmath129 for rq agn , we infer an intrinsic absorbing column of @xmath130@xmath48 .
neither of these fits is applicable over the full range : the required photon index drops to @xmath131 if we assume no intrinsic absorption , which is unrealistic considering the uniformity of the power - law slopes measured in rq agn ( @xmath132 , see , for example , figure 6 of * ? ? ?
* and references therein ) , and the best @xmath129 model including intrinsic absorption is inadequate ( @xmath133 , 24@xmath97 ) .
the full band spectrum can be represented by an intrinsically absorbed model with fixed @xmath134 and an intrinsic absorbing column of @xmath135@xmath48 ( @xmath136 , 24@xmath97 ) . for a simple intrinsically absorbed power law model , photon indices steeper than @xmath134
are excluded . if we allow an intrinsic absorber with partial coverage , the intrinsic absorbing column increases @xmath118@xmath48 , allowing a @xmath137 fit ( @xmath115 , 24@xmath97 ) .
higher signal - to - noise data is necessary to distinguish between the flat photon index , intrinsically absorbed model , and the steep photon index , partially absorbed model .
we adopt the simpler band @xmath134 model fit ( see table [ tab6 ] and figure [ xspectra ] ) and estimate the unabsorbed fluxes in the hard and soft bands : @xmath138erg@xmath48s@xmath9 and @xmath139erg@xmath48s@xmath9 .
sdssj1111@xmath174820 ( @xmath140 ) falls within the field of view of two _ xmm - newton _ observations , one of which ( obs . i d 0059750401 ) is heavily affected by background flaring .
we used the pn observation from june 2002 ( obs .
i d 0104861001 ) , which has a much longer effective exposure time , to extract and fit the spectrum of sdssj1111@xmath174820 .
the results are given in figure [ xspectra ] and in table [ tab6 ] .
the spectrum is well fit by a simple power law with galactic absorption with @xmath141 . if we extend this model to lower energies , the negative residuals below 2kev could indicate the presence of intrinsic absorption .
excess intrinsic absorption , however , is not required by the data : the spectrum also admits a power - law fit with a flatter photon index
@xmath142 and no intrinsic absorption ( @xmath143 for 48@xmath97 ) .
sdssj1134@xmath174912 ( @xmath144 ) is situated less than 10@xmath104 from another , optically fainter , point - like object in the sdss image ( [email protected] , with an @xmath145-band magnitude of 21.8 ) .
both this fainter object and the double - peaked emitter of interest to us are detected separately in the _ xmm - newton _
mos images with similar counts ; the larger pixel size of the pn ( 4.1@xmath104 for pn vs. 1.1@xmath104 for the mos detectors ) blends the two objects , with a centroid closer to the fainter mos object .
the _ rosat _ pspc ( psf fwhm of 25@xmath104 at 1kev and an observed off - axis angle of @xmath146 ) or hri ( with a 50% power radius of 9@xmath104 for an off - axis angle of @xmath147 ) lack the angular resolution to separate the two objects clearly and will not be considered here .
the measured 0.52kev flux of the double - peaked emitter is @xmath148erg@xmath48s@xmath9 for a 6@xmath104 aperture ( mos2 ) . using the epic mos on - axis psf ( which does not change much for objects at an off - axis angle of 7@xmath149 for energies
@xmath192kev ) , we estimate that the total flux of sdssj1134@xmath174912 is @xmath150erg@xmath48s@xmath9 . on account of the proximity to the second source , the small aperture used for the flux estimate , and the small number of total counts , the x - ray flux and 2kev monochromatic luminosity of this object
are uncertain . using the cash statistic
, we can obtain an acceptable fit to the mos spectrum using a power - law model with @xmath129 and an absorbing column density equal to the galactic value .
the high resolution of _ chandra _ is necessary to measure the x - ray properties of this object properly .
sdssj2304@xmath180841 is a nearby ( @xmath151 ) seyfert galaxy that has been observed repeatedly with most current and past x - ray observatories ( see appendix [ sec : app ] ) .
figure [ xspectra ] shows the spectrum in the 0.5220kev range , originally published by @xcite , overlaid with a model consisting of an exponentially cutoff power - law , reflection from an isotropically illuminated cold slab @xcite , and a gaussian fe - k@xmath5 line .
our 2220kev xspec fits confirm their results : the spectrum above 2kev is well represented by an absorbed and compton reflected power law with @xmath152 , absorption equal to the galactic column density , a 6.4kev fe k@xmath5 line with @xmath153kev and energy dispersion @xmath154ev ( corresponding to fwhm@xmath155kms@xmath9 , comparable to that of the h@xmath5 line , fwhm@xmath156kms@xmath9 ) , and a soft excess below 2kev . if we consider only the region ( see table [ tab6 ] ) , the power law fit with no intrinsic absorption and a gaussian fe k@xmath5 line has @xmath157 , consistent with the value obtained above in the 0.5220kev range .
three previous spectra observed with _
asca _ obtained in 1993 and 1997
@xcite show similar spectral shapes , with @xmath158 and evidence for some intrinsic absorption ( @xmath159@xmath48 ) in the two 1997 spectra .
judging from the relation between x - ray and optical variability , which is detailed in appendix [ sec : app ] , the presence of intrinsic absorption seems unrelated to the shape of the h@xmath5 line .
both the 1993 and one of the 1997 spectra show clear fe k@xmath5 lines , whose parameters depend on the continuum model @xcite ; the presence of a strong fe k@xmath5 line also appears unrelated to the single- or double - peaked shape of the optical balmer lines ( appendix [ sec : app ] ) .
the rest - frame uv and soft x - ray emission from agn are correlated , with more luminous agn emitting relatively less x - rays per unit uv luminosity ( e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) .
consequently a proper comparison of the x - ray emission of double - peaked emitters with those of normal agn should take this relation into account .
figure [ lxluv ] shows the 2500 vs. the 2kev monochromatic luminosities of the double - peaked emitters in comparison with those of rq agn from the @xcite sample .
the @xcite sample includes 333 optically selected rq agn with a high x - ray detection fraction ( 88% ) , and no evidence of intrinsic absorption in the uv . over half of the objects in the sample are sdss agn , directly comparable to the double - peaked emitters studied here .
the remaining objects allow us to extend the uv luminosity range of the comparison sample at each redshift .
the rq double - peaked emitters appear indistinguishable in figure [ lxluv ] from normal agn with similar uv monochromatic luminosities .
the majority of the rl double - peaked emitters show excess 2kev emission , as is characteristic of all rl agn ( e.g. , * ? ? ?
figure [ aox1 ] shows the uv - to - x - ray slope , @xmath5@xmath6 , for the double - peaked emitters in comparison with the sample of @xcite .
the majority of double - peaked agn follow the @xcite @xmath60-@xmath160 relation , @xmath161 and do not differ substantially from normal agn with comparable luminosity in the uv . due to the small range of luminosities probed at each redshift and the strong luminosity - redshift relation in the sample of double - peaked emitters , a partial - correlation analysis for the @xmath60-@xmath160 relation ( i.e. , one that takes into account the @xmath160-@xmath162 relation ) is inconclusive .
the strength of the @xmath60-@xmath160 partial correlation is only 2.3@xmath163 with partial kendall s @xmath164 , which could indicate either an inability of the partial - correlation analysis to detect a weaker @xmath60-@xmath160 anti - correlation on top of a strong @xmath160-@xmath162 correlation ( @xmath165 ) , or a genuine lack of an @xmath60-@xmath160 relation .
since we know the @xmath160-@xmath162 correlation is very strong for our sample of double - peaked emitters , we believe the former is more likely .
the top - left panel of figure [ aox2 ] presents the histogram of @xmath60 residuals for the rq main sample of double - peaked emitters , obtained by subtracting the expected @xmath166 from each observed @xmath60 , in comparison with that of the full sample and a luminosity - matched subsample ( @xmath167 ) from @xcite , shown in the middle left panel of figure [ aox2 ] . the comparison with the luminosity - matched subsample is useful , as there are some indications that the @xmath60-@xmath160 relation might be non - linear @xcite .
the kaplan - meier ( k - m ) estimator mean value of the @xmath168 residuals for the rq main sample is @xmath169 , consistent with zero , as is the equivalent k - m estimate , @xmath170 , for the luminosity - matched @xcite subsample .
both the gehan and the logrank tests ( @xmath171 , @xmath172% and @xmath173 , @xmath174% ) confirm that there is no significant difference in the cumulative distributions of the rq subsample of the main sample and the luminosity - matched @xcite subsample .
the conclusions remain unchanged if we include the full @xcite sample .
both @xmath60 residual distributions are consistent with being gaussian , with a width of @xmath175 for the rq main sample and @xmath176 for the luminosity - matched @xcite subsample and small positive means ( @xmath177 and @xmath178 , respectively ) .
the specific gaussian parameters ( especially the distribution means ) are slightly dependent on the size of the bins used and the fact that we ignore a small number of @xmath60 limits when performing the gaussian fits , but the fit parameters are generally within the 1@xmath163 errors quoted above . in all cases the rq double - peaked emitters tend to have a broader @xmath60 residual distribution than the corresponding luminosity - matched @xcite subsample distribution , but
the difference is at the 1@xmath163 level .
if this result is confirmed in larger samples , it could indicate larger x - ray and/or uv variability for the double - peaked emitters .
our current results indicate that the double - peak emitters as a class can not have dramatically different variability properties than normal active galaxies of similar optical / uv luminosity .
the left panel of figure [ aox2 ] shows that rl double - peaked emitters have a significantly different @xmath60 distribution than rq double - peaked emitters and the @xcite ( full and luminosity - matched ) samples .
because of the difference in censoring patterns between the @xcite sample and the sample of rl double - peaked emitters , the gehan and logrank tests , which assume no such difference exists , were not used in this comparison . a peto & prentice test confirms that the rl double - peaked emitters have @xmath60 residuals which are substantially different from those of normal rq agn
@xmath179 , @xmath180% . the stronger x - ray emission of rl double - peaked emitters in comparison to rq agn of similar uv luminosity is also clearly illustrated in figure [ aox1 ] , where eight of the 11 rl double - peaked emitters with @xmath181 ( where the comparison is strictly possible ) have @xmath60 values much flatter than typical rq agn .
a peto & prentice test confirms that the @xmath60 residuals of these 11 rl double - peaked emitters are significantly different than those of the rq agn in the @xcite sample ( @xmath182 , @xmath183% ) . the average @xmath60 residual for these 11 rl double - peaked emitters is 0.21 ; the median @xmath60 residual is 0.27 .
consequently , rl double - peaked emitters are on average @xmath184 more x - ray luminous than their rq counterparts , similar to the factor @xmath185 observed for rl agn as a whole ( e.g. , * ? ? ?
it is instructive to compare the positions of the higher and lower luminosity rl double - peaked emitters relative to the extrapolation of the @xcite relation from figure [ aox1 ] .
the higher luminosity rl objects tend to have flatter @xmath60 values than expected for rq objects with similar luminosities , while the liner galaxies appear to have steeper @xmath60 values .
this difference causes the apparently bimodal @xmath60-residual distribution for rl double - peaked emitters shown in in the bottom - left panel of figure [ aox2 ] . @xcite
have obtained accurate black - hole mass measurements and estimated the bolometric luminosities of eight of the rl double - peaked emitters included in our study .
the right panel of figure [ aox2 ] shows the @xmath60 residuals vs. the bolometric luminosities of these eight sources as a fraction of the eddington luminosity .
the error bars in the @xmath60 residuals include the uncertainty due to the intrinsic uv absorption correction and a constant 10% error in both the 2500 and 2kev monochromatic - luminosity measurements .
it ignores the possibly large but unknown uncertainties due to the non - simultaneity of the uv and x - ray observations . to estimate the uncertainty due to an intrinsic uv absorption correction from the inferred x - ray absorption ( or its upper limit ) we assume the @xcite extinction law and a galactic @xmath87/@xmath186 ratio .
the uncertainties due to the uv absorption corrections are typically less than a factor of 3 , except for arp 102b , where it is a factor of 20 .
the correlation between @xmath168 and @xmath187 is strong ( partial kendall s @xmath188 or @xmath189 ) but significant only at the 3.0@xmath163 or 3.2@xmath163 level when controlling for the dependence of both variables on luminosity or redshift .
this observation is consistent with our expectation of a different sed for agn with low - radiative - efficiency accretion modes in comparison to the standard thin - disk and corona models attributed to higher radiative - efficiency agn .
@xcite have used the fwhm and optical monochromatic - luminosity measurements of the s03 and @xcite samples to obtain rough estimates of the black - hole masses and accretion rates as a fraction of the eddington luminosity . even though the specific measurements of @xcite are very uncertain ( lewis & eracleous 2006 find that two of the four black - hole masses in common with their sample were overestimated by an order of magnitude ) , their general conclusions that double - peaked emitters have diverse accretion rates were confirmed by @xcite .
it is therefore encouraging that @xcite find a similar break ( at @xmath190 ) in the @xmath60 vs. @xmath187 relation ( see their figure 4 ) for all known rl double - peaked emitters .
considering the small number of objects and the large uncertainties in @xmath60 ( caused by both variability and measurement errors ) for low - luminosity agn , as well as the strong third - variable dependences , these findings need confirmation in larger samples .
figure [ gamma1 ] shows a comparison between the x - ray spectral slopes measured in the soft ( , panels a and b ) and hard ( , panels c and d , the hard - band fitting region was extended for five of the main - sample double - peaked emitters to include the soft band in order to increase the total photon counts to better model the intrinsic absorption . ] ) bands for the double - peaked emitters with those for similar broad - line agn which show no evidence of disk emission in the optical .
the comparison samples of normal agn were selected from laor et al .
( 1997 ; fwhm@xmath191@xmath192kms@xmath9 ) , walter & fink ( 1993 , fwhm@xmath191@xmath192kms@xmath9 and @xmath193 ) , brinkmann et al . ( 1997 ) , piconcelli et al .
( 2005 ; fwhm@xmath191@xmath192kms@xmath9 ) , the seyfert 1 galaxies from nandra et al . ( 1997 ) , and the blrg and rl quasars from @xcite . only broad - line agn ( fwhm@xmath191@xmath192kms@xmath9 ) were selected for this comparison , since some narrow - line seyfert 1s are known to have exceptionally steep x - ray spectral slopes ( e.g. , * ? ? ?
* ) , and their inclusion could bias the results . from figure [ gamma1 ] ,
the x - ray spectral shapes of double - peaked emitters appear similar to those of normal agn , except for the four rl double - peaked emitters in panel b. a one - dimensional k - s test confirms this statement for the hard - band power - law slopes shown in figure [ gamma1]d ( @xmath37 , @xmath194% ) .
there are only three hard - band photon index measurements for rq double - peaked emitters in figure [ gamma1]c ; consequently the k - s test results ( @xmath195 , @xmath196% ) might not be robust .
a simple student s @xmath197-test confirms that the means of the hard - band photon index distributions for the rq double - peaked emitters and normal agn are consistent , with @xmath198 and @xmath199% .
the presence of photon - index upper limits in panels a and b of figure [ gamma1 ] requires the use of the logrank , gehan , or peto & prentice comparison tests .
since the comparison samples do not have censored data , while the double - peaked emitters do , the peto & prentice test should give the cleanest results , as it is less vulnerable to different censoring distributions than the logrank or gehan tests @xcite . according to a peto & prentice test , the four rl double - peaked emitters in figure [ gamma1]b , with a k - m estimator mean of @xmath200 and a dispersion of @xmath201 ( compared to @xmath202 and @xmath203 for the normal agn ) , are significantly different from the normal agn ( @xmath204 , @xmath205% ) .
as in the case for 3c332 ( see discussion below ) , this could be an indication of the presence of intrinsic absorption .
alternatively , it may signify weaker soft excess , which is expected for the lowest eddington - ratio agn ( which might lack optically thick emission from the innermost parts of the accretion disk ) . according to a peto & prentice test ,
the soft - band photon indices of the 15 double - peaked emitters in figure [ gamma1]a are significantly different from the comparison agn ( @xmath206 , @xmath180% ) .
the evidence for this difference is however very weak , relying on the two double - peaked emitters with the flattest values of @xmath85 , both of which are limits .
if these two double - peaked emitters are removed from the comparison , the peto & prentice test gives a 8% probability that the remaining double - peaked emitters are indistinguishable from the normal agn .
the estimates of the spectral shapes in the _ rosat _ band , which are based on hardness ratios , depend sensitively on the assumption that the double - peaked emitters have no intrinsic absorption above the galactic values . as reported in
[ spgamma ] , we saw no evidence of intrinsic absorption in four of the seven main - sample double - peaked emitters .
the remaining three cases ( sdssj918@xmath175139 , sdssj0938@xmath170057 , and sdssj1111@xmath174820 ) show good evidence of intrinsic absorption , but the @xmath207 estimates are not robust , owing to the small number of counts available and the degeneracy between intrinsic absorption and spectral index in these cases .
the prototype double - peaked emitter arp102b , has a confirmed intrinsic absorbing column , equivalent to @xmath208@xmath48 @xcite .
in fact , significant intrinsic neutral absorbing columns are a general characteristic of nearby blrg ( see figure 6 of * ? ? ?
it is therefore plausible that the very hard power - law slope estimates obtained for some double - peaked emitters from the observed hardness ratios are a result of intrinsic absorption that was not taken into account .
for example , sdssj1617 + 3222 ( 3c332 ) , one of the six double - peaked emitters with power - law slope upper limits , has hr1@xmath209 , which , for a galactic obscuration of @xmath210 corresponds to @xmath211 .
the 3c332 _ rosat _ observation was previously studied by @xcite , who conclude that the hard spectral slope obtained by assuming galactic absorption is probably an indication of an intrinsic absorber .
assuming @xmath137 , @xcite derive an obscuring column of @xmath212 .
the uv - to - x - ray spectral - index of 3c332 is @xmath213 , while the value expected for a rl agn with comparable uv monochromatic luminosity is @xmath214 , indicative of weaker x - ray emission than expected and suggestive of intrinsic absorption .
there are a total of four double - peaked emitters in panels a and b of figure [ gamma1 ] with lower @xmath215 than expected , two rl agn ( 3c332 and sdssj0921 + 4538 ) with @xmath216 , and two rq agn ( sdssj1101 + 5122 and sdssj1150 + 0208 ) with @xmath215 upper limits ( @xmath217 ) . the two rq double - peaked emitters with low @xmath215 upper limits have @xmath60 values consistent within 0.05 with the @xcite relation ( indicating no intrinsic absorption ) , while sdssj0921 + 4538 has @xmath218 , with an expected value of @xmath219 ( indicating @xmath220 stronger x - ray emission than expected and no intrinsic absorption ) . taking into account the fact that three of the four double - peaked emitters with lower than expected @xmath215 upper limits have small numbers of counts and only -band detections , and that the observed scatter in the relation ( @xmath221 , * ? ? ?
* ) , we consider the above results , with the exception of those for 3c332 , to be inconclusive .
longer observations are necessary to measure accurately the power - law slopes and intrinsic absorption columns for double - peaked emitters with unusually low spectral index estimates in the 0.12kev band .
using the inner ( @xmath222 ) and outer ( @xmath223 ) radii of the h@xmath5 emission region we can estimate the amount of energy available locally in the disk due to viscous stresses and compare it to the luminosity of the h@xmath5 line .
an h@xmath5-line luminosity of 1020% or more of the locally available energy suggests the need for an external source of illumination . assuming a standard @xcite disk , @xcite and @xcite
estimate the gravitational power output of the line emitting disk annulus : @xmath224\textrm{erg\,s}^{-1 } \label{eqn2}\end{aligned}\ ] ] where @xmath225 is the luminosity in units of 10@xmath226ergs@xmath9 , and we assume for the bolometric luminosity , @xmath227 , @xmath228 and , for an accretion rate @xmath229 , @xmath230 , with the efficiency for conversion of energy into radiation @xmath231 .
the h@xmath5-line luminosities together with @xmath225 are listed in table [ tab7 ] .
since only six objects admit axisymmetric - disk profile fits , and only these allow unique estimates of the emitting region , we have no direct measurements of @xmath222 and @xmath223 for each agn in the main sample
. we do know that the main sample studied here is representative of the s03 sample and that most circular or elliptical double - peaked emitters have inner - emission radii of @xmath232 and outer - emission radii of @xmath233 .
sdssj0942 + 0900 @xcite , is an exception , with robust inner radius estimates @xmath234 .
based on both observational ( as noted above ) and theoretical ( the temperature of the disk increases for small @xmath235 and will likely ionize all hydrogen , preventing h@xmath5-line emission ) arguments , it is safe to state that @xmath236 for all known balmer - line double - peaked emitters . for the average double - peaked emitter studied by s03 , strateva et al .
( 2006 , in preparation ) and @xcite find @xmath237 and @xmath238 .
using the general sample constraints on @xmath222 and @xmath223 given above , we can now estimate @xmath239 from eqn .
[ eqn2 ] for the 23 double - peaked emitters without axisymmetric - disk fits .
figure [ diskpower ] shows the results assuming a typical double - peaked emitter with @xmath237 and @xmath238 for each of the 23 objects ( i.e. , @xmath240ergs@xmath9 ) .
only two of the 29 ( @xmath241% ) double - peaked emitters would generate enough power locally to produce h@xmath5 lines of the observed strength ( assuming that as much as 20% of the power were radiated in the h@xmath5 line ) . under the above assumptions for the extent of the emission region , the h@xmath5 line luminosity in fact exceeds the power produced locally in the disk in 16 of the 29 cases . even if all double - peaked emitters without axisymmetric - disk fits had @xmath242 ( and @xmath243 ) , @xmath244ergs@xmath9 ;
i.e. , at most 13 out of the 29 ( @xmath245% ) have enough energy generated locally to produce h@xmath5 lines of the observed strength .
we conclude that the majority of double - peaked emitters require external illumination of the disk to account for the strength of the observed h@xmath5 line emission .
we have studied the x - ray emission properties of 39 agn with double - peaked h@xmath5 lines , serendipitously observed by _ rosat _ , _ xmm - newton _ , and _ chandra _ ,
controlling for their uv and radio emission .
the main sample consists of 29 objects selected from the sdss which are representative of the sample of double - peaked emitters studied by s03 .
the remaining 10 objects ( the auxiliary sample ) are the best - studied double - peaked emitters ( eight low optical - luminosity blrg and liner galaxies , and two rl quasars ) with targeted and serendipitous observations .
the total sample includes 16 rl and 23 rq double - peaked emitters .
based on a comparison of the uv - to - x - ray index , @xmath60 , we find that rq double - peaked emitters as a class have comparable emission levels to those of normal rq agn of similar uv luminosity .
rl double - peaked emitters tend to be brighter in the band in comparison with similar - luminosity normal rq agn , as observed for rl agn in general .
the spectral shapes of the double - peaked emitters studied here are consistent with those of normal broad - line agn with similar radio - loudness , with the possible exception of four rl double - peaked emitters with unusually flat soft - band spectra .
three double - peaked emitters observed above 2kev show signs of large intrinsic absorption ( @xmath246 in two of the three cases ) .
the majority of the double - peaked emitters require external illumination of the accretion disk , as the power generated locally is insufficient to produce lines of the observed strength .
this result is more general and was used as an argument against the accretion - disk origin of broad emission lines in agn , irrespective of their profiles ( e.g. , * ? ? ?
for agn in which the balmer - lines are double - peaked , this external illumination could be associated with an elevated structure in the inner disk ( e.g. , an emitting corona , jet , or vertically extended torus ) .
however , the fact that the emission of double - peaked emitters as a class does not differ from that of normal agn with similar properties , suggests that a peculiarity of the emission structure and/or mechanism is not responsible for the occurrence of double - peaked balmer lines in agn . despite the importance of illumination for the majority of double - peaked emitters ,
the strength of the emission relative to the optical and uv emission can not be used to predict if the broad , low - ionization lines are double - peaked . by extension ,
the structure of the inner disk is unlikely to be the only ( or the dominant ) reason for the observed double - peaked profiles .
given the fact that double - peaked emitters as a class appear indistinguishable in their x - ray properties from normal agn , the best course for the future might be to study in detail extreme cases of double - peaked emitters .
our upcoming _
observations , for example , include one of the most luminous known double - peaked emitters an agn with a double - peaked line .
in addition , we hope to obtain high - quality hard - band x - ray observations of the broadest double - peaked h@xmath5-line emitters , for which the emission region of the disk is likely to be in the immediate vicinity or even coincident with the x - ray emitting structure .
0.5 in ivs , wnb and dps acknowledge the support of nasa ltsa grant nag5 - 13035 .
dps acknowledges the support of nsf grant ast03 - 07582 .
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 , cambridge university , 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 ( mpa ) , the max - planck - institute for astrophysics ( mpia ) , 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 .
this research has made use of the tartarus ( version 3.1 ) database , created by paul oneill and kirpal nandra at imperial college london , and jane turner at nasa / gsfc .
tartarus is supported by funding from pparc , and nasa grants nag5 - 7385 and nag5 - 7067 .
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 .
sdssj2304@xmath180841 is highly variable on long timescales ( 620 years ) .
@xcite cite evidence of secular changes in flux accompanied by strong , shorter lived ( @xmath11112years ) flares variability or their relation to the variability in the optical line profile . ] .
figure [ 926var ] shows the 210kev band variability ( following closely figure 1 of @xcite ) with arrows indicating whether the h@xmath5-line profiles were single or double - peaked .
the 2002 optical sdss observation shows complex h@xmath5 and h@xmath247 line profiles with at least two peaks and an extended red tail .
a 1997 observation by @xcite shows simple double - peaked profiles for both balmer lines , while a may 1993 observation of h@xmath247 by @xcite and a 1989 observation by @xcite show simple single - peaked profiles characteristic of normal agn .
the earliest known optical spectrum of sdssj2304@xmath180841 was taken in 1978 by @xcite ; judging by the spectrum in that paper the h@xmath247 line was single peaked .
the 1997 observation of a double - peaked h@xmath5 line coincides with one of the x - ray flares identified by @xcite , while the three observations of single - peaked h@xmath5 lines are in non - flare regions according to @xcite ; we have no x - ray data coincident with the sdss double - peaked profile ( the _ xmm - newton_/_bepposax _ data were taken one year before the sdss optical spectrum ) . from figure [ 926var ] it is tempting to conclude that the occurrence of disk emission in the optical is unrelated to flares in the x - ray flux .
this result , however , is highly uncertain , considering the sparse sampling time , the different x - ray bands observed , and the ( significant ) uncertainties of instrument cross - calibration .
what is obvious from figure [ 926var ] , is that the appearance of optical double - peaked lines in not predominantly associated with either low- or high - x - ray - flux states .
the x - ray spectral analysis presented in
[ spec1 ] further suggests that the appearance of optical double - peaked lines is not correlated with the steepness of the hard - band spectrum or the presence of a strong fe k@xmath5 line .
dum abazajian , k. , et al .
2005 , , 129 , 1755 argote , m. , dultzin - hyacyan , d. , de diego , j. a. , et al . 2001 , `` advanced lectures on the starburst - agn connection , '' eds . r. mujica , i. aretxaga , & d. knuthagn00/posters.html ] barth , a. j. , ho , l. c. , filippenko , a. v. , & sargent , w. l. w. 1998 , , 496 , 133 barth , a. j. , reichert , g. a. , filippenko , a. v. , ho , l. c. , shields , j. c. , mushotzky , r. f. , & puchnarewicz , e. m. 1996 , , 112 , 1829 bianchi , s. , matt , g. , balestra , i. , guainazzi , m. , & perola , g. c. 2004 , , 422 , 65 blandford , r. d. , & begelman , m. c. 1999 , , 303 , l1 brandt , w. n. , & boller , t. 1999 , asp conf .
ser . 175 : structure and kinematics of quasar broad line regions , 175 , 265 brinkmann , w. , yuan , w. , & siebert , j. 1997 , , 319 , 413 chen , k. , halpern , j. p. , & filippenko , a. v. 1989 , , 339 , 742 chen , k. , & halpern , j. p. 1989
, , 344 , 115 choi , c .- s . , dotani , t. , chang , h .- y . , &
yi , i. 2002 , journal of korean astronomical society , 35 , 1 collin - souffrin , s. , joly , m. , dumont , s. , & heidmann , n. 1980 , , 83 , 190 crawford , c. s. , & fabian , a. c. 1994 , , 266 , 669 dewangan , g. c. , griffiths , r. e. , di matteo , t. , & schurch , n. j. 2004 , , 607 , 788 durret , f. , & bergeron , j. 1988 , , 75 , 273 eracleous , m. , & halpern , j. p. 1994
, , 90 , 1 eracleous , m. , shields , j. c. , chartas , g. , & moran , e. c. 2002 , , 565 , 108 eracleous , m. , halpern , j. p. , & charlton , j. c. 2003 , , 582 , 633 eracleous , m. , & halpern , j. p. 2003
, , 599 , 886 feigelson , e. d. & nelson , p. i. 1985 , , 293 , 192 fleming , t. r. , ofallon , j. r. , obrien , p. c. & harrington , d. p. 1980 , biometrics , 36 , 607 ho , l. c. , filippenko , a. v. , & sargent , w. l. w. 1996 , , 462 , 183 ho , l. c. , et al .
2001 , , 549 , l51 isobe , t. , feigelson , e. d. , & nelson , p. i. 1986 , , 306 , 490 iyomoto , n. , makishima , k. , matsushita , k. , fukazawa , y. , tashiro , m. , & ohashi , t. 1998 , , 503 , 168 ishisaki , y. , et al .
1996 , , 48 , 237 kuiper , n. h. 1960 , proc . of the koninkl .
van wetenschappen , ser .
a. , 63 , 38 m. , & aldcroft , t. l. 2004 , , 150 , 165 laor , a. , fiore , f. , elvis , m. , wilkes , b. j. , & mcdowell , j. c. 1997 , , 477 , 93 latta , r. b. 1981 , j. am .
statistical association , 26 , 713 lavalley , m. , isobe , t. , & feigelson , e. 1992 , asp conf .
ser . 25 : astronomical data analysis software and systems i , 1 , 245 lewis , k. t. & eracleous , m. 2006 , , 642 , 711 lyons , l. 1991 , data analysis for physical science students ( cambridge : cambridge univ . press ) magdziarz , p. , & zdziarski , a. a. 1995 , , 273 , 837 marziani , p. , sulentic , j. w. , zamanov , et al .
2003 , , 145 , 199 murray , n. , & chiang , j. 1997 , , 474 , 91 nagao , t. , murayama , t. , shioya , y. , & taniguchi , y. 2002 , , 567 , 73 nandra , k. , george , i. m. , mushotzky , r. f. , et al .
1997 , , 477 , 602 narayan , r. , mahadevan , r. , & quataert , e. 1998 , theory of black hole accretion disks , 148 page , k. l. , reeves , j. n. , obrien , p. t. , et al .
2004 , , 353 , 133 page , m. j. , soria , r. , zane , s. , et al .
2004 , , 422 , 77 peacock , j. a. 1983 , , 202 , 615 piconcelli , e. , jimenez - bailn , e. , guainazzi , m. , schartel , n. , et al .
2005 , , 432 , 15 proga , d. , stone , j. m. , & kallman , t. r. 2000 , , 543 , 686 sambruna , r. m. , eracleous , m. , & mushotzky , r. f. 1999 , , 526 , 60 schneider , d. p. , et al .
2005 , , 130 , 367 seaton , m. j. 1979 , , 187 , 73p shakura , n. i. & sunyaev , r. a. 1973 , , 24 , 337 spergel , d. n. , et al .
2003 , , 148 , 175 steffen , a. , et al .
2006 , aj , submitted storchi - bergmann , t. , et al .
2003 , , 598 , 956 strateva , i. v. , et al .
2003 , , 126 , 1720 strateva , i. v. , brandt , w. n. , schneider , d. p. , et al . 2005 , , 130 , 387 sulentic , j. w. , marziani , p. , zwitter , t. , & calvani , m. 1995 , , 438 , l1 terashima , y. , kunieda , h. , misaki , k. , mushotzky , r. f. , ptak , a. f. , & reichert , g. a. 1998 , , 503 , 212 terashima , y. , ho , l. c. , ptak , a. f. , yaqoob , t. , kunieda , h. , misaki , k. , & serlemitsos , p. j. 2000 , , 535 , l79 terashima , y. , kunieda , h. , misaki , k. , mushotzky , r. f. , ptak , a. f. , & reichert , g. a. 1998 , , 503 , 212 york , d. g. , et al .
2000 , , 120 , 1579 vanden berk , d. e. , et al .
2001 , , 122 , 549 vanden berk , d. e. , et al .
2006 , , 131 , 84 vignali , c. , brandt , w. n. , schneider , d. p. , & kaspi , s. 2005 , , 129 , 2519 voges , w. 1993 , advances in space research , 13 , 391 walter , r. , & fink , h. h. 1993 , , 274 , 105 wang , t .- g . , dong , x .- b .
, zhang , x .-
, et al .
2005 , , 625 , l35 weaver , k. a. , nousek , j. , yaqoob , t. , hayashida , k. , & murakami , s. 1995 , , 451 , 147 weaver , k. a. , gelbord , j. , & yaqoob , t. 2001 , , 550 , 261 worrall , d. m. , tananbaum , h. , giommi , p. , & zamorani , g. 1987 , , 313 , 596 wu , x .- b .
, & liu , f. k. 2004 , , 614 , 91 | double - peaked balmer - line profiles originate in the accretion disks of a few percent of optically selected agn . the reasons behind the strong low - ionization line emission from the accretion disks of these objects is still uncertain . in this paper
, we characterize the x - ray properties of 39 double - peaked balmer line agn , 29 from the sloan digital sky survey and 10 low optical - luminosity double - peaked emitters from earlier radio - selected samples .
we find that the uv - to - x - ray slope of radio - quiet ( rq ) double - peaked emitters as a class does not differ substantially from that of normal rq agn with similar uv monochromatic luminosity .
the radio - loud ( rl ) double - peaked emitters , with the exception of liner galaxies , are more luminous in the x - rays than rq agn , as has been observed for other rl agn with single - peaked profiles
. the x - ray spectral shapes of double - peaked emitters , measured by their hardness ratios or power - law photon indices , are also largely consistent with those of normal agn of similar radio - loudness . in practically all cases studied here ,
external illumination of the accretion disk is necessary to produce the balmer - line emission , as the gravitational energy released locally in the disk by viscous stresses is insufficient to produce lines of the observed strength . in the appendix
we study the variability of mrk926 , a double - peaked emitter with several observations in the optical and x - ray bands . |
in this supplemental material to the letter `` scalable dissipative preparation of many - body entanglement '' we present the dynamical model of the generic system under consideration ( section [ secmodel ] ) , the coupling configurations for the described protocols ( section [ seccouplings ] ) and the resulting effective dynamics of the system ( section [ seceffective ] ) .
we discuss the engineering of the dissipative many - body interactions for ghz state preparation ( section [ secengineeredghz ] ) and strong driving effects ( section [ sectionstrongdrivingeffects ] ) . in section [ secanalysis ]
we analyze the scaling of the protocol both for weak and for strong driving .
the dynamics of the open system is modeled by a master equation of lindblad form @xmath142 + \sum_k l_k \rho l_k^\dagger - \frac{1}{2 } \left(l_k^\dagger l_k \rho + \rho l_k^\dagger l_k\right ) .
\label{eqmaster}\end{aligned}\ ] ] here , @xmath143 is the density matrix of the system and @xmath144 denotes a liouvillian of lindblad form .
the hamiltonian @xmath145 contains the coherent interactions , while sources of dissipation present in the system are represented as lindblad ( ` jump ' ) operators @xmath146 .
the hamiltonian for the coupling configurations detailed in fig .
2 b)-c ) is , in its most general form , given by @xmath147 we consider a free hamiltonian @xmath148 which contains the energies of the levels of the @xmath2 atoms and the harmonic oscillator modes , @xmath149 here we have introduced @xmath150 and made the simplifications @xmath151 and @xmath152 .
an interaction hamiltonian describes the atom - oscillator coupling , and a drive hamiltonian @xmath153 contains the fields used to perform coherent excitations of the system .
the interaction terms and drives required for ghz preparation are detailed in section [ seccouplings ] .
+ the excited degrees of freedom in the system are generally subject to dissipation . here
the excited states @xmath29 and @xmath30 of each atom undergo spontaneous emission to each of the ground states @xmath28 and @xmath3 , described by the jump operators @xmath154 where the subscript @xmath47 denotes the atom number .
the total decay rates of the excited levels are given by @xmath155 and @xmath156 .
for simplicity we will assume equal decay rates to both ground states , which is , however , not crucial for the protocol .
in addition , we assume decay of excitations of the two oscillator modes , @xmath36 and @xmath37 , represented by @xmath157 the scheme does not require oscillator decay at all ( @xmath158 ) . it may , however , still be useful to avoid heating .
the atom - oscillator couplings of the four coupling configurations in fig .
2 b)-c ) of the main text are described by the interaction hamiltonian @xmath159 here , ` z ' and ` x ' denote the coupling configurations introduced in the main text ( recall that @xmath160 .
these terms describe that an atomic excitation can be exchanged coherently with the respective harmonic oscillator with a coupling constant @xmath39 . by @xmath161 an atomic excitation in @xmath29
is exchanged with the oscillator @xmath36 , leaving the atom in @xmath3 .
@xmath162 couples the excited level @xmath30 to @xmath163 while exchanging the atomic excitation with the oscillator @xmath37 .
the coherent excitation of the atoms by classical driving fields is modeled by the drive hamiltonian @xmath164 here , we generally allow for several field tones with rabi frequencies @xmath165 and frequencies @xmath166 , from which we will deduce detunings @xmath167 of the respective fields .
in the following , we provide a detailed analysis of the effective dynamics of the system described in section [ secmodel ] . to this end , we briefly introduce the effective formalism presented in ref .
@xcite and use it to derive effective operators for the coupling configurations presented in section [ seccouplings ] .
as can be seen from section [ secmodel ] , the dissipation affects the excited levels @xmath29 and @xmath30 and the oscillator modes @xmath47 and @xmath36 . for weak driving the decaying degrees of freedom can be adiabatically eliminated from the master equation .
this is done using the effective operator formalism presented in ref .
@xcite . in this way
, the dynamics of the master equation are reduced to effective couplings between the ground states of the system , described by an effective master equation @xmath168 + \sum_{k } l_{k , \rm eff } \rho ( l_{k , \rm eff})^{\dagger } - \frac{1}{2}\left((l_{k , \rm eff})^{\dagger } l_{k , \rm eff } \rho + \rho ( l_{k , \rm eff})^{\dagger } l_{k , \rm eff}\right )
. \label{eqeffectivemaster}\end{aligned}\ ] ] since we are dealing with multiple field tones @xmath104 that give rise to the effective couplings , we use the extended formalism for many fields ( cf .
@xcite ) with @xmath169 here , @xmath170 denotes the exciting part and @xmath171 the deexciting part of the drive @xmath172 .
the non - hermitian hamiltonian @xmath173 , which contains the frequency @xmath174 , describes the time evolution of the excited states .
the drives are regarded as perturbations and thus denoted by `` @xmath175 '' ; they are defined by the drive hamiltonians @xmath153 in section [ seccouplings ] so that we use @xmath176 when we derive the effective operators for each of the coupling configurations below . based on the assumption of weak excitation
we will restrict our discussion to the states which have at most one atomic or oscillator excitation .
@xmath177 then contains the energies and couplings of the excited states .
since @xmath177 needs to be inverted to compute the effective operators , we will , for each coupling situation , start out by discussing this entity .
we begin with the z coupling configuration , which is similar to the x configuration .
both are required for the generation of ghz states .
we use @xmath178 , @xmath179 from eqs . and and the lindblad operators in eqs .
to set up the non - hermitian hamiltonian @xmath180 here we have introduced complex detunings @xmath181 and @xmath182 , where @xmath104 denotes the particular tone of the driving field .
furthermore we have changed into a frame rotating with the frequency of the drive @xmath183 . for the derivation of the effective operators we will for simplicity drop the sub- and superscripts denoting the coupling configurations and field tones . to invert the non - hermitian hamiltonian
we divide it into four blocks @xmath184 after this separation we can formally invert the hamiltonian @xmath177 , using banachiewicz theorem @xcite for the blockwise inversion of a square matrix , @xmath185 we now need to compute @xmath186 to obtain any of the above elements . as we shall see , it is possible to simplify the calculation and to obtain closed expressions for the decay rates if we separate the involved operators by the number of atoms in state @xmath3 .
this is done by introducing projection operators @xmath84 which project on the states with the same number @xmath55 ( from now on , @xmath187 ) of atoms in @xmath3 . for the z and x configurations discussed here ,
@xmath55 ( or @xmath22 for x ) , is conserved under the couplings by the coherent interactions @xmath188 and @xmath175 , but can be changed by the dissipative jump processes @xmath146 , e.g. from @xmath55 to @xmath189 in the case of z. using these projectors we can split the non - hermitian hamiltonian of the excited states and its four blocks by @xmath187 , @xmath190 the inverse of these non - hermitian hamiltonians for each @xmath187 are then found to be @xmath191 the effective operators of eqs .
are formally given by @xmath192 to obtain the effective lindblad operators and the effective hamiltonian it is thus sufficient to compute the blocks @xmath193 and @xmath194 of the inverse non - hermitian hamiltonian .
using the identities @xmath195 and @xmath196 ( where @xmath197 is the projector onto the ground states ) we obtain @xmath198 \\ b_{n } & = \frac{g}{\tilde{\delta } \tilde{\delta } } a^\dagger j_{1e } \left[j_{ee } - \left(n - \frac{\tilde{\delta } \tilde{\delta}}{g^2}\right)^{-1 } j_{1e}^\dagger j_{1e } \right]\end{aligned}\ ] ] with this , and readopting the sub- and superscripts for the configuration and the field , and changing to a more detailed notation for the effective lindblad operators we find @xmath199 to obtain this we have used the identities @xmath200 , @xmath201 , and @xmath202 . in the last expression we have also introduced the effective detunings @xmath203 and the effective couplings @xmath204 with @xmath205 as can be seen from eqs .
, dealing with multiple frequencies in the drive leads a priori to a sum over terms for all fields in the effective lindblad operators .
however , as the frequencies of these fields are well - distinguishable , we separate the lindblad operators by their driving field @xmath104 .
given the quadratic appearance of the lindblad operators in the master equation , we can also drop the exponential phase factors . for the effective lindblad operators for the fields @xmath104
we then obtain @xmath206 we also define the corresponding effective decay rates @xmath207 the operators in eqs .
are then the effective lindblad operators for the z configuration . as can be seen from the expressions in eqs .
, the effective detunings @xmath208 can be made very small by a suitable choice of the frequencies @xmath209 of the fields @xmath104 which can be used to engineer the rates @xmath210 and @xmath211 of the effective decay processes
. the engineering of the effective decay process to prepare ghz states will be subject to section [ secengineeredghz ] .
+ we now turn to the effective hamiltonian .
the effective hamiltonian is computed from eq . .
other than for the effective lindblad operators , introducing a multi - tone driving field results in cross terms between different fields , here denoted by @xmath104 and @xmath212 , @xmath213 here , all terms @xmath214 have fast rotating exponential phase factors .
restricting the treatment to @xmath215 where these terms cancel , we obtain the main contribution @xmath216 we thus find that the main effective hamiltonian processes are ac stark shifts with a magnitude @xmath217 as we will see further below , our choice of the field tones will make these hamiltonian terms compensate each other .
we perform an analogous treatment for the x pumping mediated by the excited level @xmath30 and the oscillator mode @xmath37 , using the couplings in eqs . and .
we start with the non - hermitian hamiltonian @xmath218 with the complex energies @xmath219 and @xmath220 . carrying out the derivation in the same manner as above for z , we obtain for the effective lindblad operators @xmath221 with the effective detunings @xmath222
the effective decay rates can be written as @xmath223 these operators resemble the ones for z pumping in eqs . if @xmath3 is replaced by @xmath23 .
the only difference is that spontaneous emission is still assumed to lead to the final states @xmath28 and @xmath3 . using the effective operator concept ,
we have so far reduced the dynamics of the open system to an effective master equation of the ground states .
this description is exact to second order perturbation theory .
in addition to that , we will later achieve a compensation of the effective hamiltonian , @xmath224 , so that the remaining dynamics are purely dissipative .
we can then , in another step , reduce the complexity of the dynamics to rate equations of the populations .
this is achieved by choosing subspaces of interest between which the interactions present in the system do not build up coherences .
these subspaces are defined by projection operators , e.g. @xmath225 and @xmath226 . for negligible coherences between the subspaces
, we can then trace over the liouvillian evolution to obtain decay rates from subspace @xmath227 to subspace @xmath228 @xmath229 for the subspaces we will consider , the decay rate is the same for all states in the subspace .
we can then calculate the decay rates using a single state @xmath230 , @xmath231 we will use this rate equation approach to analyze the scaling of the preparation time of the protocol in section [ secanalysis ]
. the evolution of the system is simulated by numerically solving the effective master equation , including the effective operators derived above . for the evolution due to the effective master equation of the protocol we use a trotter - like ansatz , simulating the evolution under the z and x coupling in an interchanging manner , performing base transformations between the eigenbases of @xmath232 and @xmath233 in - between . to verify our findings
, we compare our result to the solution of the master equation after truncation to one or two excitations .
in this section we show how the effective operators derived in the previous section are engineered to prepare ghz states efficiently .
as is discussed in the main manuscript , the engineered operators are used to empty certain states and transfer the population to others in such a way that in the end only the target state remains .
this is achieved using the driving fields in the coupling configurations z and x which activate the effective decay processes derived in section [ seceffective ] . by choosing suitable detunings for these driving fields we can then engineer the rates of the effective decay processes .
the choice of the rabi frequencies of these fields is subject to section [ secanalysis ] where we optimize the preparation time of the protocols .
the first mechanism to prepare a @xmath11 state described in the main manuscript is the z pumping : this process transfers the population of all states other than @xmath4 and @xmath10 to @xmath234 , and thereby , to @xmath11 , without affecting the ghz state .
we engineer this process by applying drives @xmath235 in the z configuration with detunings @xmath236 , where @xmath58 . here
, the subscript @xmath237 denotes the fields with positive detunings and @xmath238 those with negative detunings . if the field index @xmath104 coincides with the number of atoms in @xmath3 , @xmath55 , for a certain initial state and driving field , the field is resonant and the effective detunings in eqs .
become @xmath239 since we generally work in the strong coupling limit @xmath240 , the above effective detunings are small compared to those for off - resonant driving fields with @xmath241 , @xmath242 with the effective detunings we can then compute the effective decay rates in eq . for the processes with @xmath243 and for the off - resonant ones with @xmath244 , @xmath245 from these expressions we can clearly see that the first group of rates is engineered to be strong , while the second group of rates is engineered to be suppressed . using the quantities above we obtain for the effective lindblad operators in eqs . . @xmath246 p_{n_1 } + \sqrt{\kappa_{z \pm , n_1}^{(f \neq n ) } } p_{n } , ~ ~ ~ ( 1 \leq f \leq n -1 ) \\
l_{\gamma 0,a , z \pm}^{(f ) } & \approx \sum_{n_1 = 1}^{n-1 } \left[\sqrt{\gamma_{0 , z \pm , n_1}^{(f = n_1 ) } } { |0\rangle}_a { \langle1| } + \mathcal{o}\left(\frac{1}{g^2}\right)\right ] p_{n_1 } + \sqrt{\gamma_{0 , z \pm , n_1}^{(f \neq n ) } } { |0\rangle}_a { \langle1| } p_{n } , ~ ~ ~ ( 1 \leq f \leq n -1 ) \\ l_{\gamma 1,a , z \pm}^{(f ) } & \approx \sum_{n_1 = 1}^{n-1 } \left[\sqrt{\gamma_{1 , z \pm , n_1}^{(f = n_1 ) } } { |1\rangle}_a { \langle1| } + \mathcal{o}\left(\frac{1}{g^2}\right)\right ] p_{n_1 } + \sqrt{\gamma_{1 , z \pm , n_1}^{(f \neq n ) } } { |1\rangle}_a { \langle1| } p_{n } , ~ ~ ~ ( 1 \leq f \leq n -1)\end{aligned}\ ] ] from these expressions we see that for @xmath86 we will always find terms with @xmath247 to zeroth order in @xmath248 , which are much larger rates than the ones with @xmath241 , which are to second order in @xmath248 .
we can therefore drop the latter for @xmath86 .
the terms with @xmath249 which in particular affect the ghz state need , however , to be kept . since the effective detunings are engineered to depend on @xmath55 , photons scattered by resonances with different @xmath55 can be distinguished .
formally this is justified by the exponential factors @xmath250 , washing out interferences between terms ( cf .
section [ seceffective ] ) .
we can therefore separate the terms with different @xmath55 into individual lindblad operators , each acting on a set of states with @xmath55 atoms in @xmath3 .
the enhanced processes are then given by @xmath251 the weak decay processes affecting @xmath10 , and thus , @xmath11 are found to be @xmath252 the result of the engineering of the effective processes for the z configuration is thus the z pumping by which all population from states with @xmath86 is transferred to @xmath4 , and thus , to @xmath11 and @xmath7 . beside effective decay processes ,
we generally also obtain effective hamiltonian terms as given by eq . , which we write as @xmath253 with @xmath254 these terms are ac stark shifts , the magnitude of which is obtained from eq .
, @xmath255 by our combination of red - detuned drives ( @xmath256 , @xmath257 ) and blue - detuned drives ( @xmath258 , @xmath259 ) with @xmath260 and @xmath261 we achieve that the shifts compensate each other , @xmath262 given that there are no hamiltonian terms , we can turn to a description of the dynamics in terms of rate equations : in section [ secrate ] , we have described the possibility to reduce the effective dynamics further to rate equations . to perform such a step ,
we identify subspaces inside which all states have the same decay rate and in - between which no significant correlations are built up .
it is therefore important that the effective hamiltonian is zero .
we define the subspaces by the projectors @xmath263 , @xmath264 , and @xmath84 for @xmath86 , which contain the states with a certain number of atoms in @xmath3 .
the decay rates can then be calculated using eq . .
for resonant z pumping ( @xmath243 ) from a subspace with @xmath55 atoms in @xmath3 to one with @xmath12 due to spontaneous emission to @xmath28 we obtain the rate @xmath265 the subscripts in the above and the following decay rates specify the initial subspace , the final subspace , the physical process , i.e. oscillator decay ( @xmath266 ) , spontaneous emission to @xmath28 ( @xmath267 ) or @xmath3 ( @xmath268 ) , and the pumping process ( here : @xmath269 ) . as opposed to the resonant rate ( @xmath243 ) above , the decay rates due to off - resonant fields can be written as @xmath270 from these expressions follow the loss rates from @xmath11 due to z pumping : @xmath271 here we have neglected the gain of population in @xmath11 from @xmath7 .
furthermore , we note that due to its scaling with @xmath272 , loss from @xmath11 by oscillator decay should be avoided .
in addition , effective oscillator decay is not useful for the z pumping process . as we will see below ,
this is also the case for x pumping .
therefore , we will generally choose to work with @xmath273 for the ghz protocol ( a weak cooling @xmath274 may nevertheless be used to counteract heating ) . the overall loss rate from @xmath11 due to off - resonant z pumping
is then given by @xmath275 here , the question mark stands for any potential final state ; in the scaling analysis we will typically consider the worst state possible .
for the reasonable assumption of @xmath276 we obtain @xmath277 we conclude that for @xmath278 , the rates from z pumping , ultimately leading to @xmath4 , and thus to @xmath11 , are much stronger than the loss rates from @xmath11 .
+ the derived rates will be used further to analyze the error and preparation time of the protocol in section [ secanalysis ] .
@xmath7 is emptied by x pumping , as described in the main part : here we make use of the fact that in the eigenbasis of @xmath279 , consisting of @xmath19 , @xmath7 is a superposition of states with odd numbers of qubits in @xmath23 , whereas @xmath11 only contains states with an even number of qubits in @xmath23 , @xmath280 to depump @xmath7 without affecting @xmath11 we apply again red- and blue - detuned laser fields @xmath281 with detunings @xmath282 , but including only odd field indices @xmath283 , whereas for even @xmath284 we use @xmath285 . from eqs .
we find for the effective detunings for resonant ( @xmath286 ) and off - resonant ( @xmath287 ) excitation @xmath288 for the effective decay rates we find from eqs .
, @xmath289 for resonant excitation ( @xmath286 ) and for off - resonant ( @xmath287 ) excitation . with these rates and eqs
we obtain the effective lindblad operators @xmath290 p_{n_- } + \sum_{{\rm even } ~ n_- } \sqrt{\kappa_{x \pm , n_-}^{(f \neq n_- ) } } p_{n_- } , ~ ~ ~ ( f = 1 , 3 , 5 , ... , ( f \leq n ) ) , \\
l_{\gamma 0,a , x \pm}^{(f ) } = & \sum_{{\rm odd } ~ n_- } \left[\sqrt{\gamma_{0 , x \pm , n_-}^{(f = n_- ) } } { |0\rangle}_a { \langle-| } p_{n_- } + \mathcal{o}\left(\frac{1}{g^2}\right)\right ] + \sum_{{\rm even } ~ n_- } \sqrt{\gamma_{0 , x \pm , n_-}^{(f \neq
n_- ) } } { |0\rangle}_a { \langle-| } p_{n_- } , ~ ~ ~ ( f = 1 , 3 , 5 , ... , ( f \leq n ) ) , \\
l_{\gamma 1,a , x \pm}^{(f ) } = & \sum_{{\rm odd } ~ n_- } \left[\sqrt{\gamma_{1 , x \pm , n_-}^{(f = n_- ) } } { |1\rangle}_a { \langle-| } p_{n_- } + \mathcal{o}\left(\frac{1}{g^2}\right)\right ] + \sum_{{\rm even } ~ n_- } \sqrt{\gamma_{1 , x \pm , n_-}^{(n_- \neq f ) } } { |1\rangle}_a { \langle-| } p_{n_- } , ~ ~ ~ ( f = 1 , 3 , 5 , ... , ( f \leq n)).\end{aligned}\ ] ] again , we separate the effective lindblad operators by the frequencies of the resonances , this time depending on @xmath22 .
we obtain similar lindblad operators as for z pumping , with enhanced terms to zeroth order in @xmath248 , @xmath291 and suppressed terms acting on the target state @xmath292 it can be seen that for @xmath278 these operators make @xmath7 ( with only odd @xmath22 ) decay rapidly .
the losses from @xmath11 ( with only even @xmath22 ) are caused by the off - resonant drives with @xmath287 .
since these terms are of second order in @xmath248 , they are suppressed for @xmath278 .
again , the effective hamiltonian is compensated by the red- and blue - detuned fields , @xmath293 we can therefore describe the dynamics in terms of rates : with the effective operators from eqs .
we find the decay rates from @xmath7 , @xmath294 for the reasonable assumption of @xmath295 the total rate is therefore given by @xmath296 the expression in eq .
is the total decay rate from the @xmath7 state which is approximately given by the sum of all enhanced decay rates , weighted with the number of states with the same excitation .
the x pumping distributes the population of @xmath7 over all other states which we here denote by a question mark and later on replace by a worst - case assumption .
+ the losses from @xmath11 are only caused by the off - resonant drives with @xmath287 .
using eqs . and ignoring negligible gain processes we obtain the loss rates @xmath297 here , the binomial coefficients originate from the number of states with the same number of atoms in @xmath23 . for the total loss rate from @xmath11 through x pumping with both red- ( @xmath298 ) and blue - detuned ( @xmath299 ) fields we approximately find @xmath300 from this expression we see that the loss terms from @xmath11 due to x pumping are of second order in @xmath248 and thus suppressed for @xmath278 . beside decay out of @xmath7
the x pumping also causes losses from states with odd @xmath22 which have an overlap with states with @xmath86 in the z basis .
this affects the transport from @xmath301 to @xmath13 by the z pumping and thus to @xmath11 by imposing a loss rate @xmath302 we will refer to this process as `` x toss '' below . as we will show below this does not , however , have a significant effect on the scaling of the preparation time and the error , if the strength of the x process is chosen properly .
we conclude that the combination of z and x pumping results in the preparation of a ghz state from any initial state . for @xmath278
the gain rates are engineered to be strong , while the loss rates from @xmath11 are suppressed .
the scaling of the error and the preparation time of the protocol are analyzed in section [ secanalysis ] .
the operator formalism used to derive effective couplings of the ground states in the system is built on the assumption of a perturbative drive which is much smaller than the couplings , decay , or detunings of the excited states , @xmath303 .
effects from saturating the excited states that could possibly slow down the preparation process are thus not included in the rates derived so far . on the contrary
, the dependency of all decay rates on the drives @xmath304 suggests that the optimal rabi frequencies @xmath122 of the drives are infinitely strong and thereby outside the perturbative regime .
a proper assessment of the driving strength @xmath122 therefore requires the inclusion of strong driving effects which begin to play a role for @xmath305 . in the following ,
we include such effects in the dynamics in an approximate way . here ,
we discuss power broadening and population of the excited states and adjust the rates derived from the effective operators to account for power broadening of the excited states .
they will later be used to analytically derive the optimal rabi frequencies @xmath122 and the scaling of the protocol for strong driving .
in addition , in our numerical simulations we use effective operators where we have included the power broadening terms manually .
while this treatment is not rigorous , it provides a convenient tool for rapidly determining the optimal parameters .
these are used to simulate the effective dynamics of the system beyond the regime of weak driving and to match it with the simulations describing a larger hilbert space including excitations .
in addition , we take into account the effect of population of the excited states in our numerics . we first address the effect of power broadening ( or ` line ' broadening ) , regarding a simple model situation : a ground state @xmath28 is resonantly coupled by a field with a rabi frequency @xmath122 to an excited level @xmath29 . in total
, @xmath29 decays at a rate @xmath306 , where @xmath268 is the decay rate into @xmath3 .
we perform adiabatic elimination by setting the derivative of the density matrix to zero , @xmath307 . in the weak driving regime , where the broadening of the excited level can be neglected ,
this yields an excited population of @xmath308 and thus an effective decay rate from @xmath28 into @xmath3 of @xmath309 .
the population gain of state @xmath3 is then given by @xmath310 the same result is obtained when using the effective operators . performing adiabatic elimination with a stronger drive
we need to take into account the population of the excited level .
this yields an excited population of @xmath311 .
then , as the coupling of the ground state @xmath28 to @xmath312 is increased , we need to take into account the population of the coupled subspace of @xmath28 and @xmath29 rather than of state @xmath28 only .
this leads to @xmath313 in the weak driving limit @xmath314 , this new decay rate @xmath315 matches the previous result .
the decay rate for strong driving can thus obtained from the rate which was derived from the effective operators by `` broadening '' the natural line width of the excited state , @xmath316 . in the limit of strong driving @xmath317
, the population is found in @xmath29 with a probability of @xmath318 such that here @xmath319 approaches the constant value of @xmath320 .
therefore , the effective decay rate from @xmath28 to @xmath3 can not be increased beyond half of the line width of the level mediating it .
more generally , we now seek to include power broadening in the effective operators .
this is done by replacing the complex effective detunings , e.g. @xmath208 , here generally denoted as @xmath321 and @xmath322 , by `` power broadened '' ones .
the replacements are made such that the rates obtained from the effective operators agree with the ones derived by adiabatic elimination . to this end
, we make the replacements @xmath323 and @xmath324 , where @xmath187 is the number of atoms that can be excited by the drive . from our numerical simulations
it turns out that the action of power broadening needs to be doubled in the effective operators for z and x to achieve high accuracy between the evolution due to the effective master equation and the more complete master equation in eq . .
this can be attributed to interference between the blue- and red - detuned drives . indeed , considering coherent excitation of a bright state consisting of both
the blue- and the red - shifted dressed state suggests an increase of the broadening by a factor of two and thus the replacements of the effective detunings @xmath325 and @xmath326 .
while these replacements are less rigorous , they are supported by our numerical simulations : it can be seen from fig . 3 in the main manuscript that ( 1 ) the analytical scalings derived using the strong driving operators comprise upper bounds to the numerically obtained scalings and that ( 2 ) the effective operators for strong driving agree well with simulations of the more complete master equation .
+ following the reasoning above , the power broadened decay rate for the z pumping from @xmath55 to @xmath12 is found to be @xmath327 for the x depumping of @xmath7 we have @xmath328 and for the rate of the x process when acting on other states with @xmath55 ( the `` x toss '' rate ) @xmath329 as a consequence of the inclusion of terms for power broadening , increasing the driving strengths in the desired processes also increases the effect of power broadening in the desired decay rates . in the off - resonant decay rates
@xmath330 the effect of power broadening is , on the other hand , negligible ( given that @xmath331 ) .
the detrimental rates thus still increase for a growing @xmath122 while the desired rates saturate .
this is the limiting factor for the drive @xmath122 which we will use to derive the possible @xmath122 further down .
an additional reduction in the fidelity for strong driving comes from the population of the excited induced by the drive . in the following ,
we investigate this effect : + the ghz state , through its contribution from @xmath10 , is coupled to an excited state @xmath332 . despite being off - resonance , all tones of the driving field couple to the transition from @xmath11 to the dressed states of @xmath333 and @xmath334 , with driving strengths of @xmath335 and detunings @xmath336 . for each tone
the excited population is then approximately given by @xmath337 we also consider the excited population caused by x pumping .
this requires representing the ghz state in the x basis as in eq . and leads to an expression @xmath338 these expressions are included in our numerical simulations of the effective dynamics in the strong driving regime and provide another limitation on @xmath122 .
our analytical derivation of the scaling , on the other hand , involves upper bounds to certain error processes and turns out to take place in a parameter regime where the population of the excited states is not significant .
therefore , in the analytical considerations below we leave out its effect .
in the following , we provide an analytical study of the scaling of the ghz scheme .
we derive an expression for the preparation time and optimize it by the choice of the parameters .
the analysis is performed both for weak driving , beginning in section [ secpreparation ] , and for strong driving , in section [ secghzstrong ] .
the different schemes presented in the main manuscript have in common , that the population of a nearly exponential number of states is pumped to subspaces with a polynomial number of states in a number of steps linear in the size of the system . for ghz state preparation
we have engineered strong decay processes from @xmath339 .
the decay is achieved using the z pumping which is only sensitive to @xmath55 and reduces this to @xmath12 , finally leading to @xmath13 . in order to assess the performance of the schemes it is thus important to know the time for a concatenated process consisting of many consecutive z pumping steps .
the rate of each individual decay is given by ( see eq .
( [ gamma0processesforzpm ] ) ) : @xmath340 the average time for this decay to occur is given by the inverse decay rate , @xmath341 for the total time required for pumping from an @xmath55 to an @xmath342 we add the average times for the intermediate steps @xmath343 we can also assign a total decay rate from an initial state with @xmath55 to a final state with @xmath342 , @xmath344 .
it is however more useful to use the total preparation time and to minimize it by the choice of available parameters . here ,
in particular the rabi frequencies @xmath345 of individual field tones @xmath104 can be chosen , as well as the tunable decay rate @xmath346 .
since the pumping occurs from a maximal @xmath55 of @xmath347 to @xmath12 , the worst case preparation time for z pumping is ( using eq . ) found to be given by @xmath348 the time from @xmath73 to @xmath11 differs from that to @xmath13 by a factor of @xmath349 .
this is due to the fact that only half of the population is pumped to @xmath11 and the other half to @xmath7 , which is continuously depumped by the x pumping discussed below .
therefore , on average two attempts are required so that the preparation time is doubled , @xmath350 as anticipated already before eq .
( [ gammaminusformula ] ) , for ghz preparation we will from now on choose the parameter values @xmath351 , @xmath352 , and abbreviate @xmath353 .
in order to obtain dimensionless optimization variables , we will furthermore write @xmath354 ( for @xmath355 ) with nonnegative dimensionless variables @xmath356 .
the quantity @xmath122 is a dimensionful frequency parameter , whose size has been chosen such that ( for the weak driving calculation ) all @xmath357 have to satisfy @xmath358 , i.e. such that the dimensionless parameters @xmath356 satisfy @xmath359 $ ] . in this notation ,
the above ghz preparation time reads : @xmath360 where we have defined the function @xmath361 our goal for now is to find parameters @xmath364 that minimize the @xmath269-error for any given value of the ghz preparation time ( or , equivalently , minimize the ghz preparation time for any fixed error value ) .
minimizing @xmath365 under the constraint @xmath366 by the method of lagrange multipliers leads to @xmath367 approximating the latter harmonic sum gives roughly @xmath368 , and we would have @xmath369 for some @xmath104 ( in particular , for @xmath370 ) if @xmath371 , i.e. if @xmath372 .
however , one can still find the optimal assignment @xmath364 minimizing @xmath365 while obeying @xmath373 and @xmath374 : @xmath375 where we defined the `` ceil-1 '' function @xmath376 and @xmath377 needs to be adjusted such that @xmath378 .
the assignment ( [ afceilingassingment ] ) means that @xmath379 for @xmath380 and @xmath381 for @xmath382 .
that ( [ afceilingassingment ] ) is the unique optimal solution can be checked by the karush - kuhn - tucker ( kkt ) conditions ( * ? ? ?
* section 5.3.3 ) , using that the function @xmath365 to be minimized and the constraint function @xmath383 are both strictly convex in their arguments . for @xmath384 $ ] ( such that the selection between the two cases in ( [ ceil1function ] ) happens at some @xmath385 $ ] ) one can thus compute : @xmath386 and @xmath387 for @xmath388 , the error in both estimates is @xmath389 as @xmath390 and thus irrelevant compared to the @xmath391 terms . in particular , if one wants the numerical factor @xmath145 in the preparation time ( [ restatetaupessimist ] ) to scale like @xmath391 ( like optical pumping ) , then one needs @xmath392 , which we achieve by letting @xmath393 as @xmath390 .
the error @xmath394 can generally be obtained by comparing the preparation time of the desired state of the protocol and the loss rate out of it . using eqs .
( [ restatetaupessimist ] ) and ( [ restategammaminus ] ) , the error @xmath394 is found to be given by @xmath395 thus , to achieve a desired error @xmath394 ( which may be @xmath2-dependent , i.e. @xmath396 ) , one can adjust @xmath397 appropriately . if we assume a relation @xmath398 in order to limit @xmath122 to the weak driving regime ( and where the number @xmath399 may or may not depend on @xmath2 , i.e. @xmath400 ) , we can plug this back into the worst - case preparation time ( [ restatetaupessimist ] ) to obtain @xmath401 where the last estimate holds for @xmath393 as @xmath390 and we have neglected lower - order terms in @xmath2 .
the prefactor is minimized for @xmath402 , leading to a minimal preparation time ( for the desired error @xmath394 ) : @xmath403 * the @xmath404 for @xmath405 have to be chosen as follows ( cf .
( [ afceilingassingment ] ) with the optimal choice @xmath402 ) : @xmath406i.e .
@xmath407 for @xmath408 , and @xmath409 for @xmath410 .
this leads to : @xmath411 * the choice of @xmath412 , and consequently of @xmath413 and @xmath414 , is given by : @xmath415 we also set @xmath351 .
this leads to : @xmath416 the appearance of @xmath417 in the scaling of @xmath418 can be traced back to the fact that the noise @xmath419 acts on each of the @xmath2 atoms ( i.e. the prefactor of @xmath2 in eq .
( [ gammaminusformula ] ) ) .
while so far we have only discussed the z pumping , we will now make a simplified model of both processes , z and x , that create the @xmath11 state .
this model is more pessimistic w.r.t .
the preparation time and the error treatment than the actual ( lindbladian ) dynamics described in section [ secengineeredghz ] . nevertheless , this model will still yield a scaling of the preparation time as @xmath420 with the number of qubits @xmath2 and the error @xmath394 , just like eq .
( [ tau1nforoptimalconsteta2 ] ) , which thus shows that this is indeed the best achievable scaling . our simplified model
is shown in fig .
[ comp4pessfig ] and consists of four compartments as we explain now .
we split the @xmath421-dimensional hilbert space into the subspaces with @xmath422 , counting the number of @xmath3-states appearing in a computational basis state ( as in section [ secengineeredghz ] ) .
all states with @xmath423 are put into compartment 1 , as they are furthest away ( in pumping time ) from the ghz states ; the states with @xmath424 are then put into compartment 2 .
the two hilbert space dimensions belonging to @xmath425 and to @xmath426 are split into the compartments @xmath427 and @xmath428 of fig .
[ comp4pessfig ] , corresponding to the @xmath7 and the @xmath11 states , respectively .
the desired z pumping process creates each of those ghz states with equal rate @xmath429 out of compartment 2 where using eq .
( [ taun1ton1minus1 ] ) we find @xmath430 here , we have again plugged in @xmath431 , @xmath432 and the optimal @xmath433 from eq .
( [ finalresultforafzpumping ] ) and followed the same computation as for ( [ hforgeneralconsteta ] ) , neglecting lower - order terms .
similarly , the rate from compartment 1 to compartment 2 is given by eq .
( [ rateasdfasdfasdfabasdbasdbasb ] ) : @xmath434 as compartment 1 is the furthest away from the desired ghz state , we pessimistically model the @xmath18-toss ( i.e. the action of the x process on the states other than @xmath11 and @xmath7 , see eq . ) to throw any state back to the @xmath423 compartment , with a rate @xmath435 to be computed below .
similarly , we model the errors @xmath436 and @xmath437 affecting @xmath11 such that they move the @xmath11-state back to compartment @xmath438 . by the same rationale ,
even the `` good '' x process @xmath439 is modelled to take the un - wanted state @xmath7 back to compartment 1 .
+ the detrimental z rate for the compartment model as computed with the parameters from the previous subsection , eq .
( [ finalresultforafzpumping ] ) ( see also eqs .
( [ restategammaminus ] ) , ( [ asdfasdfzuio ] ) , and ( [ gforgeneralconsteta ] ) with the optimal @xmath402 ) is @xmath440 for the x rates , it was found in section [ subsectiondepumpingghzx ] that only the rabi oscillations @xmath94 with _ odd _ index @xmath104 should be turned on . similar to section [ secpreparation ] , we write @xmath441 with the dimensionless parameters @xmath442 which we discuss below .
the `` good '' x rate @xmath443 for the compartment model of fig .
[ comp4pessfig ] is then given by eq .
( [ gammaxtotal ] ) : @xmath444 since the ( normalized ) probability distribution @xmath445 is strongly peaked around the values @xmath446 ( similar to the full binomial distribution ) , the exact functional choice of the x coefficients @xmath442 ( for odd @xmath104 ) does not really matter in the limit of large @xmath2 , as long as it is not exponentially fine tuned .
since a similar binomial distribution occurs in the x error rates below as well and as the same reasoning applies , we can take all @xmath442 to be equal as a very good approximation : @xmath447 with this , the rate @xmath443 evaluates to ( with exponentially good accuracy for large @xmath2 ) : @xmath448 as we will see below , with this choice the error induced by the x error rate is smaller than the one from the z process and thus not a major limitation to the preparation of the entangled state . the x error rate @xmath437 is given by ( [ gammaxminus ] ) , and we simplify it again with our above parameter choices , making approximations of the binomial distribution and the other sum which are good in the large-@xmath2 limit : @xmath449_{even}\right)^2}.\end{aligned}\ ] ] in the expression above , @xmath450_{even}$ ] denotes the next higher integer number of @xmath451 . having the limit of large @xmath2 in mind , we evaluate the last sum as follows : @xmath452_{even}\right)^2}=\sum_{1\leq f\leq[n/2]_{even}-1}^{f odd}\frac{f}{\left(f-[n/2]_{even}\right)^2}+\sum_{[n/2]_{even}+1\leq f\leq n}^{f odd}\frac{f}{\left(f-[n/2]_{even}\right)^2}\\ = & \sum_{1\leq f\leq[n/2]_{even}-1}^{f odd}\frac{[n/2]_{even}}{\left(f-[n/2]_{even}\right)^2}+\sum_{1\leq f\leq[n/2]_{even}-1}^{f odd}\frac{f-[n/2]_{even}}{\left(f-[n/2]_{even}\right)^2}\\ & + \sum_{[n/2]_{even}+1\leq f\leq n}^{f odd}\frac{[n/2]_{even}}{\left(f-[n/2]_{even}\right)^2}+\sum_{[n/2]_{even}+1\leq f\leq n}^{f odd}\frac{f-[n/2]_{even}}{\left(f-[n/2]_{even}\right)^2}\\ = & [ n/2]_{even}\left(\frac{1}{1 ^ 2}+\frac{1}{3 ^ 2}+\ldots+\frac{1}{([n/2]_{even}-1)^2}\right)-\left(\frac{1}{1}+\frac{1}{3}+\ldots+\frac{1}{([n/2]_{even}-1)}\right)\\ & + [ n/2]_{even}\left(\frac{1}{1 ^ 2}+\frac{1}{3 ^ 2}+\ldots+\frac{1}{(n-[n/2]_{even}-1)^2}\right)+\left(\frac{1}{1}+\frac{1}{3}+\ldots+\frac{1}{(n-[n/2]_{even}-1)}\right)\\ \approx&2\cdot\frac{n}{2}\sum_{n=1,3,5,\ldots}^{\infty}\frac{1}{n^2}=2\cdot\frac{n}{2}\cdot\frac{3}{4}\sum_{n=1,2,3,4,\ldots}^\infty\frac{1}{n^2}=n\frac{3\pi^2}{4\cdot6}\approx1.23n\approx\frac{5n}{4},\end{aligned}\ ] ] neglecting subleading terms in @xmath2 .
this finally gives : @xmath453 finally , the x toss rate @xmath435 in fig .
[ comp4pessfig ] is given by eq .
( [ xtossrate ] ) , which with our parameter choices becomes [ note that ( [ xtossrate ] ) is half of the `` good '' rate ( [ gammaxtotal ] ) , which we have computed in eq . ( [ goodxrate2 ] ) already ] : @xmath454 from the process in fig .
[ comp4pessfig ] one can see that the optimal parameters @xmath455 have to be chosen such that the good @xmath18-process and the good @xmath269-process have about the same rate : if the x pumping rate @xmath443 is too weak , population will accumulate in @xmath7 by the z pumping ( @xmath456 ) . on the other hand
, a too strong x pumping will hinder the preparation mechanism through the x toss effect .
we thus set the rates for the desired processes , z pumping and x depumping to be equal , @xmath457 this results in : @xmath458 ( note that , for all @xmath459 , this choice is consistent with the requirement @xmath460 for the weak - driving analysis ) .
* for the @xmath269-pumping , we make the choice of parameters found to be optimal in eq .
( [ finalresultforafzpumping ] ) : @xmath461 meaning that @xmath462 .
* for the @xmath18-pumping we take ( see eq .
( [ finalchoiceax ] ) ) : @xmath463 * then one obtains for the effective rates in fig .
[ comp4pessfig ] : @xmath464 * the total error rate ( leading from compartment 4 to compartment 1 ) is thus @xmath465 the transition matrix for the 4-compartment model of fig .
[ comp4pessfig ] is : @xmath466 the steady - state population @xmath467 is given as the solution ( normalized to the sum of entries being 1 ) of the equation @xmath468 .
this gives : @xmath469 the _ steady - state fidelity _ is just @xmath470 , and the _ error _ is thus @xmath471 ( here the approximation in the second line was made for analytical convenience and gives a slightly pessimistic bound . ) in the scaling with large @xmath2 , this expression for @xmath394 agrees with the one implied by eq .
( [ gammafinalfromghz - zpumpingonly ] ) that was found by other means before , and the prefactor is similar .
thus , to achieve a desired stationary error @xmath394 , we need to adjust @xmath397 such that : @xmath472^{-1/2}\,.\label{gammawithxandzpumpingasdfqer}\end{aligned}\ ] ] below we will use this expression instead of the results obtained in ( [ gammafinalfromghz - zpumpingonly ] ) which were derived by considered only the z pumping .
so far we have discussed the z pumping separately from the x pumping , deriving an individual characteristic time @xmath473 .
it now remains to derive an analytical expression for the total ghz pumping time that is obtained in the presence of x pumping , using the parameters ( [ gammawithxandzpumpingasdfqer ] ) and ( [ finalresultforafzpumping ] ) . in order to factor out the dependence of the stationary error @xmath394 and to obtain a tractable analytical expression ,
we make ( for the computation of the ghz preparation time ) an approximation to the transition matrix ( [ transitionmatrix4x4pess ] ) by dropping the small terms leading out of the ghz state ( these terms vanish in the limit of @xmath474 ) .
that is , we set the fourth column in ( [ transitionmatrix4x4pess ] ) to zero : @xmath475 for the initial population vector @xmath476 , which corresponds to the whole population being in the worst state @xmath7 of fig .
[ comp4pessfig ] , we have the following evolution : @xmath477 thus , on time - scales larger than any fixed @xmath478 , the transition from the worst state @xmath7 to the desired state @xmath11 happens at least as fast as in an exponential decay process , with an approximation @xmath106 for the effective exponential rate computed as : @xmath479 note that the fraction in the last expression will depend both on @xmath480 and on @xmath2 , since the @xmath2-dependence of the transition matrix @xmath481 in eq .
( [ transitionmatrix4x4pessgood ] ) can not be factored out completely . to get a meaningful expression for the exponential rate ,
the timescale @xmath478 should be chosen comparable to the other relevant timescales of the process .
for definiteness we will thus set @xmath482 throughout .
the characteristic preparation time of @xmath11 in the sense of an exponential rate is then : @xmath483}{=}~\frac{3\gamma}{2\omega^2}\log n\,\left(\frac{-\log\left(1-p_4(t_0)\right)}{\gamma_z^+t_0}\right)^{-1}\\ & \stackrel{[\omega=\alpha\gamma]}{=}~\frac{3}{2\alpha^2\gamma}\log n\,\left(\frac{-\log\left(1-p_4(t_0)\right)}{\gamma_z^+t_0}\right)^{-1}\\ & \stackrel{[\text{eq.}~(\ref{gammawithxandzpumpingasdfqer})]}{=}~\frac{n^{1/2}\log^2n}{\alpha^2g\sqrt{e}}\,\left[\frac{9\sqrt{3}}{8}\,\frac{\gamma_z^+t_0}{-\log\left(1-p_4(t_0)\right)}\,\sqrt{\left(1 + \frac{5}{9\log^2n}\right)\left(1+\frac{1}{3\log n}\right)}\,\right]\ , .
\label{upperboundontaughzinpess4compmodel}\end{aligned}\ ] ] in the limit of large @xmath2 , the square root inside the bracketed expression will tend to @xmath438 . and
also the fractional expression involving @xmath484 will tend to a constant number independent of @xmath2 , since large values of the first column in the matrix in ( [ transitionmatrix4x4pessgood ] ) mean that the transition time out of the first compartment is insignificant compared to the other transition times , which are all
_ _ in__dependent of @xmath2 .
this is also easily seen numerically . when taking power broadening into account ( see section [ powerbroadeningdescriptionsubsection ] ) , then instead of eq .
( [ rateasdfasdfasdfabasdbasdbasb ] ) from the weak driving scenario , the favorable transition rates of the z process are now given by eq .
( [ gammagoodpower ] ) ( for @xmath422 ) : @xmath486 where we have again used the parameter values and abbreviations @xmath353 , @xmath431 , @xmath487 as in section [ secpreparation ] .
thus , instead of ( [ worstcasezpreptimewithfactor2 ] ) , the z pumping time is now : @xmath488 the error rate from ( [ restategammaminus ] ) remains unchanged : @xmath489 as below ( [ asdfasdfzuio ] ) , we now minimize @xmath490 ( as a function of the variables @xmath491 ) while keeping @xmath473 constant .
this leads , by the method of lagrange multipliers , to : @xmath492 where @xmath493 is a lagrange multiplier ( that has units of frequency@xmath494 ) , which we will determine later . plugging this back into ( [ tau1nghzpowerasdf ] ) and ( [ gammaminusghzpowerasdfqwer ] )
, we get : @xmath495 @xmath496 thus , the stationary z error is ( cf .
( [ errorinzpumping ] ) ) : @xmath497 thus , when the desired stationary z error @xmath498 is given , then @xmath493 and @xmath397 are determined by each other ( since @xmath39 and @xmath2 are fixed ) : @xmath499\right)\,.\label{lambdaghzlangranemultiplier}\end{aligned}\ ] ] plugging this back into ( [ tau1nghzpowerwithlambda ] ) , we get : @xmath500\,.\end{aligned}\ ] ] for fixed @xmath2 , @xmath39 , @xmath498 , and @xmath397 , this is the minimal z pumping time ( i.e. minimized over all choices of pumping rates @xmath491 ) . since we do not want to fix @xmath397 a priori , we minimize the last expression over @xmath397 , finding that the optimal choice is @xmath501 the corresponding pumping strengths can now be computed from ( [ omegafghzpowerbroaderninglambda ] ) and ( [ lambdaghzlangranemultiplier ] ) : @xmath502 and the optimal z pumping time for given @xmath2 , @xmath39 , @xmath498 is then ( compare to ( [ tau1nforoptimalconsteta2 ] ) without power broadening ) : @xmath503 from now on we set @xmath504 , as this neglects only subleading terms .
we also take power broadening into account for the desired x pumping rate ( see ( [ goodxratepowerbroadening ] ) ) and for the x toss rate ( see [ tossxratepowerboradeningasdfqwerasdf ] ) , which we both compute as in ( [ startcomputinggoodxrateasdfqwerasdf])([goodxrate2 ] ) and ( [ xtossrate2 ] ) , and we take the detrimental x rate from ( [ badxrate ] ) : @xmath505 now we adjust the x parameters @xmath122 and @xmath506 such that the x rates agree with the corresponding z rates ( cf . section [ subsectioncompartmentmodel ] ) , i.e. @xmath507 and @xmath508 . to solve this for @xmath506 and @xmath122 exactly
, one would have to solve a cubic equation .
as this is quite cumbersome and uninformative , we are looking for solutions which satisfy the following scaling ansaetze : @xmath509 and @xmath510 .
then , one finds actually _ two _ possible solutions for @xmath122 and @xmath506 , which lead to the desired scaling of @xmath439 and @xmath437 ( as @xmath2 becomes large ) .
one of the solutions is given by @xmath511 formally , another solution exists , but this is in the extremely saturated regime where our effective operators do not apply .
finally , we need to find the relation between the above rate @xmath512 and the total ghz preparation time @xmath513 ( which includes the errors and x toss ) and the total stationary error @xmath394 , just as in section [ subsectiontransitionmatrixetc ] . for this , we consider a simplified 3-compartment model which constitutes a very good approximation in the large-@xmath2-limit , in which the parameters ( [ twodifferentchoicesforgammaomegaasdfqwer ] ) were computed in the first place . here , compartment a comprises the states with @xmath422 , compartment b is the state @xmath7 and compartment c the state @xmath11 .
then the transition matrix is ( cf .
( [ transitionmatrix4x4pess ] ) and fig .
[ comp4pessfig ] ) : @xmath514 again with @xmath515 , and again as in section [ subsectiontransitionmatrixetc ] we denote by @xmath481 the transition matrix without the `` bad '' rates @xmath419 . from the stationary vector @xmath516
satisfying @xmath468 we can again compute the total stationary error @xmath394 : @xmath517 we therefore have to set @xmath518 in all previous expressions .
the relation between @xmath456 and @xmath519 is similar to eq .
( [ rateasdfqwerycxvasdfqutqw ] ) : @xmath520 where again @xmath521 denotes the population in the ghz state when starting from @xmath7 state . computing this for the above transition matrix and plugging in the values for @xmath456 and @xmath394 from above
, we obtain : @xmath522 thus , in the large-@xmath2 limit , the final ghz preparation time @xmath523 is : @xmath524 this is achieved by the following parameter choices in terms of @xmath39 , @xmath2 , @xmath394 : @xmath525 again , as in eq .
( [ relationstatanddynerror ] ) , if one is interested in the dynamical error @xmath526 instead of the static error @xmath394 , one should everywhere set @xmath527 for @xmath528 ( or @xmath529 for @xmath530 ) .
furthermore , the ghz preparation time is then prolonged by an additional factor of @xmath531 ( see eq .
( [ asdfqweryxcfv ] ) ) .
if the maximal preparation time @xmath533 of the preparation procedure is limited , e.g. by experimental constraints or by the available coherence times of the underlying hardware , we can still try to adjust the remaining parameters of the preparation scheme in such a way , that the error @xmath534 after this fixed time @xmath532 is minimized .
we first take the same setup of the dynamical problem as in subsection [ appdynamical ] , i.e. as in eqs .
( [ gammaindynamicalsectionasdf])-([gammaminuswithh ] ) and ( [ erasdfwerasdfr ] ) : @xmath535\,,\label{asdfqewrasdfjknkjhhjhj}\end{aligned}\ ] ] where we again assumed a relation @xmath398 between @xmath122 and @xmath397 with a fixed parameter @xmath399 .
it is now our task to minimize the preparation error @xmath536 for each fixed @xmath532 , @xmath2 , @xmath39 , and for the given constants and functions @xmath399 , @xmath266 , @xmath537 , and @xmath538 .
the free parameter is thus @xmath397 ( which determines @xmath398 ) .
this optimization of the error @xmath536 for fixed preparation time @xmath532 can be done similarly as the optimization in subsection [ appdynamical ] . to find the optimal parameters @xmath397 and @xmath122 , we demand @xmath539 , which becomes : @xmath540 this is equivalent to @xmath541 whose solution is again given by the lambert w function : @xmath542 where this time we have to take the branch @xmath543 since both sides are nonnegative .
we can now solve the condition ( [ bringtolambertwform ] ) for the optimal @xmath544 and plug this back into the optimal preparation error @xmath536 from ( [ asdfqewrasdfjknkjhhjhj ] ) together with the solution in terms of the lambert w function .
then we obtain for the optimal preparation error @xmath536 after a fixed time @xmath532 : @xmath545 and the optimal rate @xmath397 is determined by ( [ blasdfqewradsfb ] ) .
as we have seen previously , plausible behaviours of the underlying rates @xmath546 and @xmath419 are given by a @xmath2-independent @xmath547 and a linear @xmath548 .
the expression for the optimal error @xmath536 at fixed time @xmath532 and particle number @xmath2 can then be evaluated numerically . unlike in subsection [ appdynamical ]
however , an asymptotic formula for the lambert w function can not be applied here since the optimal solution lies right in the middle of the competing polynomially growing ( in @xmath397 ) and exponentially decaying behaviours in ( [ asdfqewrasdfjknkjhhjhj ] ) . to obtain more analytical insight , we will thus now upper - bound the particle number @xmath2 , for a given fixed preparation time and prescribed maximal error @xmath526 . for concreteness of the exposition we will also take @xmath549 and @xmath550 in ( [ asdfasdfasdfasdfasdfasdfasdfasdf])-([asdfqewrasdfjknkjhhjhj ] ) , although the computation is easily generalized to more general behaviours of @xmath546 and @xmath419 . if we demand @xmath551 , then from ( [ asdfqewrasdfjknkjhhjhj ] ) we certainly need to have @xmath552 from the last inequality , we obtain the condition @xmath553 , i.e. the requirement @xmath554 on @xmath397 . using this in the first of the two inequalities , we finally obtain : @xmath555 thus ,
if the maximal preparation time @xmath532 is given and a certain maximal error @xmath526 is to be achieved , then the maximal size of the ghz state is limited to a number @xmath2 of particles that is proportional to the square of the `` dimensionless coherence time '' @xmath556 and also proportional to the allowed error @xmath526 ( up to logarithmic factors ) .
one can notice that the bound on @xmath2 from ( [ fiadfqewrxcvasdfqwer ] ) could also be simply obtained ( up to logarithmic factors ) by solving the main result ( [ asdfqweryxcfv ] ) from subsection [ appdynamical ] for @xmath2 .
thus , if we now want to take the effect of powerbroadening into account , we solve the main final result from section [ secghzstrong ] for @xmath2 , namely eq .
( [ finalghzpowerasdfqwerycxv ] ) , to obtain ( up to logarithmic factors ) : @xmath557 note that this same result would be obtained from the above discussion by a choice of @xmath537 and @xmath538 satisfying @xmath558 , i.e. for example by rates @xmath546 and @xmath419 given as @xmath559 and @xmath560 . | we present a technique for the dissipative preparation of highly entangled multiparticle states of atoms coupled to common oscillator modes . by combining local spontaneous emission with coherent couplings we engineer many - body dissipation that drives the system from an arbitrary initial state into a greenberger - horne - zeilinger state .
we demonstrate that using our technique , highly entangled steady states can be prepared efficiently in a time that scales polynomially with the system size .
our protocol assumes generic couplings and will thus enable the dissipative production of multiparticle entanglement in a wide range of physical systems . as an example
, we demonstrate the feasibility of our scheme in state - of - the - art trapped - ion systems .
multiparticle entanglement is an essential resource for quantum computation and information @xcite , e.g. in quantum error correction @xcite , quantum memories @xcite , and entanglement - enhanced quantum measurement schemes @xcite . among the most important states which exhibit genuine multiparticle entanglement
are greenberger - horne - zeilinger ( ghz ) states @xmath0 deterministic preparation of such states has so far been performed using time - dependent unitary gates @xcite , which have recently yielded impressive progress towards entangling larger numbers of qubits @xcite , and feedback control schemes @xcite . these operations however suffer from quantum noise , causing decoherence and dissipation so that it remains difficult to prepare high - fidelity multiparticle entangled states with these methods @xcite .
recently , dissipative state preparation has been proposed as an alternative approach where the dissipative environment is actively engineered and used to prepare states relevant for quantum information and simulation @xcite .
numerous theoretical studies on the production of bipartite entangled states @xcite have since been performed and the first experimental demonstrations @xcite have been realized .
more recently , also dissipative schemes for the generation of multipartite entangled states @xcite have become available , e.g. for the preparation of states stabilized by local interactions @xcite .
it has however remained a challenge to prepare in a scalable way states , like ghz , that are highly entangled in the sense that they can not be stabilized by local operators @xcite . in this letter , we extend the range of the dissipative approach by demonstrating a scalable technique for the dissipative preparation of highly entangled states of many particles . we show that , by using local spontaneous emission as a generic source of dissipation , we can engineer nontrivial many - body dissipative interactions @xcite which are tailored to produce multiparticle ghz states . our scheme is deterministic and operates by continuous optical driving from an arbitrary initial state towards the desired steady state using weak classical fields .
the preparation time of our protocol is found to exhibit a favorable polynomial scaling with the number of qubits .
in addition to our generic system - independent scheme , we describe an implementation in a system of trapped ions .
qubits is realized by two operations : _
( i ) _ `` z - pumping '' of states with other than @xmath1 or @xmath2 atoms in state @xmath3 to @xmath4 . _
( ii ) _
@xmath4 is a superposition of @xmath5 and @xmath6 so that @xmath7 needs to be removed by the parity - selective `` x - pumping '' . ] in our protocol , preparation of a steady ghz state of @xmath2 qubits starting from an arbitrary initial state is accomplished by two simultaneous operations shown in fig .
[ figprotocol ] : _ ( i ) _ pumping all states with more than zero but less than all qubits in state @xmath3 ( @xmath8 ) to the state @xmath4 , which can be written as a superposition @xmath9 , and _
( ii ) _ removing the ghz state with the wrong phase , @xmath6 from the subspace spanned by @xmath4 and @xmath10 . operation _
( i ) _ is implemented such that it fulfills the main requirement of a dissipative protocol : it has to pump an exponential number of states efficiently , i.e. in polynomial time . in principle ,
standard optical pumping @xcite satisfies this criterion as well , but would also erase the state @xmath10 , thus ruling out the possibility to prepare @xmath11 with high fidelity .
instead , we design a new procedure which is selective in the number of atoms in @xmath3 .
this allows us to pump only states with @xmath8 to states with @xmath12 , and thus eventually to @xmath4 ( @xmath13 ) .
we refer to this operation shown in fig .
[ figprotocol ] as `` z - pumping '' , since it is based on counting the number of atoms in the eigenbasis of @xmath14 . a second operation _
( ii ) _ is required to remove the undesired @xmath7 state in a continuous manner , as also illustrated in fig .
[ figprotocol ] . to this end
, we perform a pumping process selective in the parity @xmath15 ( @xmath16 ) . here
, we apply the recipe for stabilizer pumping from ref .
@xcite to the case of the parity stabilizer , @xmath17 : expressed in terms of the eigenstates of @xmath18 , @xmath19 , any state @xmath20 with @xmath21 , such as @xmath11 , is a superposition of only those product states that contain an even number , @xmath22 , of @xmath23 qubits ; e.g. , for @xmath24 , @xmath25 . on the other hand
, @xmath22 is odd for any state @xmath20 with @xmath26 , such as @xmath7 ; for @xmath24 , @xmath27 . by pumping all states with odd @xmath22 to other states , in the following referred to as `` x - pumping '' , we achieve the depumping of @xmath7 .
the two operations required for our protocol can be realized using a generic setup as described next .
we assume a general system of @xmath2 particles ( `` atoms '' ) , shown in fig .
[ figrealization ] ( a)-(c ) .
each atom supports two stable ground states @xmath28 and @xmath3 and two excited states @xmath29 and @xmath30 . the atoms are driven by classical multi - tone driving fields , with identical amplitudes on all ions , as described by the hamiltonians @xmath31 with strengths @xmath32 and detunings @xmath33 , where @xmath34 denotes the desired operation and @xmath35 the field tone .
the transitions of the atoms are collectively coupled to two harmonic oscillator modes , @xmath36 and @xmath37 , e.g. , two resolved resonator modes in cavity or circuit qed @xcite , or two phononic modes in an ion trap setup @xcite , as modeled by the hamiltonians @xmath38 with @xmath39 being the coupling constant . for the dissipative process
we consider decay by spontaneous emission from the excited states @xmath40 to the ground states @xmath41 at a rate @xmath42 ( with @xmath43 ) , described by jump operators @xmath44 for @xmath45 , @xmath46 acting incoherently on each atom @xmath47
. qubits . setup ( a ) .
we consider a chain of @xmath2 sub - systems ( `` atoms '' ) with four levels , coupled to two common harmonic oscillators . as dissipative processes we assume spontaneous emission from the excited levels @xmath48 .
we apply coupling configurations ` z ' ( b ) , and ` x ' ( c ) consisting of @xmath49 driving tones @xmath50 with detunings @xmath51 in ( b ) , @xmath52 tones with @xmath53 and @xmath54 in ( c ) , and couplings of the atoms to the oscillator modes @xmath39 .
( d ) `` z - pumping '' towards @xmath4 using the couplings in ( b ) .
ground states are coupled to atom - excited states by weak driving .
depending on the number @xmath55 of atoms in @xmath3 , the excited states form dressed states with oscillator - excited states at energies @xmath56 . by applying fields with detunings @xmath57 for @xmath58 (
only the red - detuned fields are shown ) , all states except @xmath5 and @xmath59 are pumped to @xmath60 .
@xmath7 is emptied by the parity - selective x - pumping as described in the text . ]
we realize the z- and x - pumping operations by engineering selected transitions to be driven resonantly , while suppressing others due to off - resonant driving . for z - pumping
we use the coupling configuration in fig .
[ figrealization ] ( b ) . here
, a coupling @xmath39 of the oscillator @xmath36 to the transition @xmath61 and a weak drive on the same transition are used to effectively `` count '' the number of atoms in state @xmath3 . in fig .
[ figrealization ] ( d ) , we illustrate this mechanism for @xmath62 qubits .
the weak driving tones @xmath63 couple the ground states to atom - excited states .
for example we consider @xmath64 which is coupled to @xmath65 .
this state is in turn coupled to an oscillator - excited state @xmath66 by the atom - oscillator coupling .
because of constructive interference between the two terms in @xmath67 this coupling has a strength of @xmath68 . as a consequence of the strong atom - oscillator coupling ,
the atom- and the oscillator - excited state form dressed states @xmath69 at detunings @xmath70 , see fig . [ figrealization ] ( d ) . applying a weak driving field with a detuning @xmath71 ,
one thus excites ground states with @xmath72 to excited states like @xmath67 .
since @xmath29 is subject to spontaneous emission to @xmath28 and @xmath3 , @xmath67 decays either back to the manifold of states with @xmath72 or `` forward '' to states with @xmath73 .
in general , the couplings of the atom- and oscillator - excited states have a strength of @xmath74 that depends on the number @xmath55 of atoms in @xmath3 so that the dressed states are shifted by @xmath75 .
this creates a resonance condition depending on @xmath55 of the initial state which we can use to selectively drive a manifold . applying a drive with @xmath76 , states with @xmath77
are thus pumped to @xmath13 , and thereby to @xmath11 and @xmath7 . on the other hand , with the detunings @xmath78 and @xmath79 , the state @xmath10 is excited only off - resonantly and thus decays slowly ( see fig .
[ figrealization ] ( d ) , so that @xmath11 remains almost unaffected . to realize the full z - pumping process based on the mechanism above
, we apply a weak drive consisting of @xmath49 tones @xmath63 with detunings @xmath80 ranging from @xmath81 to @xmath82 .
this gives rise to effective decay processes @xcite described by @xmath83 here , @xmath84 is the projector onto the ground states with @xmath55 atoms in state @xmath3 , @xmath47 denotes the atom subject to decay , and @xmath85 are the strongly enhanced decay rates of the states which are resonantly excited .
each ground state with @xmath86 then decays towards one with @xmath12 by the engineered spontaneous emission in eq . .
the concatenation of these decay processes causes a continuous drift towards states with smaller @xmath55 , finally ending at the state @xmath4 with @xmath13 . for a suitable choice of the @xmath87 @xcite , the preparation time @xmath88 of @xmath4 and the corresponding rate @xmath89 can then have a favorable scaling @xmath90 , which is similar to optical pumping where the transition rate of each level is proportional to the number @xmath55 of its excitations
however , as opposed to standard optical pumping , the z - pumping is engineered such that it does not affect @xmath10 , and thus neither @xmath11 , since resonant excitation out of @xmath10 would require the tone @xmath91 . by excluding such tones from the drive the excitation out of @xmath10 is off - resonant and thus much weaker .
the weak off - resonant excitation of @xmath10 by the other driving tones leads only to a small leakage from @xmath11 . since the energy gap between the resonances and the driving tones @xmath92 decreases with @xmath2 , the leakage rate from @xmath11 increases with the number of qubits , and can be estimated to be @xmath93 .
the resulting error can , however , be compensated by reducing the speed of the protocol by a small polynomial factor as we discuss below . having operation _ ( i ) _ realized , we now turn to operation _
( ii ) _ , the x - pumping , which removes states with an odd number @xmath22 of atoms in @xmath23 . to implement it we use a similar mechanism as was used to `` count '' @xmath55 above , except that now we need to do this in a different basis .
we achieve this operation using the coupling configuration in fig .
[ figrealization ] ( c ) : coupling the transitions from @xmath28 to @xmath30 and from @xmath3 to @xmath30 by fields of the same strength , but with the opposite phase , results in a coupling of the transition from @xmath23 to @xmath30 . to avoid interference with the z - pumping
, we consider a second excited state @xmath30 and a second oscillator mode @xmath37 .
it is nonetheless possible to implement the described operations with a single excited level and a single oscillator mode in a stroboscopic manner , resulting in a quasi - steady state . in the x - pumping ,
the transitions between the excited level @xmath30 and the ground levels @xmath28 and @xmath3 are coherently coupled to the harmonic oscillator @xmath37 and excited by a multi - tone drive @xmath94 .
opposite phases on both transitions result in a coupling of the transition @xmath95 , similar to the coupling @xmath61 in the z configuration . in this way
, we make the x - pumping selective in @xmath22 in a similar manner as the z - pumping is selective in @xmath55 : applying @xmath96 field tones with detunings @xmath97 for @xmath98 we resonantly excite @xmath59 to dressed states which lie at @xmath99 for @xmath100 .
thereby we make @xmath7 decay to random states by effective spontaneous emission with a strong rate @xmath101 , similar to the z - pumping , the decreasing energy gap between the dressed states gives rise to a leakage rate from ghz , @xmath102 ( using @xmath103 for odd @xmath104 and @xmath105 for even @xmath104 ) , which increases with the number of qubits @xmath2 .
the simultaneous action of the z- and the x - pumping prepares @xmath11 and maintains it as the unique steady state of the dissipative dynamics .
however , since the z - pumping is disturbed by the x - pumping , the latter has to be sufficiently weak so that the z - pumping has a sizable probability of reaching the final state @xmath4 before being subject to x - pumping ; picking similar rates for the x - pumping and the total z - pumping @xmath106 , this requirement does not slow down the preparation process significantly @xcite . to find the preparation time we can consider a simple model where the rate of pumping to @xmath4 is determined by @xmath107 , and where @xmath11 and @xmath7 are pumped out with rates @xmath108 and @xmath109 , respectively . further details on the model
are given in the supplementary , but in short we find that the steady state fidelity @xmath104 is determined by the ratio of the decay out due to off - resonant excitation at the rate @xmath110 , and the effective preparation rate @xmath111 ( using @xmath112 , @xmath113 for @xmath114 , and @xmath115 for @xmath116 ) . this gives a steady state error @xmath117 which is approached exponentially in time @xmath118 . to avoid having an increasing error with the qubit number @xmath2 we assume that we can control the decay rates of the excited states @xmath29 and @xmath30
this is for instance the case if the states are metastable states coupled to higher lying unstable states with a laser field @xcite .
the increase of the error with @xmath2 can then be compensated by having a sufficiently low decay rate @xmath119 , which keeps the error @xmath120 constant for growing @xmath2 , but prolongs the necessary preparation time @xmath121 .
these considerations , however , assume a weak driving @xmath122 , whereas for strong driving the pumping rate becomes limited by power broadening .
we therefore need to use a suitably low driving strength @xmath123 which sets a limit on the preparation rate .
these considerations and parameter values can be turned into a rigorous upper bound on the preparation time @xcite : @xmath124 we can thus prepare a ghz state at any desired fidelity @xmath125 within a preparation time that has a low - order ( @xmath126 ) polynomial scaling in the number of qubits , for a coupling @xmath39 independent of @xmath2 .
if instead the total preparation time is restricted to @xmath127 , then the preparation error of an @xmath2-particle ghz state will necessarily obey @xmath128 and @xmath129 , limiting the achievable fidelity to @xmath130 @xcite .
( black dashed line ) within as short a time as possible . in ( b ) we show the scaling of the preparation time with the number of qubits .
both panels ( a ) and ( b ) show different degrees of truncation of the hilbert space ( dash - dots / blue squares effective ground state dynamics after adiabatic elimination , solid lines / green circles one excitation , dashes / red triangles two excitations ) . in ( b ) , small symbols represent analytically optimized parameters and large symbols numerically optimized parameters .
we find a polynomial scaling of the preparation time which is within our analytical bound ( black dashed line in ( b ) ) . ] to confirm the results of the simple model we simulate the protocol numerically : in fig .
[ figcurves ] ( a ) we plot the time - evolution towards a steady ghz state for @xmath131 qubits resulting from our protocol .
the plots are obtained by numerically simulating an effective master equation @xcite as well as the complete master equation truncated to one or two excitations . here
we optimize the available parameters ( driving strengths , tunable decay rates ) to reach a fidelity @xmath132 of the ghz state in minimum time .
the resulting preparation times and the analytical bound from eq . is shown in fig .
[ figcurves ] ( b ) .
these results confirm that the scheme exhibits a low - order polynomial scaling of the preparation time with the number of qubits .
in contrast to the scheme in @xcite , our protocol requires only two operations for a ghz state of @xmath2 qubits . furthermore , the highly directed z - pumping is only weakly perturbed by the polynomially weaker x - pumping and thus allows for a polynomial scaling of the protocol .
the ingredients necessary for our scheme are available in trapped ion experiments .
one suitable setup consists of a chain of @xmath2 trapped ions , each with two ( meta-)stable ground levels @xmath28 and @xmath3 and two auxiliary levels , @xmath29 and @xmath30 .
tunable decay of the auxiliary levels by spontaneous emission can be realized by a repumper to a higher lying rapidly decaying state @xcite .
two phononic modes , cooled to the ground state , and coupled to the sidebands of the ions transitions , can be used as the harmonic oscillators @xmath36 and @xmath37 . for the pumping , we require @xmath49 field tones in the z configuration and @xmath52 tones in the x configuration .
an alternative stroboscopic implementation requires only a single auxiliary level , interchanging between the roles of @xmath29 and @xmath30 , a single phononic mode , interchanging between being @xmath36 and @xmath37 , and a single field tone with tunable detuning applied on either the transition @xmath133 or @xmath134 . with @xmath135 khz ,
typical preparation times are @xmath136 ms . on such timescales collective
dephasing needs to be taken into account , but can be overcome by switching the role of @xmath28 and @xmath3 in the z - pumping for every second ion , thereby preparing @xmath137 which is in a decoherence free subspace and is equivalent to a ghz state @xcite , or by using clock states @xcite .
fluctuations of @xmath39 of @xmath138 result in a reduction of the fidelity by @xmath139 for @xmath140 qubits , whereas fluctuations of @xmath141 have an effect at the sub - percent level .
ac stark shift fluctuations are suppressed since both red- and blue - detuned driving tones ( e.g. @xmath75 ) are applied .
heating of the motion would constitute another error on the timescale of the scheme , but this can be made negligible for cryogenic traps @xcite .
we have shown that dissipative state preparation can be extended to the efficient generation of highly entangled steady states of many particles .
we achieve this by engineering complex multiparticle dissipation which deterministically drives the system into a desired steady state within a time scaling only polynomially with the size of the system .
the generic couplings assumed in our approach can be found in a variety of physical systems , such as trapped ions where the basic ingredients of the scheme have already been demonstrated @xcite . as specific examples
we have considered the preparation of highly entangled ghz states , which are paradigmatic multiparticle entangled states , but the developed techniques are applicable to a range of other quantum information tasks . particularly relevant further possibilities are the construction of quantum error correcting codes @xcite or the observation of exotic phase transitions @xcite induced by multiparticle dissipation .
the research leading to these results has received funding under the european union s seventh framework programme ( fp/2007 - 2013 ) through the erc grant agreement qios ( grant no .
306576 ) and through siqs ( grant no .
600645 ) .
this work was also supported by the villum kann rasmussen foundation . f.r .
acknowledges helpful conversations with yiheng lin and support from the studienstiftung des deutschen volkes .
d.r . was supported by the marie curie fellowship quintyl under grant no .
298742 .
99 m. a. nielsen and i. l. chuang , _ quantum computation and quantum information _
( cambridge university press , cambridge , 2000 ) .
p. w. shor , _ phys .
rev .
a _ * 52 * , 2493 ( 1995 ) .
a. m. steane , _ phys .
rev .
lett . _ * 77 * , 793 ( 1996 ) .
m. fleischhauer and m. d. lukin , phys .
rev .
a * 65 * , 022314 ( 2002 ) .
v. giovannetti , s. lloyd , and l. maccone , science * 306 * , 1330 ( 2004 ) .
g. toth and i. apellaniz , j. phys . a : math
. theor . * 47 * , 424006 ( 2014 ) .
a. rauschenbeutel _ et al .
_ , science * 288 * , 2024 ( 2000 ) .
c. a. sackett _
et al . _ ,
nature * 404 * , 256 ( 2000 ) .
f. schmidt - kaler , h. hffner , m. riebe , s. gulde , g. p. t. lancaster , t. deuschle , c. becher , c. f. roos , j. eschner , and r. blatt , nature * 422 * , 408 ( 2003 ) .
l. dicarlo _
et al . _ ,
nature * 467 * , 574 ( 2010 ) ; m. neeley _
et al . _ ,
nature * 467 * , 570 ( 2010 ) .
d. leibfried _ et al .
_ , nature * 438 * , 639 ( 2005 ) .
h. hffner _
et al . _ ,
nature * 438 * , 643 ( 2005 ) .
t. monz , p. schindler , j. t. barreiro , m. chwalla , d. nigg , w. a. coish , m. harlander , w. hnsel , m. hennrich , and r. blatt , phys .
rev .
lett . * 106 * , 130506 ( 2011 ) . h. m. wiseman and g. j. milburn , _ quantum measurement and control _
( cambridge university press , cambridge 2009 ) . v. bergholm and t. schulte - herbrueggen , arxiv:1206.4945 ( 2012 ) .
d. riste , m. dukalski , c. a. watson , g. de lange , m. j. tiggelman , y. m. blanter , k. w. lehnert , r. n. schouten , and l. dicarlo , nature * 502 * , 350 ( 2013 ) .
d. p. divincenzo , fortschr . phys . * 48 * , 771 ( 2000 ) .
t. d. ladd , f. jelezko , r. laflamme , y. nakamura , c. monroe , and j. l. obrien , nature * 464 * , 45 ( 2010 ) .
a. beige , phys .
rev .
a * 69 * , 012303 ( 2004 ) .
b. kraus , h. p. bchler , s. diehl , a. kantian , a. micheli , and p. zoller , phys .
rev .
a * 78 * , 042307 ( 2008 ) .
f. ticozzi and l. viola , ieee t. automat . contr . * 53 * , 2048 ( 2008 ) . s. diehl , a. micheli , a. kantian , b. kraus , h. p. bchler , and p. zoller , nat . phys . * 4 * , 878 ( 2008 ) .
f. verstraete , m. m. wolf , and j. i. cirac , nat . phys . * 5 * , 633 ( 2009 ) .
j. f. poyatos , j. i. cirac , and p. zoller , phys .
rev .
lett . * 77 * , 4728 ( 1996 ) .
m. b. plenio , s. f. huelga , a. beige , and p. l. knight , phys .
rev .
a * 59 * , 2468 ( 1999 ) .
a. beige , d. braun and p. l. knight , new j. phys .
* 2 * , 22 ( 2000 ) .
x. t. wang and s. g. schirmer , arxiv:1005.2114 .
m. j. kastoryano , f. reiter , and a. s. srensen , phys .
rev .
lett . * 106 * , 090502 ( 2011 ) .
m. gullans , t. g. tiecke , d. e. chang , j. chang , j. feist , j. d. thompson , j. i. cirac , p. zoller , and m. d. lukin , phys .
rev .
lett . * 109 * , 235309 ( 2012 ) . a. w. carr and m. saffman , phys .
rev .
lett . * 111 * , 033607 ( 2013 ) .
m. j. a. schuetz , e. m. kessler , l. m. k. vandersypen , j. i. cirac , and g. giedke , phys .
rev .
lett . * 111 * , 246802 ( 2013 ) .
h. krauter , c. a. muschik , k. jensen , w. wasilewski , j. m. pedersen , j. i. cirac , and e. s. polzik , phys .
rev .
lett . * 107 * , 080503 ( 2011 ) .
j. t. barreiro , m mller , p. schindler , d. nigg , t. monz , m. chwalla , m. hennrich , c. f. roos , p. zoller , and r. blatt , nature * 470 * , 486 ( 2011 ) .
y. lin , j. p. gaebler , f. reiter , t. r. tan , r. bowler , a. s. srensen , d. leibfried , and d. j. wineland , _ nature _ * 504 * , 415 ( 2013 ) .
s. shankar , m. hatridge , z. leghtas , k. m. sliwa , a. narla , u. vool , s. m. girvin , l. frunzio , m. mirrahimi , and m. h. devoret , nature * 504 * , 419 ( 2013 ) . s. schneider and g. j. milburn , phys .
rev .
a * 65 * , 042107 ( 2002 ) .
g. vacanti and a. beige , new j. phys .
* 11 * , 083008 ( 2009 ) .
h. weimer , m. mller , i. lesanovsky , p. zoller , and h. p. bchler , nat . phys .
* 6 * , 382 ( 2010 ) .
k. g. h. vollbrecht , c. a. muschik , and j. i. cirac , phys .
rev .
lett . * 107 * , 120502 ( 2011 ) .
j. cho , s. bose , and m. s. kim , phys .
rev .
lett . * 106 * , 020504 ( 2011 ) .
r. n. stevenson , j. j. hope , and a. r. r. carvalho , phys .
rev .
a * 84 * , 022332 ( 2011 ) .
m. j. kastoryano , d. reeb , and m. m. wolf , j. phys .
a * 45 * , 075307 ( 2012 ) .
m. foss - feig , a. j. daley , j. k. thompson , and a. m. rey , phys .
. lett . * 109 * , 230501 ( 2012 ) .
a. gonzalez - tudela and d. porras , phys . rev . lett . *
110 * , 080502 ( 2013 ) . c. cormick , a. bermudez , s. f. huelga , and m. b. plenio , new .
j. phys .
* 15 * , 073027 ( 2013 ) .
m. hning , d. muth , d. petrosyan , and m. fleischhauer , phys .
rev .
a * 87 * , 023401 ( 2013 ) .
s. k. lee , j. cho , and k. s. choi , new j. phys .
* 17 * , 113053 ( 2015 ) .
f. ticozzi and l. viola , quantum inf . comput . * 14 * , 0265 ( 2014 ) .
d. d. b. rao and k. mlmer , phys .
rev .
a * 90 * , 062319 ( 2014 ) .
h. pichler , t. ramos , a. j. daley , and p. zoller , phys .
rev . a * 91 * , 042116 ( 2015 ) .
t. e. lee , f. reiter , and n. moiseyev , phys .
rev .
lett . * 113 * , 250401 ( 2014 ) .
w. happer , rev .
mod .
phys . * 44 * , 169 ( 1972 ) .
see supplementary information for the detailed analysis .
f. reiter and a. s. srensen , phys .
rev .
a * 85 * , 032111 ( 2012 ) .
t. p. harty , d. t. c. allcock , c. j. ballance , l. guidoni , h. a. janacek , n. m. linke , d. n. stacey , and d. m. lucas , phys .
rev .
lett . * 113 * , 220501 ( 2014 ) .
s. puntanen and g. styan , _ the schur complement and its applications _ , ( springer , new york , 2005 ) .
s. boyd and l. vandenberghe , _ convex optimization _ ( cambridge university press , cambridge , 2004 ) [ [ section ] ] |
stimulated raman adiabatic passage ( stirap ) resulting in the population transfer between the states which are not directly coupled @xcite , electromagnetically induced transparency ( eit ) via adiabatic following of the dark , non - absorbing state @xcite , nonresonant pulse excitation of the two - level atom @xcite are just a short list of phenomena in quantum optics where adiabatic processes are considered .
generally , one can find in any part of physics problems concerned with adiabaticity , which are treated almost similarly . among them we can mention multiphoton resonances induced in atoms and molecules by a strong low - frequency field @xcite-@xcite , wave - packet dynamics in physics and chemistry or , so - called , femto - chemistry @xcite-@xcite and slow atomic and molecular collisions @xcite-@xcite .
a fully adiabatic process takes place if the hamiltonian of the problem is diagonal .
then any time dependence of this diagonal hamiltonian does not change the state of the quantum system except its phase .
in general , the hamiltonian , describing the interacting systems , has both diagonal and nondiagonal components . in most cases dealing with adiabatic excitation , the problem can be reduced to the consideration of the two - level system perturbed by the diagonal and nondiagonal interactions .
then , both components of the hamiltonian , i.e. , the two - level splitting and the nondiagonal perturbation , are time dependent .
one can find instantaneous eigenstates of such a hamiltonian .
these states are called adiabatic states if the system , being initially in one of them , follows this state .
in contrast , the initial states , where the transition takes place , are called diabatic states .
the diagonalization of the instantaneous hamiltonian takes into account the diabatic states coupling , incorporating the interaction parameter into the instantaneous eigenvalues .
however , a new coupling appears again because of the time dependence of the transformation from the initial , diabatic basis to the adiabatic basis ( see , for example , @xcite ) .
inasmuch as the coupling between instantaneous eigenstates does not vanish , the following of the adiabatic states can never be perfect .
since it is almost impossible to get rid of the state coupling by the time dependent transformation of the basis functions , the choice of the adiabatic basis is aimed to minimize the transition probability between the states , almost approaching in this sense to the fully adiabatic hamiltonian ( i.e. , the diagonal one ) when the coupling can be neglected .
the estimation of the fraction of the system that is transferred from one adiabatic state to another during the interaction ( nonadiabatic correction ) is a main problem in this approach .
this nonadiabatic correction is a measure of the nonadiabaticity of the process .
it should be obvious that the searching for the adiabatic basis is an approximate approach .
it was developed for the first time to treat the behavior of the two - level system subject to an excitation that couples the states if both the two - level splitting and the excitation are time dependent .
only a few models of such a two - level system have analytical solutions . among them
are the rosen zener model with @xmath0 and @xmath1 , the demkov
kunike model with @xmath2 and @xmath3 or @xmath4 , the landau zener model with @xmath5 and @xmath6 , and the parabolic model with @xmath7 and @xmath6 .
although the last model has not an exact solution , it has been studied comprehensively ( see , for example , ref .
@xcite ) . here , @xmath8 is the amplitude of the diagonal component of the hamiltonian ( i.e. , the two - level splitting ) and @xmath9 is the amplitude of the nondiagonal component ( i.e. , the coupling perturbation ) .
the symbols @xmath10 , @xmath11 , @xmath12 , @xmath13 can represent arbitrary constants and @xmath14 is time .
one can find these models and their solutions in , for example , suominen95 .
the advantage of these models comes from the possibility to reduce their schrdinger equations to second order differential equations ( with time dependent coefficients ) whose solutions are known .
for example , the schrdinger equation for the rosen - zener model can be reduced to an equation whose solution is the hypergeometric function and the schrdinger equation for the landau - zener model can be solved in terms of parabolic cylinder functions . for an arbitrary hamiltonian ,
the quasi - classical approach based on the wentzel - kramers - brillouin ( wkb ) approximation was developed by dykhne @xcite@xcite . in this method
, the amplitude of the nonadiabatic transition is calculated under the assumption that the adiabatic levels cross in the complex time plane and the main contribution of the nonadiabatic coupling is given at the time of the level crossing .
the latter is usually the branch point of the phase integral of the energies .
the result becomes universal , i.e. , independent of the nonadiabatic coupling model and it is sufficient to take the leading term of the coupling at the crossing point .
in this paper we consider adiabatic following of the dark , nonabsorbing state in the three - level system excited by two resonant fields .
the adiabatic following of the dark state results in stirap ( population transfer ) .
it was proposed in ref .
@xcite where the so called counterintuitive raman pulse sequence emerged as a result of the search for the generalization of the liouville equations for the @xmath15-level system using @xmath16 coherence vector theory @xcite@xcite .
this pulse sequence consists of two fields coupling two low energy levels @xmath17 and @xmath18 with one common excited state @xmath19 .
if one of the levels ( for example , @xmath18 ) is initially empty , then applying first the field which couples this state with level @xmath19 , and then the field coupling the populated state ( for example , @xmath17 ) with @xmath19 , it is possible to transfer the population of state @xmath17 to state @xmath18 , without appreciably populating the intermediate state @xmath19 . later , the importance of the adiabaticity in the counterintuitive pulse sequence development was realized in @xcite carrhioe90 .
there is a particular superposition of states @xmath17 and @xmath18 that does not interact with the coupling fields and , if the development of the field amplitudes in time is properly chosen , this superposition state changes from state @xmath17 to state @xmath18 .
if the three - level atom follows this superposition state , the atom population is transferred from state @xmath17 to state @xmath18 by the pulse sequence without populating the intermediate excited state @xmath19 .
this noncoupled superposition state was first introduced by arimondo @xcite to explain qualitatively the dark resonance as population trapping in this state @xcite@xcite .
the noncoupled state is often referred to as a dark state . it is obvious that , if the dark state changes in time , a process must exist which tends to empty this state .
the condition minimizing the dark state depopulation is formulated in @xcite .
this is done _ ad hoc _ , without an estimation of the nonadiabatic correction for the excited probability amplitude .
however , numerical calculations show that , if this condition is satisfied , the adiabatic population transfer @xmath20 is almost perfect .
some attempts were undertaken to find a rigorous justification of the intuitively found adiabaticity condition and to estimate the amplitude of the excited state during the stirap pulse sequence ( see , for example , refs .
@xcite@xcite ) . in stenholm96 ,
the so called ramp pulses were considered , which allow an exact solution .
furthermore , with the help of the method developed for the two - level system by dykhne ( see the discussion above and refs .
pechukas76@xcite ) , the nonadiabatic amplitude of the excited state @xmath19 is estimated for gaussian and hyperbolic secant pulses .
this is possible because the liouville equation for the two - level atom in terms of the @xmath21 coherence vector ( bloch - vector model , see , for example , alleneber75 ) coincides with the schrdinger equation for the state probability amplitudes of the three - level atom excited by two resonant fields ( see , for example , @xcite , @xcite , shakhmuratov2001 ) .
however , the authors of @xcite admit that the analytical approximations for the nonadiabatic corrections _ have been introduced ad hoc without derivation _ and they _ would really like to see more detailed investigations of the analytic behavior of the system discussed _ .
fleischhauer with coauthors @xcite developed a different approach introducing higher - order trapping states .
they define a @xmath22th order generalized adiabatic basis , which is similar to the superadiabatic basis introduced for the two - level system in @xcite@xcite . by successive transformations from the diabatic basis to the adiabatic basis , then from this adiabatic basis , which can be considered as the adiabatic basis of the first order , to the next , i.e. , the second order adiabatic basis , etc . , the solution is presented as an infinite product of transformations .
the general expression for the @xmath22-th transformation matrix is presented . however , it is hard to implement this scheme for an arbitrary pulse sequence . in this paper
we develop a new method that allows us to calculate the adiabatic and nonadiabatic components of the solution with controllable accuracy .
the adiabatic component describes the part of the atomic probability amplitude visiting the excited atomic state during the pulse train and coming back to the dark state when the pulses are gone .
the atom evolution following the adiabatic component of the solution resembles the excitation de - excitation process induced by a soliton in the two - level atom .
the nonadiabatic component is that part which is lost from the dark state and describes the fraction of the atomic probability amplitude that is left excited after the pulse train .
we compare our result with the previous theories reported in @xcite and @xcite .
we found corrections to the theory presented in @xcite and show that our result corresponds to the calculation of the infinite number of transformations proposed in @xcite .
as an example of the validity of the method , we present also the solution of the rosen zener and landau - zener models by our method ( see the appendix and the end of sections vi and vii ) .
the paper is organized as follows . in sec .
ii we present the general formalism employed in the description of the three - level atom excited by two resonant fields .
the transformation to the basis of the bright and dark states is derived .
it is shown that the system evolves between two states , i.e. , bright and common
( they are specified in section ii ) . in sec .
iii we consider the time evolution of the atomic state vector for the @xmath23-scheme of excitation if the field amplitudes are time dependent .
bloch - like equations are derived . in sec .
iv the adiabatic following approximation is presented .
the adiabatic solution for the stimulated raman adiabatic passage ( stirap ) is found in sec .
v. nonadiabatic corrections are described in sec .
the case when the raman rabi frequency changes in time is considered in sec .
we consider a three - level atom shown in fig .
1 ( a ) . the arrows indicate the transitions induced by the coherent fields .
this excitation scheme is classified as the @xmath24 scheme .
we define the level that is common for both transitions as @xmath19 .
the others are designated by the numbers @xmath17 and @xmath18 , level @xmath18 being of higher energy than @xmath17 and initially not populated .
the dynamic part of the hamiltonian of this three - level atom , excited by two resonant fields @xmath25 and @xmath26 , is @xmath27where @xmath28 is the energy of the state @xmath22 .
planck s constant is set equal one ( @xmath29 ) for simplicity .
the operators @xmath30 are defined by @xmath31 , where @xmath32 and @xmath33 are bra and ket vectors of the states @xmath22 and @xmath34 in the schrdinger representation .
the interaction constant ( rabi frequency ) @xmath35 depends on the dipole - transition matrix element , taken real so , @xmath36 , and on the field amplitude @xmath37 .
the rotating wave approximation is taken into account . if the fields @xmath38 and @xmath39 are in exact resonance with the relevant transitions , the hamiltonian , eq .
( [ eq1 ] ) , can be made slowly varying by transforming it into the interaction representation ( ir ) .
this representation is defined by a canonical transformation by means of the unitary operator @xmath40the wave function of the atom in the ir is defined by @xmath41 , where @xmath42 is the wave function in the schrdinger representation .
@xmath43 satisfies the schrdinger equation with the effective hamiltonian @xcite @xmath44this hamiltonian has the explicit form @xmath45or in matrix notation @xmath46 .
\label{eq5}\]]the @xmath47-operators are defined for the vectors @xmath48 and @xmath49 of the interaction representation differing from the states @xmath33 and @xmath50 by the phase factors @xmath51 and @xmath52 .
assume that the field amplitudes @xmath53 , @xmath54 and the phases @xmath55 , @xmath56 are constant .
then the hamiltonian , eqs .
( [ eq4])([eq5 ] ) , is diagonalized by a unitary transformation @xmath57where @xmath58 .
\label{eq7}\]]and @xmath59 .
the new ket vectors @xmath60 are related to the former ones @xmath61 , defined in the interaction representation , as follows @xmath62 .
the explicit relations are @xmath63@xmath64@xmath65the basis @xmath66 , in which the hamiltonian is diagonal , is called the basis of the quasi - energy states @xcite , shakhmuratov77 .
it coincides with the basis of the dressed states if the limit of the classical field is taken for them .
states @xmath67 and @xmath68 are mixtures of all unperturbed states @xmath69 , @xmath70 and @xmath71 , whereas state @xmath72 does not contain the common state @xmath71 . if the atom is in state @xmath73 , a mixture of the ground state sublevels @xmath17 and @xmath18 , the atom does not leave this state since it is an eigenstate of the interaction hamiltonian , eq .
( [ eq6 ] ) .
the bichromatic field @xmath74 does not interact with such an atom , and , consequently , it is not excited
. therefore , we call this state a dark state and designate it @xmath75 .
this state was introduced for the first time by arimondo in ref .
@xcite , who called this state a non - coupled state .
he introduced the coupled state as well , which interacts with the bichromatic field @xmath76 .
following arimondo we define the state @xmath77which is orthogonal to the dark state .
we call this state a bright state , since for the @xmath24-scheme , if the atom is in this state , it is excited by the bichromatic field and then the luminescence from state @xmath78 may follow .
excitation can take place because the bright state is not an eigenstate of the interaction hamiltonian , eq .
( [ eq6 ] ) .
the states @xmath75 , @xmath79 and @xmath78 are mutually orthogonal and can be chosen as a new basis .
we call this basis @xmath80 , designating the state @xmath71 by the letter @xmath81 , since it is common for the induced transitions .
because we will often refer to these states , they are presented below in a common set of equations to simplify further citation @xmath82@xmath83@xmath84the interaction hamiltonian , eqs .
( [ eq4])([eq5 ] ) , is transformed in this basis as follows @xmath85or in a matrix form @xmath86 , \label{eq16}\]]where @xmath87 and the unitary operator of the canonical transformation is @xmath88 .
\label{eq17}\]]we call this basis the @xmath80-representation .
figure 1 ( b ) shows the excitation scheme in the @xmath80basis where only the @xmath89 and @xmath90 states are coupled .
the @xmath91 and @xmath90 states do not correspond to a defined energy . to present
schematically the @xmath80 states and the transitions between them , we choose the state population before the excitation for their relative position in the diagram .
so , the vertical scale in fig .
1 ( b ) represents the initial population of the states counted from the bottom to the top .
the hamiltonian in the @xmath80-basis , eq .
( [ eq15 ] ) , resembles the interaction hamiltonian of the two - level system @xmath92 , excited by one resonant field with an effective interaction constant @xmath93 .
the dynamic evolution of the atom is described by the schrdinger equation with this hamiltonian containing only the transition operators between states @xmath91 and @xmath94 . if , for example , initially the @xmath95 and @xmath91 states are populated and @xmath90 is not , @xmath96the dynamics of the three - level atom follows the very simple formula @xmath97 + c_{d}\left| d\right\rangle , \label{eq19}\]]where @xmath98 is the raman rabi frequency .
this rabi frequency is defined as twice the interaction constant in the hamiltonian , eq .
( [ eq15 ] ) , according to the conventional definition adopted for the two - level system @xcite . the initial condition specified in eq .
( [ eq18 ] ) has a simple relation to the ir basis @xmath99if the initial condition is @xmath100 , then @xmath101 and @xmath102 , which means that the phase relation between the amplitudes of @xmath91 and @xmath95 states is fixed as @xmath103 .
suppose that the pulse duration of the bichromatic field , @xmath104 , is chosen such that the effective two - level system @xmath92 experiences a so called @xmath105-pulse , i.e. , @xmath106 . during the first half of the pulse , the population of level @xmath91 is completely transferred to state @xmath90 . during the second half of the pulse
, the population of the excited state @xmath107 is transferred back to state @xmath91 .
the state vector of the atom at the end of the pulse is @xmath108equation ( [ eq21 ] ) shows that the @xmath105-pulse results in a @xmath109 phase shift of the @xmath91 state probability amplitude .
usually , for the simple two - level atom , this @xmath109 phase shift of the ground state after a @xmath105-pulse is not revealed in the observables . in our case , apart from the two levels @xmath91 and @xmath90 , we have the third level @xmath95 whose phase is well defined with respect to the phase of state @xmath91 .
the phase definition is determined by the phase relation of the components of the bichromatic field [ see comments after eq .
( [ eq20 ] ) ] . in section
iv we show the importance of this point .
if during the pulse excitation the field amplitudes @xmath110 , @xmath111 are time dependent and they have the same time evolution satisfying the condition @xmath112 , then the developments of the @xmath75 , @xmath79 states in the ket - vectors @xmath113 , @xmath70 have constant coefficients and the @xmath114-transformation is time independent . in this case
the solution of the schrdinger equation is simple and the development coefficients of the state vector @xmath115 , @xmath116are @xmath117 , @xmath118 $ ] and @xmath119 $ ] , where @xmath120 , @xmath121 are the initial values of the probability amplitudes and @xmath122 is the pulse area of the bichromatic field .
this case corresponds to so called matched pulses @xcite@xcite . if there is a time shift @xmath123 between the pulses @xmath38 and @xmath124 , one can again introduce the bright and dark states , employing the time dependent @xmath114-transformation .
then the evolution of the state vector of the atom in the @xmath80-basis is given by @xmath125where @xmath43 is the state vector in the interaction representation . taking the time derivative of the equation ( [ eq23 ] ) , one can obtain the schrdinger equation @xmath126with the modified hamiltonian @xmath127where @xmath128 is defined in eq .
( [ eq15 ] ) and @xmath129 \label{eq26}\]]the first part , @xmath128 , of the modified hamiltonian induces transitions between the @xmath79 and @xmath81 states with the rate @xmath98 , and the second part , @xmath130 , induces transitions between states @xmath75 and @xmath131 with the rate @xmath132 ( see fig .
the development coefficients of the state vector @xmath133 satisfy the equations @xmath134@xmath135@xmath136where the substitution @xmath137 , @xmath138 , @xmath139 is made to deal with real numbers . here the phases @xmath140 and @xmath56 are assumed to be constant throughout the excitation process .
equations ( [ eq27])([eq29 ] ) remarkably coincide with the bloch equations for an abstract two - level system @xmath141 excited by a field with frequency @xmath142 slightly tuned from resonance ( @xmath143 and @xmath144 being ground and excited states , respectively ) @xcite .
expressed in terms of the bloch - vector components , which are the following combinations of the density matrix of the two - level system @xmath145 : @xmath146 and @xmath147 , these equations are , ref .
@xcite , @xmath148@xmath149@xmath150here @xmath151 is the detuning from resonance and @xmath152 is the rabi frequency . if one makes the substitution @xmath153 , @xmath154 , @xmath155 for the variables and @xmath156 , @xmath157 for the parameters , both sets of equations are identical .
the coincidence of the equations enables us to use the adiabatic following approach developed by crisp @xcite to describe nonresonant excitation of the two - level system .
a similar approach was applied by laine and stenholm ( ls ) @xcite based on the ideas of the adiabatic following developed by dykhne @xcite,@xcite for the two - level system with time dependent splitting and coupling parameters . in the ls approach,@xcite ,
the instantaneous eigenstates of the three - level atom excited by two resonant pulses are found .
this basis is called the adiabatic representation .
the transformation to the instantaneous basis [ see eq .
( [ eq7 ] ) ] and the instantaneous hamiltonian diagonal in this basis [ see eq .
( [ eq6 ] ) ] are time dependent
. therefore , the adiabatic states coupling , @xmath158 , also appears in their consideration . in spite of being different in structure , our schrdinger equation in the changing @xmath80-basis and the ls equation in the instantaneous eigenstate basis
can both be reduced to the equation for the two - level system . however , our equations ( [ eq27])([eq29 ] ) are in a one to one corespondence to the bloch equations while the ls equations match the equations for the two - level density matrix elements @xmath159 , @xmath160 and @xmath161 .
although this difference is not crucial , it brings , however , some convenience in our case , because our equations are expressed for real quantities .
crisp gave in his paper @xcite a qualitative and quantitative description of the adiabatic following approximation for the two - level atom excited by a nonresonant field .
the qualitative explanation employs the bloch - vector model as follows .
the bloch equations can be written in the form @xmath162if the relaxation times @xmath163 and @xmath164 are very long compared with the pulse duration . the bloch vector @xmath165 is given by@xmath166and the effective field @xmath167 is @xmath168equation ( [ eq33 ] ) has the geometric interpretation that the bloch vector tries to precess about the effective field as @xmath167 varies both in magnitude and direction .
if the condition @xmath169 is satisfied , the precession frequency of the bloch - vector , @xmath170 , will be large compared with the rate of change of the field @xmath171 .
if initially the bloch - vector points parallel or antiparallel to @xmath172 [ @xmath173 and @xmath174 , then precessing with high frequency @xmath175 about the effective field , the bloch - vector remains nearly parallel or antiparallel to the effective field as it moves adiabatically .
qualitatively , adiabatic movement of the vector @xmath167 means that the bloch - vector does not slip off from the direction of the effective field if the rotation of the bloch - vector around the @xmath167field is much faster than the change of the @xmath167field direction .
quantitatively , this sets a condition that the precession period @xmath176 is to be much shorter than the time scale of the effective field variation . in this limit
, the angle between the bloch vector and the effective field will remain very small .
the deviation of the bloch vector from the effective field is described by the nonadiabatic corrections . in the next section
we consider how small this deviation angle is for particular pulses .
the approximation of the adiabatic following allows also a quantum mechanical interpretation .
the hamiltonian in the @xmath80 time varying representation , eq .
( [ eq25 ] ) , consists of two parts .
one part couples two , initially unpopulated states , @xmath91 and @xmath90 .
the coupling parameter @xmath93 is time varying .
the other part couples the populated state @xmath95 with the unpopulated state @xmath91 .
the coupling parameter @xmath177 is also time varying . in order that the atom is subject to the adiabatic following of the changing @xmath95 state (
the atom stays in this state ) , one has to choose the pulse train with the raman rabi frequency @xmath93 much larger than @xmath177 when the latter takes its maximum value shakhmuratov2001 .
also the time variation scale of @xmath177 must be much longer than the raman rabi period .
these conditions can be made plausible by the following speculations .
since initially only state @xmath95 is populated , the atom starts to evolve when @xmath178 becomes nonzero . when the slow process @xmath179 brings a small fraction of the population to state @xmath91 , the fast rabi oscillation via the @xmath105-cycle ( see the discussion at the end of section ii ) changes the sign of the @xmath180-amplitude from plus to minus , reversing back the population transfer process from @xmath181 to @xmath182 . as a result ,
the population of the @xmath91-state does not increase and remains small .
one can roughly estimate its mean value by solving the equations ( [ eq27])([eq29 ] ) with the initial condition @xmath183 , @xmath184 and assuming that @xmath185 and @xmath186 are constant , i.e. , neglecting the derivatives of these parameters , @xmath187@xmath188@xmath189the solution shows that the change of the amplitude of the @xmath95-state is small if @xmath190 .
this is exactly the condition of the adiabatic following formulated in ref .
kuklgaubhioebergm89 for stirap .
it should be emphasized that the higher time derivatives of the mixing parameter @xmath191 , neglected above , play a crucial role in the adiabatic population transfer .
however , the solution ( [ eq36])(eq38 ) gives some qualitative hint on the choice of the relation between @xmath93 and @xmath177 .
the adiabatic following approximation can be applied to the consideration of stirap and electromagnetically induced transparency ( eit ) because in both cases the atom follows the dark , noncoupled state . a first attempt to study adiabatic following in eit
was undertaken in ref .
@xcite . in this section
we consider the application of this approach to stirap .
the stimulated raman adiabatic passage assumes that before the application of the @xmath192 and @xmath193 pulses the atom is in state @xmath69 .
the duration of the excitation as well as the pulse sequence must be chosen such that at the end of the pulse train the atom is left in state @xmath70 .
it is expected that during this process the atom stays in the dark state . however , this state itself changes since the coefficients @xmath194 and @xmath195 of the development of the dark state in the vectors @xmath113 and @xmath196 [ see eq .
( [ eq12 ] ) ] change in time .
the parameter @xmath197 rises from zero to @xmath198 , so @xmath199 before the excitation and @xmath200 after it . since @xmath201 ,
the condition imposed on @xmath202 means that the @xmath203-pulse must be delayed with respect to the @xmath204-pulse .
we have to emphasize that the phase of the final state @xmath18 after the pulse train is related to the phase of the initial state @xmath17 according to eq .
( eq12 ) .
if we choose two identical , bell - shaped , delayed pulses having a hyperbolic secant shape , then the interaction constants evolve in time as follows @xmath205 , \label{eq39}\]]where @xmath206 or @xmath18 , @xmath207 is the time when the @xmath22-pulse has maximum amplitude , and @xmath13 is the rise and fall rate of the pulse edges .
the mixing parameter @xmath191 increases monotonously if the condition @xmath208 is satisfied .
this pulse sequence was considered by laine and stenholm in ref .
@xcite .
let us analyze the constraints imposed on the parameters @xmath13 , @xmath209 and @xmath210 to have @xmath202 changed from zero to @xmath198 .
the mixing angle @xmath197 varies according to the relation @xmath211 . for the pulse sequence specified above , an explicit form of this relation
is @xmath212}{1-d\tanh [ r(t - t_{0 } ) ] } , \label{eq40}\]]where @xmath213 , @xmath214 is the mean time of the maxima and @xmath215 is the time interval between the maxima of the pulses .
suppose that , initially , the atom is in the ground state @xmath17 and we start the atom evolution from the initial mixing parameter @xmath216 satisfying the condition @xmath217 , which means that essentially @xmath218 and @xmath219 .
we stop the atom evolution at @xmath220=100 $ ] .
so , if the final state coincides with state @xmath18 then @xmath221 and @xmath222 , where the phase change of state @xmath18 is taken into account . from eq .
( [ eq40 ] ) it follows that @xmath223 and at the condition imposed on @xmath224 and @xmath225 we have @xmath226 . since state @xmath91
is strongly coupled with state @xmath90 [ the coupling is @xmath227 , we have to keep the initial population of state @xmath91 as small as possible [ @xmath228 must be close to zero ] .
otherwise , the probability amplitude @xmath229 , if not infinitely small , will be spread among the @xmath91 and @xmath90 states by the pulse train and population transfer @xmath230 will be imperfect .
therefore , on the one hand , the initial value of the mixing parameter @xmath197 must be as small as possible to have complete population transfer @xmath231 .
for a small initial mixing angle @xmath224 , the relation between @xmath224 and @xmath123 becomes simple , i.e. , @xmath232 . the smaller the initial value of the mixing angle @xmath202 ,
the larger the product @xmath233 or the pulse separation @xmath123 is . on the other hand , if the distance between the pulses is large , the value of the coupling @xmath234 at @xmath235 becomes small : the larger the distance , the smaller the coupling .
however , the adiabatic following demands a large coupling @xmath93 at @xmath235 .
therefore , one has to choose the optimum value of the pulse spacing satisfying two conditions simultaneously .
the distance between pulses is to be as large as possible to have a small value of @xmath236 and , at the same time , @xmath237 should be kept as large as possible .
below we give some arguments how to find this optimum value . if @xmath238 , then according to eq . ( [ eq40 ] )
we have @xmath239 throughout the excitation and the dark state does not change in time .
it has the probability amplitude @xmath240 .
state @xmath91 , having the initial population @xmath241 , is depopulated with the rate @xmath93 [ @xmath242 , @xmath243 , see the definition of @xmath244 in section iii ] . for small values of @xmath233 ,
the time interval of the @xmath202 change is small .
for example , if @xmath245 , then @xmath246 , @xmath247 and @xmath248 , @xmath249 , which corresponds to the change of @xmath202 from @xmath250 to @xmath251 during the pulse train . in this case
only the fraction @xmath252 of the atomic population is transferred to state @xmath70 if the atom
adiabatically follows the dark state .
another fraction @xmath253 of the atomic population participates in the process of excitation and de - excitation between states @xmath91 and @xmath90 .
this means that the population transfer via the change of the amplitude of the dark state components is possible only if the time interval between the pulses exceeds a certain value .
we choose the value @xmath254 since in this case the initial population of state @xmath91 is @xmath255 so that we neglect this population in our further consideration .
figure 3 ( a ) shows the pulse train with @xmath254 along with the change of the mixing parameter @xmath202 during the excitation . on the same plot the dependence of the derivative @xmath256}{1+d^{2}\tanh ^{2}% \left [ r\left ( t - t_{0}\right ) \right ] } = \frac{r\sinh ( rt)}{1+\cosh ( rt)\cosh [ 2r(t - t_{0 } ) ] } \label{eq41}\]]is shown .
this derivative takes a maximum value @xmath257 at @xmath258 .
for example , if @xmath254 , then @xmath259 . in fig . 3 ( b ) we compare the time evolution of the rabi frequency @xmath260 with the evolution of the derivative @xmath261 .
the rabi frequency determines the transition rate between states @xmath91 and @xmath90 , whereas the derivative of the state mixing angle @xmath202 specifies the transition rate between states @xmath95 and @xmath91 [ see eqs .
( [ eq27])(eq29 ) ] .
since at @xmath258 the parameter @xmath261 takes its maximum value and @xmath234 has its minimum , the adiabatic following condition is @xmath262 or , explicitly , @xmath263 .
figure 4 ( solid lines ) shows the time dependence of the amplitudes @xmath264 ( plot a ) , @xmath180 ( plot b ) and @xmath265 ( plot c ) obtained by the numerical solution of the equations ( eq27)([eq29 ] ) for the pulse train with @xmath266 , @xmath267 , @xmath268 , @xmath269 , @xmath270 .
the relation between @xmath271 and @xmath13 corresponds to the ratio @xmath272 .
we see that the deviation of the @xmath265amplitude from its initial value ( @xmath273 ) and the maximum absolute value of the @xmath180amplitude ( @xmath274 ) are different from the values predicted by the simplified solution , eqs .
( eq36)([eq38 ] ) .
only the maximum value of @xmath264 ( @xmath275 ) fits the value of the simplified solution .
to estimate the actual values of the probability amplitudes of the @xmath80 states during the excitation and find the borders within which the adiabatic following of the dark state takes place , we follow the theory developed by crisp @xcite for the case when the condition @xmath276 is well satisfied . in this case
we can use the expansion in a power series of the parameter @xmath178 for the solution of the equations ( [ eq27])-([eq29 ] ) .
then the first term of the expansion is found by setting @xmath277 in the equations ( [ eq28]),([eq29 ] ) and then solving them .
the solution is@xmath278 \stackrel{.}{\alpha } ( \tau ) d\tau , \label{eq42}\]]@xmath279 \stackrel{.}{\alpha } ( \tau ) d\tau .
\label{eq43}\]]this is the part of the general solution that is linear in @xmath280 .
@xmath265 satisfies the equation @xmath281which is the formal solution of eq.([eq27 ] ) .
therefore the change of the dark state amplitude @xmath265 is nonlinear in @xmath261 and can be presented in products of @xmath177-s .
the first contribution of @xmath261 to @xmath265 can be found as the square of @xmath177 , i.e. , in the second term of the expansion . substituting the corrected @xmath265 into eqs .
( [ eq28]),(eq29 ) ( instead of @xmath282 ) , one can find the next term in the expansion of @xmath283 and @xmath284 .
then the substitution of the thus found @xmath283 into eq .
( [ eq44 ] ) gives the next term of the expansion of @xmath265 , etc . in this paper
we consider only the linear corrections to @xmath180 and @xmath264 .
our solution , given by the equations ( [ eq42])([eq44 ] ) , differs from crisp s solution in the @xmath93-time dependence .
he considered the case when @xmath93 is constant and @xmath177 is time dependent .
4 ( a c ) shows the comparison of the numerical solution of the equations ( eq27)([eq29 ] ) with eqs .
( [ eq42])([eq44 ] ) .
they areindistinguishable and shown by the same solid lines .
we find the adiabatic and nonadiabatic components of the analytical solution by applying two different procedures .
the adiabatic components of the equations ( [ eq42])([eq43 ] ) are calculated by integrating them by parts .
for example , the first step of the @xmath285 calculation is@xmath286 \right ) = \frac{\stackrel{.}{\alpha } } { b}-\int\nolimits_{-\infty } ^{t}\cos % \left [ \int_{\tau } ^{t}b(\tau _ { 1})d\tau _ { 1}\right ] \left ( \frac{\stackrel{.% } { \alpha } } { b}\right ) _ { \tau } ^{\prime } d\tau .
\label{eq45}\]]where @xmath287 is the derivative with respect to @xmath288 .
repeating these steps several times , we obtain@xmath289@xmath290where @xmath291 , @xmath292 and it is assumed that @xmath261 is a bell - shaped function with first and higher derivatives equal zero at @xmath293 . to find a similar expansion of equation ( [ eq44 ] ) , we introduce the function @xmath294 .
then the equation for @xmath295 takes the form@xmath296with@xmath297@xmath298the functions @xmath299 and @xmath300 are reduced to the single integrals@xmath301@xmath302this can be done since they have the structure@xmath303where @xmath304 is @xmath305and the index @xmath90 stands for the cosine function and index @xmath306 for sine . integrating these integrals by parts
, we obtain@xmath307 , \label{eq55}\ ] ] equations ( [ eq46],[eq47 ] ) and ( [ eq55 ] ) give the probability amplitudes of the dark , bright and common states .
they coincide with those one obtains if the successive transformations @xmath308 to the new set of @xmath309 -states are performed , as was done by fleischhauer and coauthors in ref .
@xcite , i.e. , @xmath310 , where @xmath311 is the initial state [ @xmath312 , @xmath313 . the transformation @xmath114 is defined in eq .
( [ eq17 ] ) and the other transformations ( @xmath314 ) are specified below .
these transformations have a simple meaning .
since the @xmath315 and @xmath316 transitions are excited simultaneously by the @xmath177 and @xmath93 `` fields '' , one can make a @xmath317-transformation to the new set of dark , bright and common states , @xmath318 , @xmath319 and @xmath320 , where @xmath318 is a particular mixture of the former @xmath95 and @xmath90 states , @xmath319 is a state orthogonal to state @xmath321 , and @xmath320 coincides with state @xmath91 .
the new mixing angle is @xmath322 .
repeating this procedure @xmath22-times , one can get our solution if the conditions @xmath323 and @xmath324 are applied at each step .
figure 4 ( a - c ) shows the comparison of the numerical solution ( solid lines ) with the expansions given by the equations ( [ eq46 ] ) , ( [ eq47 ] ) and ( [ eq55 ] ) ( dashed lines ) . the parameters of the pulse train are specified above and they are the same as in fig .
3 ( a , b ) .
only the first two terms of the expansions are taken into account for each plot , which is justified because @xmath325 .
for the @xmath264 and @xmath180 components , the maximum absolute values of the second terms of the expansions ( i.e. , the adiabatic terms ) are already comparable with the amplitude of the oscillations ( i.e. , the nonadiabatic contribution , coming from the summation of an infinite number of the expansion terms ) . of course _ , nonadiabatic oscillations are not described by the main part of the adiabatic solution presented by a few leading terms of the expansion_. concluding this section , we refer to a particular case when @xmath234 and @xmath186 have the same time dependence .
this is again the case of matched pulses ( see the beginning of section iii ) , however in the @xmath80 basis .
fleischhauer and coauthors , ref .
@xcite , classify this case as second order matched pulses .
the solution of the equations ( eq27)([eq29 ] ) is trivial since these equations in terms of a new variable @xmath326 can be reduced to a set of differential equations with constant coefficients , where @xmath327 . for the first time , this analytically solvable model was considered by vitanov and stenholm in ref .
we would classify the case as nonadiabatic , however in the second order @xmath328-basis , where the transition takes place . if the generalized pulse area is properly chosen in this basis , the population transfer between the diabatic states @xmath17 and @xmath18 is complete . for these particular pulse areas ,
there are no nonadiabatic corrections , which is typical for the resonant nonadiabatic transitions .
all adiabatic terms tend to zero at @xmath329 , which secures for the three - level atom the adiabatic following of a particular state coinciding with state @xmath95 at @xmath329 .
however , as it will be shown below , if we sum all these infinitely small terms , the result will be finite .
the net value of the small contributions of each adiabatic term is a nonadiabatic contribution , which specifies the excited probability amplitude left by the pulse train . to estimate this value , we rewrite the solution , eqs .
( [ eq42 ] ) and ( [ eq43 ] ) , as follows@xmath330@xmath331when @xmath329 , the function @xmath332 tends to a finite value whereas @xmath333 tends to zero since in the corresponding integrals [ see eq .
( [ eq54 ] ) ] @xmath334 is an even function of time and @xmath335 is an odd function . as a result ,
at the end of the pulse train @xmath336 and @xmath285 oscillate as @xmath337 and @xmath338 , respectively , and they have constant amplitudes @xmath339 .
the value of @xmath340 defines the probability amplitude left by the pulse sequence in states @xmath91 and @xmath90 .
the excitation process is adiabatic if the nonadiabatic part @xmath341 is small .
if the nonadiabatic part becomes comparable with the main term @xmath342 of the adiabatic expansion , then we can not describe the excitation process as adiabatic .
figure 5 ( a ) shows the numerically found dependences of the @xmath342 term ( dashed line ) and the nonadiabatic contribution @xmath340 ( solid line ) on the maximum amplitude @xmath271 of the pulses for the pulse train with @xmath343 .
the adiabatic and nonadiabatic parts of the solution become comparable if @xmath344 .
the semilogarithmic plot of @xmath340 versus @xmath271 , figure 5 ( a ) , clearly demonstrates that the nonadiabatic part decreases exponentially with the increase of @xmath271 .
laine and stenholm @xcite also found an exponential decrease of the nonadiabatic deviation from the ideal population transfer with increase of @xmath345 ( @xmath271 is the amplitude of each pulse at the maximum and @xmath123 is the interpulse distance , fixed in our case by the relation @xmath343 , so @xmath346 is a variable ) .
they followed the calculation procedure proposed by davis and pechukas @xcite employing the dykhne model @xcite . according to ref .
@xcite , one has to find the zeros of the function @xmath234 in a complex plane @xmath14 and take the one , @xmath347 , that is nearest to the real axis .
then the population of the dark state after the pulse train is@xmath348 .
\label{eq58}\ ] ] we propose another calculation procedure of the nonadiabatic deviation .
it will be shown that there are two nonadiabatic contributions , one coming from the rabi frequency @xmath234 and another from the mixing parameter derivative @xmath178 .
only the cooperative contribution of both determines the net nonadiabatic correction , while the pechukas - dykhne theory , taking into account only the @xmath234-change , underestimates the nonadiabatic correction . to show this
, we express @xmath332 via the fourier transform of @xmath349:@xmath350@xmath351let us consider first the case if the raman rabi frequency is constant , i.e. , @xmath352 . then @xmath353 and@xmath354 \,a(\omega ) = a(\omega ) |_{\omega = \beta _ { 0 } } , \label{eq61}\]]where @xmath355 is the dirac delta function and @xmath356 is an even function of @xmath142 .
the fourier transform of the mixing parameter derivative for a secant hyperbolic pulse train can be found , for example in ref .
this function is@xmath357 } { \sinh \left ( \frac{\pi \omega } { 2r}\right ) } .
\label{eq62}\]]if we take @xmath358 , which is the value of the raman rabi frequency at time @xmath258 when @xmath178 takes its maximum , then for our numerical example specified above we obtain the amplitude of the nonadiabatic contribution @xmath359 .
this value is four times smaller than the amplitude of the nonadiabatic oscillations on the right tail of the functions @xmath285 and @xmath336 ( which is @xmath360 ) , shown in figures 4 ( a , b ) , i.e. , four times smaller than the nonadiabatic contribution given by the numerical calculations of the solution of eqs . ( [ eq27])([eq29 ] ) . to explain this difference and clarify the origin of the nonadiabatic contribution , we recall the interaction hamiltonian in the @xmath80-basis , eq .
( [ eq25 ] ) [ see also eqs .
( [ eq15 ] ) and ( [ eq26 ] ) ] .
this hamiltonian resembles the interaction representation hamiltonian of the three - level system excited by two resonant `` fields '' with amplitudes @xmath234 and @xmath178 ( see fig .
2 ) . assume , first , that the couplings @xmath361 and @xmath261 are absent and the system is in state @xmath362 .
then , states @xmath95 , @xmath91 and @xmath90 can be considered as having the same energies in the interaction representation .
switching on the coupling @xmath363 [ here @xmath364 is the heaviside step function and @xmath365 is an arbitrary constant ] mixes states @xmath95 and @xmath91 or in other words induces the transition @xmath181 .
if the @xmath93field is also present and its amplitude @xmath366 is constant , i.e. , @xmath367 , this @xmath93field mixes states @xmath91 and @xmath90 producing a new couple of states @xmath368 and @xmath369 , which are the states @xmath67 and @xmath68 [ see eqs .
( [ eq8 ] ) and ( [ eq10 ] ) ] .
this couple is split by the energy gap @xmath370 [ see eq .
( [ eq6 ] ) ] .
this is the so called mollow splitting @xcite or quasi - energy splitting @xcite,@xcite .
assume that without the @xmath93field the @xmath177field is in resonance with levels @xmath95 and @xmath91 .
the switching on of the @xmath93field mixes levels @xmath91 and @xmath90 , producing an additional splitting .
level @xmath91 moves on the frequency @xmath93 out of resonance with the @xmath177field . if the @xmath186field would have a @xmath371spectrum ( in the case specified above it has only a zero frequency component ) , then it would not interact with the atom . however , because of the finite spectral width of the @xmath177field , its spectrum has a component with frequency @xmath372 on the far tail which is in resonance with the new position of level @xmath91 .
only this spectral component excites the atom if the @xmath93field is on . with increase of the @xmath93field , the component of the @xmath177field spectrum , which interacts with the atom , shifts further to the tail of the spectrum .
if the @xmath93field amplitude changes in time , several spectral components of the @xmath177field interact with the atom since at each instant of time some particular spectral component is in resonance .
the process of the sweeping of the splitting @xmath234 along the tail of the @xmath177field spectrum involves a broader band of the @xmath177spectrum in the interaction . to find the net atom excitation in this case
, we have to calculate the integral @xmath339 where the change of @xmath234 is taken into account .
this is done in the next two sections . concluding this section
, we apply our spectral method to find the approximate solution of the rosen - zener model . the comparison of our solution with the exact analytical solution helps to find the condition when our approximation is valid . referring to the bloch equations ( [ eq30])([eq32 ] ) discussed in section iii
, we define the splitting of the two - level system as @xmath373 and the coupling field amplitude as @xmath374 , where @xmath375 is the amplitude of the secant hyperbolic pulse of width @xmath13 .
according to the exact solution ( see , for example , ref .
@xcite ) , the population of the excited state @xmath144 after the pulse is@xmath376if initially the system is in the ground state @xmath143 , i.e. , @xmath377 . employing the correspondence between the bloch equations for the two - level system and the schrdinger equation for the three - level system discussed in section iii
, we make the substitution @xmath378 and @xmath379 in eq .
( [ eq54 ] ) where @xmath380 becomes a simple function of time , i.e. , @xmath381 . then ,
if @xmath329 , the solution is @xmath382 , @xmath383 [ see eqs .
( [ eq56])([eq57 ] ) and the discussion following these equations ] .
@xmath384 since @xmath380 is the odd function .
@xmath340 is simply the fourier transform @xmath385 [ defined as a nondimensional value according to eqs .
( [ eq59 ] ) and ( [ eq61 ] ) ] of the function @xmath171 at the frequency @xmath386 , i.e. , @xmath387 , [ see eqs .
( [ eq61 ] ) and ( eq59 ) ] , or explicitly@xmath388this integral also can be found in ref .
@xcite : @xmath389 .
the population of the excited state @xmath144 of the two - level system can be expressed as follows @xmath390 .
since there is a reciprocity between the bloch - vector components of the two - level system and the state probability amplitudes of the three - level system ( @xmath153 , @xmath154 , @xmath391 ) , the excited state population is @xmath392 .
the probability amplitudes are normalized , @xmath393 , and hence @xmath394 . assuming that @xmath395 , we obtain@xmath396thus , our approximate result , eq .
( [ eq65 ] ) , coincides with the exact solution , eq .
( [ eq63 ] ) , if the population change is small and the sine function can be represented by the first term of its expansion in a power series .
these conditions specify the validity of the adiabatic approximation .
what is remarkable here is that the dependence of the excited state population on the two - level splitting @xmath151 in the approximate solution coincides with that in the exact solution .
this shows that our spectral approach in the description of the nonadiabatic corrections is valid . if @xmath397 , the nonadiabatic transition between states @xmath143 and @xmath144 takes place according to the straightforward solution of the bloch - equation for the two - level system : @xmath398 $ ] , where @xmath399 . at @xmath400
we have @xmath401 , which reproduces the exact rosen - zener solution ( [ eq63 ] ) if @xmath402 .
if the raman rabi frequency changes during the development of the mixing parameter @xmath261 , those rabi frequencies sweeping the frequency bandwidth of @xmath261 contribute to the nonadiabatic corrections . to take this process into account , we have to convolute the spectral content of both the mixing parameter and the raman rabi frequency . in general , this calculation is nontrivial . for the case of secant hyperbolic pulses [ see the time development of the pulses , the raman rabi frequency and the mixing parameter in fig .
3 ( a , b ) ] , it is possible to simplify the problem . since @xmath234 has a minimum value of @xmath366 at @xmath403 where @xmath261 has its maximum , only the spectral components of @xmath356 with @xmath404
because time and frequency domains are inversely proportional , the main part of the nonadiabatic contribution appears in a short time interval around @xmath235 , corresponding to the far tail of the spectral distribution @xmath356 .
this means that the time scale of the nonadiabatic change is shorter than @xmath405 , while , according to the condition imposed on the parameters , the time variation of @xmath280 is much longer than the rabi oscillation period @xmath406 .
thus , a nonadiabatic transition takes place almost jumpwise compared to the time scale of the variation of @xmath261 . due to this circumstance
, we can simplify the calculation of the integral @xmath339 by expanding @xmath234 in a power series of @xmath14 near @xmath235 and retaining only the first two terms of the expansion@xmath407 , \label{eq66}\]]where @xmath408 /2 $ ] .
we verified this approximation by comparing numerically two integrals @xmath340 and @xmath409 calculated with @xmath234 and its approximated value , eq .
( [ eq66 ] ) , respectively .
the comparison is shown in fig .
5 b , where @xmath410 ( dashed line ) is the approximation .
the dependences of both integrals on the amplitude @xmath271 are indistinguishable .
the approximation of @xmath234 by a parabolic function helps to express @xmath339 via the airy integral .
further , to simplify the notations we set @xmath267 .
then the phase @xmath380 is@xmath411and @xmath412 is expressed as@xmath413 \int_{-\infty } ^{+\infty } d\omega \,a(\omega ) e^{-i\omega \tau } .
\label{eq68}\]]evaluating the time integral in eq .
( [ eq68 ] ) , we obtain@xmath414 , \label{eq69}\]]where @xmath415 is airy integral and @xmath416{g\beta _ { 0}}$ ] .
we derived this equation assuming that @xmath356 is an even function .
the airy integral has a different dependence for positive and negative arguments ( see , for example ref .
@xcite ) . if @xmath417 , we have@xmath418 = \frac{1}{3}\sqrt{\frac{% \beta _ { 0}+\omega } { 3g\beta _ { 0}}}\left\ { i_{-\frac{1}{3}}\left [ \frac{2}{3}% \sqrt{\frac{(\beta _ { 0}+\omega ) ^{3}}{g\beta _ { 0}}}\right ] -i_{\frac{1}{3}}% \left [ \frac{2}{3}\sqrt{\frac{(\beta _ { 0}+\omega ) ^{3}}{g\beta _ { 0}}}\right ] \right\ } , \label{eq70}\]]where @xmath419 is the modified bessel function of order @xmath420 . if the argument is negative , @xmath421 , then@xmath418 = \frac{1}{3}\sqrt{\frac{% \left| \beta _ { 0}+\omega \right| } { 3g\beta _ { 0}}}\left\ { j_{-\frac{1}{3}% } \left ( \frac{2}{3}\sqrt{\frac{\left| \beta _ { 0}+\omega \right| ^{3}}{g\beta _ { 0}}}\right ) + j_{\frac{1}{3}}\left ( \frac{2}{3}\sqrt{\frac{\left| \beta _ { 0}+\omega \right| ^{3}}{g\beta _ { 0}}}\right ) \right\ } , \label{eq71}\]]where @xmath422 is the bessel function of order @xmath423 . for large arguments
both functions have a simple asymptotic behavior@xmath418 \approx \frac{\exp \left [ -\frac{2}{3}\sqrt{\frac{(\beta _ { 0}+\omega ) ^{3}}{g\beta _ { 0}}}\right ] } { 2% \sqrt{\pi } \left [ g\beta _ { 0}(\beta _ { 0}+\omega ) \right ] ^{\frac{1}{4 } } } , \label{eq72}\]]if @xmath424 is positive , and@xmath418 \approx \frac{\cos \left ( \frac{2}{3}\sqrt{\frac{\left| \beta _ { 0}+\omega \right| ^{3}}{g\beta _ { 0}}}-% \frac{\pi } { 4}\right ) } { \sqrt{\pi } \left ( g\beta _ { 0}\left| \beta _ { 0}+\omega \right| \right ) ^{\frac{1}{4 } } } , \label{eq73}\]]if @xmath424 is negative .
if the argument is zero , @xmath425 , this function is@xmath418 |_{\omega = -\beta _ { 0}}=% \frac{\gamma ( \frac{1}{3})}{2\pi \sqrt[3]{\sqrt{3}g\beta _ { 0 } } } , \label{eq74}\]]where @xmath426 is the gamma function .
if the @xmath177-field does not depend on time and has a value @xmath427 , then @xmath428 and the main contribution to the nonadiabatic part is given by the airy integral@xmath429which has the explicit form @xmath430expressed via approximation ( [ eq72 ] ) . in this section
we show how the approximation , specified above , is related to the dykhne - pechukas model dykhne61 - 62,@xcite
. for large @xmath366 , the airy integral can be calculated by the _ saddle point methods _ as shown , for example , in ref .
first , the stationary or saddle point @xmath431 is found where the phase @xmath380 becomes stationary : @xmath432 . in this point , the raman rabi frequency becomes zero since @xmath433 .
for the case of the positive argument of the airy integral , the saddle point is in the complex plane and @xmath431 has only an imaginary component [ @xmath434 , i.e.,@xmath435 in fact , there are two saddle points : @xmath436 .
the negative sign is chosen to avoid exponentially increasing numbers .
then the method of stationary phase ( one of the saddle point _
_ methods ) is applied .
this method is applicable to integrals of the form@xmath437where @xmath366 is large and @xmath438 is a path in the complex plane such that the ends of the path do not contribute significantly to the integral .
the idea of the method is to deform the contour @xmath438 so that the region of most of the contribution to @xmath439 is compressed into as short a space as possible .
this compression occurs at the saddle point .
since the main contribution to this integral comes from the vicinity of the saddle point , one may conclude that in the adiabatic limit the atom abruptly makes a nonadiabatic transition between adiabatic states .
this is very much like crossing a stokes line for the asymptotic series , and resembles the turning point problem in the wkb theory ( see , for example , refs .
suominen95 , @xcite ) . in the wkb method ,
the one - dimensional schrdinger equation is considered for a particle with mass @xmath34 in a potential @xmath440:@xmath441\psi .
\label{eq78}\]]the turning point is the one where the energy of the particle coincides with the potential : @xmath442 .
if @xmath443/\hbar ^{2}\propto x$ ] , the solution of the equation ( [ eq78 ] ) can be expressed via the airy integral walker65 . in this approach the wkb connection formula , relating the exponentially small solution on one side of a turning point to an oscillatory solution on the other side ,
is derived quite naturally .
the pechukas - dykhne @xcite recipe of the calculation of the nonadiabatic contribution is similar to the method described above since the definitions of the saddle point @xmath431 and the crossing point @xmath347 , where @xmath444 is zero , are the same .
however , in our consideration we substituted the expression for the raman rabi frequency@xmath445by the expansion near time @xmath235 , eq .
( [ eq66 ] ) .
therefore , we have only two saddle or crossing points , ( [ eq78 ] ) , whereas equation ( eq79 ) has an infinite number of crossing points in the complex plane .
according to pechukas , one has to take the crossing point that is nearest to the real axis
. for the secant hyperbolic pulse train , this point , as shown by stenholm @xcite , has only an imaginary part @xmath446 , i.e.,@xmath447 . \label{eq80}\]]reformulating the dykhne approach @xcite
, we conclude : if there are no spectral components of @xmath356 matching the frequency gap between the quasi - energy levels split by the @xmath93-field , the @xmath448-field comes to resonance with the @xmath449 transition at the imaginary time @xmath347 when @xmath93 is zero .
this approach disregards the spectral content of @xmath261 .
if we take @xmath427 throughout the excitation , this method gives also an underestimated value of the nonadiabatic contribution .
for our numerical example , equation ( eq76 ) gives @xmath450 , which is @xmath451 times smaller than the value of the nonadiabatic contribution given by the numerical calculation of the schrdinger equation ( [ eq27])(eq29 ) .
figure 6 ( a ) shows the comparison of the numerically calculated nonadiabatic component @xmath340 ( solid line ) , eq .
( [ eq54 ] ) , with that calculated for the case if the time dependence of the @xmath452-field is neglected ( dashed line ) .
the latter is given by eq.([eq76 ] ) .
the plots obviously demonstrate that this approximation underestimates the nonadiabatic contribution . concluding this section
, we demonstrate the application of our method to the landau - zener model . in this model
the parameters of the corresponding two - level system are @xmath453 and @xmath454 . here
, the level crossing takes place at @xmath455 and the nonadiabatic correcton appears in the vicinity of the crossing point . after the substitution specified in section iii and repeating the same procedure
as described at the end of section iv , we find that@xmath456@xmath457the value @xmath458 is not zero for this model since @xmath459 is even function of time .
the integrals in eqs .
( eq81)-([eq82 ] ) are expressed via fresnel integral asymptotes ( see , for example , @xcite ) : @xmath460 . recall that @xmath461 .
then in the adiabatic limit , @xmath462 , we obtain@xmath463this expression coincides with the leading term of the exact solution , ref .
@xcite , expansion , that is@xmath464 we have to emphasize that our solution is valid only if the atom adiabatically follows the ground state @xmath143 .
therefore , the nonadiabatic correction ( [ eq83 ] ) , specifying the population probability of the excited state @xmath144 , coincides with the landau - zener solution , eq .
( [ eq84 ] ) , if @xmath465 , i.e. , when the population of the @xmath144 and @xmath143 states does not change appreciably . in this case
our three - level system @xmath80 , equivalent to the corresponding two - level system @xmath466 , also follows adiabatically state @xmath95 .
the opposite case is considered in the appendix . there
the new adiabatic states @xmath328 are introduced whose population also does not change appreciably .
this transformation is necessary if @xmath467 , because the reciprocal three - level system @xmath80 does not follow state @xmath95 , making a transition from state @xmath95 to state @xmath90 and back at the level crossing .
the necessity of the transformation can also be formulated in a different way . to apply the adiabatic solution
, we need a certain relation between the interaction parameters @xmath93 and @xmath177 .
then the expansion in @xmath202 s , where each next term of the expansion is smaller than the previous one [ see eqs .
( [ eq42])-([eq55 ] ) ] , is applicable .
if @xmath468 , this expansion can be made in the @xmath80-basis . in the opposite case , this expansion is not applicable
. therefore , we have to find new @xmath328-basis where the new interaction parameters @xmath469 and @xmath470 are related such that the expansion in @xmath202 s converges .
in previous sections we showed that the contribution to the nonadiabatic corrections of the rabi frequency chirping or the time dependence of the @xmath178-field , being taken into account separately , was not enough to have the right nonadiabatic correction for stirap induced by secant hyperbolic pulses . only both processes taken together , i.e. , the rabi frequency @xmath234 chirping and the participation of all spectral components @xmath471 of the @xmath261-field engaged by this chirping to excite the atom , give the right value of the nonadiabatic component .
this value is given by eq .
( [ eq69 ] ) . to calculate analytically the convolution integral of the two spectra , @xmath356 and the airy integral describing the process of the rabi frequency sweeping , we make two approximations .
first , we take the approximation of the airy integral , given by eq .
( [ eq72 ] ) in the frequency domain @xmath472 .
this gives a small overestimation of the integrand near @xmath473 .
for this reason , we start the integration from this value , not from @xmath474 . the oscillating part of the airy integral for @xmath475 gives a much smaller contribution than the main part , which is located between @xmath476 and zero , @xmath477 .
so , neglecting the part @xmath478 , we compensate the overestimation near @xmath473 .
second , we approximate the spectrum @xmath356 in the domain @xmath477 by@xmath479where @xmath480\right ) /2r\approx \pi /4r$ ] .
this approximation also gives a slight overestimation of the integrand near @xmath481 . to compensate this
, we stop the integration at @xmath482 .
the integrand has a maximum between @xmath483 and @xmath482 . to calculate the contribution of this part
, we use the modified method of the saddle point method . usually , in the method of the saddle point , the integration near the point is extended to @xmath484 . in our case , to avoid the overestimation of the integrand we limit the integration by finite boundaries .
the calculation of these boundaries for the deformed integration contour @xmath438 is simple because the deformed @xmath438 stays on the real axis .
the result of the integration is@xmath485 , \label{eq86}\]]where @xmath486 is a correction factor , which takes into account the finite integration boundaries to avoid the overestimation of the integral .
explicitly , we have@xmath487 , \label{eq87}\]]@xmath488 } , \label{eq88}\]]@xmath489and @xmath490 is the error function .
figure 6 ( b ) shows the comparison of the true nonadiabatic contribution @xmath340 ( solid line ) with our approximate calculation of @xmath412 , given by eq .
( [ eq86 ] ) ( dashed line ) .
figure 6 ( c ) compares the same dependencies if the correction factor is @xmath491 ( dash dotted line ) . for our numerical example
specified above , the approximated value of the nonadiabatic contribution is @xmath492 .
as was discussed above , this contribution appears in a very short time interval around @xmath235 .
therefore , we can approximate the solution of the schrdinger equations ( [ eq27])([eq29 ] ) by@xmath493@xmath494where the index @xmath495 stands for the approximation , @xmath496 is the heaviside step function , @xmath497 and @xmath93 has its exact value [ not the approximation ( [ eq66 ] ) ] .
figures 7 ( a , b ) show the comparison of the numerically found solutions of the schrdinger equation ( solid lines ) with the approximation ( [ eq90])([eq91 ] ) ( dashed lines ) .
the fit of the solutions is striking .
this means that the nonadiabatic contribution really appears in a quite short time range around @xmath403 .
the introduction of the basis of the dark and bright states facilitates the understanding of the physical processes in the three - level atom excited by a bichromatic field .
the dynamic evolution of the atom between the dark , bright and common states is as simple as the dynamics of the two - level atom excited by one field with rabi frequency @xmath93 .
this is because the evolution of the three - level atom can be effectively reduced to the evolution between two states , i.e. , the bright and common states .
the reduction of the three - level model to the two - level one allows also the application of the bloch - vector model and bloch equations for the treatment of the three - level atom excitation by the bichromatic field .
this quite simple algebra is applicable for the case of matched pulses .
if the pulses do not match in shape and have different time dependencies , one can also reduce the consideration to the bloch - vector model since there is a similarity between the schrdinger equations for the probability amplitudes of the dark , bright and common states and the bloch equation for an effective two - level system .
the effective detuning of the two - level system from resonance is @xmath234 and the rabi frequency is @xmath280 , which is the derivative of the mixing angle in the dark state development in states of the reciprocal three - level system .
this similarity allows a simple interpretation of the physical processes in the three - level system in case of adiabatic following of the dark state . in terms of the bloch - vector model , the adiabatic following criterion is formulated resulting in three conditions .
first , the initial state must coincide with the bloch - vector aligned along an effective field at @xmath293 .
this field is determined by the value @xmath498 .
second , the effective field makes a small angle with its initial position when the reciprocal parameter @xmath261 responsible for the atom - field interaction takes its maximum .
third , the time scale of the change of the pulse , determined by the scale of the variation @xmath261 , is much longer that the period of oscillations induced by the effective field @xmath234 .
if all three conditions are met , the dynamic evolution of the three - level atom excited by a sequence of pulses is described by simple algebra .
the adiabatic population transfer has a nice interpretation in terms of dark , bright and common states .
we developed an approximation describing the adiabatic interaction of the three - level atom with two resonant pulses .
the method of estimation of the nonadiabatic correction is presented .
it is applied to the case of two secant hyperbolic pulses .
the adiabatic part of the solution describes the excitation and de - excitation processes of the three - level atom .
the time dependence of the adiabatic part is smooth and follows the derivatives of the mixing angle @xmath191 .
both parts , excitation and de - excitation , are symmetric in time with respect to @xmath455 where @xmath261 takes its maximum and they exactly compensate each other . in this respect
the adiabatic following resembles the soliton - like interaction with the field .
the nonadiabatic part appears in a short time interval in the vicinity of the maximum of the mixing parameter @xmath261 .
it contains the information about the excitation left in the atom by the pulses because of the imperfect following of the dark state . since this excitation lasts only a short time , we consider the nonadiabatic process as a transition between the ground and excited states , which takes place like a jump .
the origin of the transition has a simple interpretation . the parameter @xmath93 defines the coupling strength of the bright ( @xmath91 ) and common ( @xmath94 ) states of the three - level system in the @xmath80basis .
the parameter @xmath178 defines the coupling strength of the bright and dark ( @xmath95 ) states .
it is assumed that , initially , the system is in the dark state and any time we have @xmath499 .
the @xmath93coupling moves the bright state from the resonance with the @xmath280coupling to the value determined by the frequency @xmath93 .
if the spectral content of the @xmath261coupling has a component with frequency @xmath93 , the nonadiabatic transition takes place .
if @xmath93 changes in time , several spectral components of the @xmath261coupling contribute to the transition .
our results are compared with those of laine - stenholm @xcite and fleischhauer et al .
@xcite .
we applied also our method to the rosen - zener and landau - zener models .
the comparison of our approximate solution with the exact solutions available for these models is quite encouraging .
we obtained exponentially small nonadiabatic corrections for the case of a symmetric time dependence of the excitation parameters @xmath234 and @xmath261 with respect to @xmath500 ( our model , the rosen - zener model and the landau - zener model with @xmath501 , and a small power dependent nonadiabatic correction for the case of an asymmetric time dependence of @xmath234 ( landau - zener model with @xmath465 ) .
we considered also the excitation of the three - level system by asymmetric pulses and found a power dependent nonadiabatic correction .
the results on asymmetric pulses , exciting the three - level atom , will be presented in another paper .
the simplified algebra developed in this paper could be useful for the description of the atom state manipulation by coherent fields .
this work was supported by the fonds voor wetenschappelijk onderzoek vlaanderen , the uiap program of the belgian government , istc ( 2121 ) and the russian foundation for basic research .
in this section we compare our approximate solution with an exact solution of the landau - zener model if the parameters of the corresponding two - level system @xmath453 and @xmath454 , satisfy the condition @xmath502 .
this condition is opposite to the one considered at the end of section vii .
there , since the coupling @xmath375 is small compared to the sweeping rate @xmath503 of the two - level system splitting @xmath151 ( @xmath465 ) , the system adiabatically follows the ground state @xmath143 and the net population of the excited state @xmath144 is small for @xmath329 .
if @xmath504 , states @xmath143 and @xmath144 are not an adiabatic basis . to apply the adiabatic approximation in this case , we have to choose a new basis that is adiabatic .
this can be done if we use again the similarity between the two - level and three - level systems discussed in section ii .
we consider the three - level system @xmath80 described by the hamiltonian ( [ eq25 ] ) with the parameters @xmath505 and @xmath506 .
then we make a @xmath317transformation to the new basis @xmath328 @xmath507 @xmath508 @xmath509 where @xmath510 and the new rabi frequency is @xmath511 . a new mixing angle @xmath512 takes the values @xmath513 at @xmath514 .
thus , it varies from @xmath515 to @xmath516 with time and , hence , state @xmath318 coincides with @xmath517 and @xmath518 before and after the crossing of levels @xmath143 and @xmath144 , which takes place at @xmath500 .
if the system adiabatically follows the changing state @xmath318 , no transition to the @xmath319 state takes place and , finally , the system is left in the initial state @xmath95 but with the @xmath109-shift of its phase .
this phase shift originates from the transition @xmath519 in the @xmath520basis .
one can find that , according to eq.([eq92 ] ) , the system is in state @xmath95 at @xmath514 when @xmath521 , and in state @xmath94 at @xmath455 when @xmath522 if the system follows state @xmath318 .
we know from the exact solution of the landau - zener model that @xmath523 and @xmath524 if @xmath525 [ see eq .
( [ eq84 ] ) ] .
thus , if @xmath526 , the population of the ground state @xmath143 of the corresponding two - level system is exponentially small at @xmath329 . with the help of our method for the estimation of the nonadiabatic correction , we can find the probability amplitudes of states @xmath319 and @xmath320 left by the interaction @xmath152 ( coupling states @xmath95 and @xmath91 ) after the level crossing of the equivalent two - level system .
these probability amplitudes allow to estimate the extent of the depopulation of state @xmath318 . if the system adiabatically follows the @xmath318 state , the state - vector changes from @xmath517 to @xmath518 .
since the development coefficient @xmath527 of the state vector in the @xmath80basis corresponds to the @xmath528component of the bloch - vector of the equivalent two - level system , the change of the @xmath529 sign means the flip of the @xmath528component .
this flip means a population inversion of the two - level system , i.e. , if before level crossing ( at @xmath293 ) the system is in the ground state @xmath143 , after the level crossing ( at @xmath530 ) it is in the excited state @xmath144 . we can estimate the amount of population left in the ground state , which is the nonadiabatic correction . to find the nonadiabatic correction ,
we apply the @xmath317transformation@xmath531 \label{eq95}\]]to the hamiltonian ( [ eq25 ] ) with the parameters @xmath532 and @xmath533 .
the result is@xmath534where@xmath535and@xmath536here , for simplicity , we set the phases [ see eq .
( [ eq1]),([eq4 ] ) ] @xmath537 and @xmath56 equal to zero .
the development coefficients @xmath538 , @xmath539 , @xmath540 of the state vector @xmath541 in the @xmath328basis satisfy equations that are similar to equations ( [ eq27])([eq29 ] ) if the substitution @xmath542 , @xmath543 and @xmath544 is made .
these equations are@xmath545@xmath546@xmath547the derivative of the mixing parameter @xmath512 has an explicit form@xmath548this function has a bell shape with maximum value @xmath549 at @xmath455 .
the modified rabi frequency @xmath550 has a minimum value @xmath551 at @xmath455 and increases to infinity at @xmath514 .
qualitatively , the time dependence of the interaction parameters resembles what we have for the stirap induced by the secant hyperbolic pulses , i.e. , @xmath234 and @xmath261 .
therefore , we can apply the same procedure of the calculation of the nonadiabatic correction and the same approximations .
first , we approximate the rabi frequency by its expansion in a power series of @xmath552 , retaining only the first two terms@xmath553where @xmath554 and @xmath555 [ see eq .
( [ eq66 ] ) ] .
rescaling time as @xmath556 , we can define the nonadiabatic correction @xmath412 as follows@xmath557where@xmath558and @xmath559 , @xmath560 [ compare the function @xmath561 with the one in eq .
( [ eq67 ] ) ] .
the fourier transform@xmath562of the coupling parameter @xmath563 is@xmath564following the same procedure as the one described in section vii , we obtain@xmath565 d\omega . \label{eq108}\]]employing the same approximation of the airy function by equation ( [ eq72 ] ) for positive arguments and taking into account only the negative wing of the function @xmath566 , we limit the integration in eq .
( [ eq108 ] ) by the boundaries @xmath567 .
these boundaries are taken to reduce the overestimation of the integrand introduced by the approximation applied to the airy function and @xmath566 function .
the arguments justifying this procedure are absolutely similar to those that are applied in section ix to the case of the secant hyperbolic pulses . after some algebra , we obtain@xmath568where@xmath569 , \label{eq110}\]]and@xmath570 }
, \label{eq111}\]]@xmath571this result can be obtained from eqs .
( [ eq86])([eq89 ] ) by the straightforward substitution @xmath572 , @xmath573 and @xmath574 .
we take @xmath575 or @xmath576 and @xmath577 as initial condition , which means that the reciprocal two - level system is initially in the ground state @xmath143 .
the final state is defined by the relation @xmath578 since @xmath579 keeps its sign , @xmath527 changes sign .
the population difference of the equivalent two - level system @xmath147 is related to @xmath580 as @xmath581 . if @xmath527 becomes negative , then @xmath582 . in this case , to estimate the population fraction left in the ground state @xmath143 , we have to use the expression@xmath583after the substitution @xmath584 into eq .
( [ eq113 ] ) and expanding the square root in a power series , we find the nonadiabatic correction for the rosen - zener model@xmath585figure 8 shows the comparison of our approximate solution , eq .
( [ eq114 ] ) , with the exact solution of the landau - zener model@xmath586the difference is small and it originates from the approximations made . to estimate the accuracy of our result , we compare eqs .
( [ eq114 ] ) and ( eq115 ) in logarithmic scale as follows@xmath587 = -\eta \pi \approx -\eta \left [ \frac{% 10}{3}-\frac{2}{\eta } \ln \left ( \frac{\pi q_{1}}{2}\right ) \right ] .
\label{eq116}\]]for large @xmath588 , the accuracy of the exponential factor calculation , expressed in the square brackets in eq .
( [ eq116 ] ) , is 6% .
see , for example , the review by k. bergmann , h. theuer , and b. w. shore , rev .
* 70 * ( 1998 ) 1003 .
r. n. shakhmuratov , j. odeurs , r. coussement , p. mgret , g. kozyreff , and p. mandel , phys .
* 87 * ( 2001 ) 153601 .
m. d. crisp , phys .
rev . * a 8 * ( 1973 ) 2128 .
l. v. keldysh , jetp * 47 * ( 1964 ) 1945
. a. m. perelomov , v. s. popov , and v. s. terentyev , jetp * 50 * , 1393 ( 1966 ) , _
ibid _ * 51 * ( 1966 ) 309 .
a. i. nikishov , and v. i. ritus , jetp * 50 * ( 1966 ) 255 , ibid * 52 * ( 1967 ) 223 .
a. bychkov , and a. m. dykhne , jetp * 58 * ( 1970 ) 1734 .
a. h. zewail , sci . am .
* 263 * ( 1990 ) 40 , j. phys . chem . * 97 * ( 1993 ) 12427 .
g. beddard , rep .
* 56 * ( 1993 ) 63 .
b. m. garraway , and k .- a .
suominen , rep .
* 58 * ( 1995 ) 365 .
l. d. landau , phys .
* 1 * ( 1932 ) 89 , _ ibid _ * 2 * ( 1932 ) 46 . c. zener , proc .
* 137 * ( 1932 ) 696 .
e. c. stueckelberg , helv .
acta * 5 * ( 1932 ) 369 . v. k. bykhovsky , ye .
ye . nikitin , and m. ya .
ovchinnikova , jetp * 47 * ( 1964 ) 570 .
d. s. f. crothers , adv . phys .
* 20 * ( 1971 ) 86 , 405 ; j. phys .
* a 5 * ( 1972 ) 1680 , j. phys . *
b 9 * ( 1975 ) 442 , _
ibid _ * b 9 * ( 1976 ) 635 .
a. d. bandrauk , and m. s. child , molec
* 19 * ( 1970 ) 95 .
m. s. child , molec .
* 31 * ( 1976 ) 4 , 1031 .
w. r. thorson , j. b. delos , and s. a. boorstein , phys .
rev . * a 4 * ( 1971 ) 1052 .
j. b. delos , and w. r. thorson , phys . rev .
* a 6 * ( 1972 ) 720 , 728 .
g. v. dubrovskiy , and i. fisher - hjalmars , j. phys .
* b 7 * ( 1974 ) 892 .
m. demkov , and v. i. osherov , jetp * 53 * ( 1967 ) 1589 .
m. demkov , and m. kunike , vestn .
fiz . khim .
* 16 * ( 1969 ) 39 .
k - a suominen , opt .
commun . * 93 * ( 1992 ) 126 .
a. m. dykhne , sov .
jetp * 11 * , 411 ( 1960 ) , jetp * 41 * , 1324 ( 1961 ) , sov .
jetp * 14 * ( 1962 ) 941 .
a. m. dykhne and a. v. chaplik , jetp * 43 * ( 1962 ) 889 .
j. oreg , f.t .
hioe , and j.h .
eberly , phys .
rev . a * 29 * ( 1984 ) 690 .
hioe , and j.h .
eberly , phys .
* 47 * ( 1981 ) 838 .
hioe , phys . rev . a * 28 * ( 1983 ) 879 ; _ ibid .
_ a * 29 * ( 1984 ) 3434 , _ ibid_. a * 30 * ( 1984 ) 3097 .
kuklinsky , u. gaubatz , f.t .
hioe , and k. bergmann , phys .
a * 40 * ( 1989 ) 6741 .
hioe and c.e .
carroll , phys .
a * 37 * ( 1988 ) 3000 .
carroll and f.t .
hioe , phys .
a * 42 * ( 1990 ) 1522 .
e. arimondo , coherent population trapping in laser spectroscopy , in progress in optics xxxv , e. wolf ed . ( 1996 ) , p.p . 257 - 354 .
e. arimondo and g. orriols , nuovo cimento lett .
* 17 * ( 1976 ) 333 .
gray , r.m .
whitley , and c.r .
stroud jr .
* 3 * ( 1978 ) 218 .
g. orriols , nuovo cimento b * 53 * ( 1979 ) 1 .
t. a. laine and s. stenholm , phys .
rev . a * 53 * ( 1996 ) 2501 .
m. fleischhauer , r. unanyan , b. w. shore and k. bergmann , phys .
a * 59 * ( 1999 ) 3751 .
j. p. davis and p. pechukas , j. chem .
* 64 * ( 1976 ) 3129 .
hwang , and p. pechukas , j. chem .
* 67 * ( 1977 ) 4640 .
l. allen and j. h. eberly , _ optical resonance and two - level atoms _ ( wiley , new york , 1975 ) . m. v. berry , proc .
london , ser . a * 414 * ( 1987 ) 31 ; * 429 * ( 1990 ) 61 .
r. lim and m. v. berry , j. phys .
a * 24 * ( 1991 ) 3255 .
m. v. berry and r. lim , j. phys .
a : math . gen .
* 26 * ( 1993 ) 4737 .
r. lim , j. phys .
a : math . gen . * 26 * ( 1993 ) 7615 .
a. abragam , _ the principles of nuclear magnetism _ ( clarendon press , oxford 1961 ) . ya
. b. zeldovich , uspekhi fiz .
nauk * 110 * ( 1973 ) 139 .
a. r. kessel and r. n. shakhmuratov , physica c * 92 * ( 1977 ) 132 .
s. e. harris , phys .
lett . * 70 * , 552 ( 1993 ) ; * 72 * ( 1994 ) 52 .
s. e. harris , and z .- f .
luo , phys .
a * 52 * ( 1995 ) r928 .
n. v. vitanov and s. stenholm , phys .
a * 55 * ( 1997 ) 648 .
a. erdlyi , w. magnus , f. oberhettinger , f. g. tricomi , h. bateman manuscript project , _ tables of integral transforms _ , vol .
a. erdlyi ( mcgraw - hill book company , inc .
, new york , toronto , london 1954 ) .
b. r. mollow , phys .
* 188 * ( 1969 ) 1969 .
m. abramovitz and i. a. stegun , _ handbook of mathematical functions _ ( dover pub . ,
, new york 1965 ) .
j. mathews and r. l. walker , _ mathematical methods of physics _ ( w. a. benjamin , inc , new york 1965 ) p. 111
fig.1 . the excitation scheme of the three - level atom by two resonant fields with interaction constants @xmath53 and @xmath54 : @xmath24-scheme ( a ) , and the same excitation scheme in the basis of dark @xmath95 , bright @xmath91 and common @xmath107 states ( b ) .
@xmath87 is the generalized interaction constant for the raman excitation .
the vertical scale in diagram b is not the energy and represents the initial population of the levels before the excitation .
the higher populated level is put at the bottom and the lower or unpopulated level is at the top .
the bichromatic excitation scheme of the three - level atom in the @xmath589 basis if the state developments of @xmath75 and @xmath131 change in time .
the coupling parameter of states @xmath95 and @xmath91 is @xmath177 , the derivative of the mixing angle ( see the text ) .
the pulse train with corresponding interaction parameters @xmath590 and @xmath204 evolving in time ( bold lines ) .
they are normalized by the maximum value @xmath271 .
the delay between pulses is @xmath591 , @xmath592 .
the time is scaled in units of @xmath13 .
the time dependence of the mixing parameter @xmath202 is shown by the thin line .
the dashed line shows the dependence of the mixing parameter derivative @xmath177 normalized by @xmath13 . in ( b ) , the time dependence of the bichromatic rabi frequency @xmath593 ( solid line ) and the mixing parameter derivative ( dashed line ) are shown for comparison .
4 . ( a ) the evolution of the amplitudes @xmath264 ( plot a ) , @xmath180 ( plot b ) , @xmath265 ( plot c ) for a pulse train @xmath203 , @xmath204 with the parameters @xmath270 , @xmath266 , @xmath594 .
solid lines are the numerical solution of equations ( [ eq27])([eq29 ] ) .
dashed lines are analytical approximations given by the first two terms of equations ( eq46,[eq47 ] ) and the first two terms in each of the parentheses of equation ( [ eq55 ] ) .
5 . ( a ) comparison of the dependence of the adiabatic @xmath595 ( dashed line ) and nonadiabatic @xmath340 ( solid line ) parts of the analytical solution of eqs .
( [ eq27])([eq29 ] ) on the maximum pulse amplitudes @xmath271 . in ( b ) , the comparison of the nonadiabatic part , numerically calculated using the actual rabi frequency @xmath234 ( solid line ) and its parabolic approximation , eq .
( [ eq66 ] ) ( dashed line ) , are shown .
@xmath340 is the true nonadiabatic part and @xmath412 is an approximation .
( a ) comparison of the dependencies of the nonadiabatic contributions @xmath340 ( solid line ) and @xmath412 ( dashed line ) versus the pulse amplitude @xmath271 .
the first is calculated with and the second without taking into account the time dependence of the mixing parameter derivative @xmath261 .
( b ) the plots of the true , @xmath339 , and the approximate , @xmath412 , nonadiabatic contributions versus @xmath271 , where @xmath412 ( long dashed line ) is calculated taking into account the time dependence of @xmath596 using the approximation described in the text .
( c ) the same plots as in ( b ) , except @xmath412 ( dot dashed line ) , where the correction factor @xmath486 is dropped ( see the text ) .
numerical solution of the schrdinger equation ( solid lines ) for the amplitudes @xmath285 ( plot a ) , @xmath336 ( plot b ) and the approximation given by equations ( [ eq86])([eq87 ] ) ( dashed lines ) for the amplitudes @xmath597 ( plot a ) , @xmath598 ( plot b ) .
comparison of the exact solution of the landau - zener model ( dashed line ) with our approximate solution ( solid line ) for the ground state population of the reciprocal two - level system after level crossing .
the dependence of the exact solution of the landau - zener model for the ground state population , @xmath599_{lz}$ ] , on the parameter @xmath600 is shown by the dashed line [ see eq .
( [ eq115 ] ) ] . the same dependence of our approximate solution , @xmath599_{a}$ ] , is shown by the solid line .
our solution is given by eq .
( [ eq114 ] ) . | an accurate theory describing adiabatic following of the dark , nonabsorbing state in the three - level system is developed .
an analytical solution for the wave function of the particle experiencing raman excitation is found as an expansion in terms of the time varying nonadiabatic perturbation parameter .
the solution can be presented as a sum of adiabatic and nonadiabatic parts .
both are estimated quantitatively .
it is shown that the limiting value to which the amplitude of the nonadiabatic part tends is equal to the fourier component of the nonadiabatic perturbation parameter taken at the rabi frequency of the raman excitation .
the time scale of the variation of both parts is found .
while the adiabatic part of the solution varies slowly and follows the change of the nonadiabatic perturbation parameter , the nonadiabatic part appears almost instantly , revealing a jumpwise transition between the dark and bright states .
this jump happens when the nonadiabatic perturbation parameter takes its maximum value . |
quasiclassical methods , developed several years ago and used successfully to describe non - equilibrium states in superconductors and superfluids ( see , e.g. , @xcite ) , have recently been extended to meso- and nanoscopic systems .
this became possible after the formulation of boundary conditions , including hybrid structures and interface roughness @xcite . in recent developments we focussed on spin - effects and spatial confinement , e.g. , concerning the spin - hall effect in a 2d electron gas @xcite , and spin relaxation in narrow wires in the presence of spin - orbit coupling @xcite . within the quasiclassical approach ,
it is also possible to study the influence of disorder and coulomb interaction on the same footing @xcite , which is manageable since the method works on an intermediate level : microscopic details of the wave - functions , on the scale of the interatomic distance , are integrated out , leaving equations of motion which can be solved with sufficient accuracy . as an illustrative example , we apply in this paper the quasiclassical method to determine the equilibrium green s function , and hence the density of states ( dos ) , of the ( spinless ) 1d model known as luttinger liquid @xcite .
this model contains two fermionic branches , linearized near the fermi points , with @xmath1 and @xmath2 the bare fermi velocity and dos .
only interaction processes with small ( @xmath3 ) momentum transfer are considered : standard parameters are @xmath4 and @xmath5 , for scattering processes involving only one or both branches , respectively .
the dimensionless quantities @xmath6 and @xmath7 are useful .
the spectrum of charge fluctuations is linear , @xmath8 ; the renormalized velocity , @xmath9 , and the parameter @xmath10 , given by @xmath11^{1/2 } \ ; , \;\ ; k = \left [ \frac{1+\gamma_4-\gamma_2}{1+\gamma_4+\gamma_2 } \right]^{1/2 } \ ; , \ ] ] are often called luttinger liquid parameters . in the following sect .
2 we discuss the solution of the equation of motion for the quasiclassical green s function in the presence of a fluctuating potential , the latter representing the fermion - fermion interaction .
the keldysh technique is employed throughout .
considering the average with respect to the fluctuating field of the green s function and an analogy to the @xmath0 theory , the field fluctuations are related to an effective impedance of the luttinger liquid , and finally to the dos ( sect .
3 ) . in sect .
4 we consider an insulating boundary , i.e. a reflecting wall , and determine the dos boundary exponent . a concluding discussion is given in sect .
in essence , the keldysh technique @xcite differs from the zero - temperature and the matsubara approach by employing time - ordering along a contour which runs from @xmath12 to @xmath13 and back to @xmath12 . rewriting the theory in terms of standard ( @xmath14 ) integrations , a 2@xmath152 matrix structure results ; e.g. , the green s function can be cast into the form @xmath16 accordingly , considering some general 2-particle interaction @xmath17 and decoupling this interaction using the hubbard - stratonovich transformation within the path - integral formalism , two auxiliary fields are necessary . as a result , the green s function can be expressed as an average of the free - particle green s function in the presence of fluctuating potentials as follows ( see , e.g. , ref .
@xcite ) : @xmath18 \ , { \rm e}^{{\rm tr}\ln ( \check{1 } + \check{g}_0 \check{\phi } ) } \rangle_{0 , \phi } = \langle \check{g}_0 [ \phi ] \rangle_\phi\ ] ] here @xmath19 is the free - particle green s function , and @xmath20 = \check{g}_0^{-1 } + \check{\phi}$ ] , where @xmath21 ; @xmath22 and @xmath23 denote the unit and the first pauli matrix , and @xmath24 and @xmath25 correspond to the two potentials mentioned above . furthermore , @xmath26 denotes an average which is gaussian by construction , and defined such that @xmath27 where for clarity space ( @xmath28 ) and time ( @xmath29 ) arguments are included .
the average @xmath30 , defined through eqs . (
[ av ] ) and ( [ phi ] ) , can be non - gaussian .
( the spin will be suppressed througout this paper . )
note that a `` real '' electrical potential can be included by the replacement @xmath31 in @xmath32 $ ] of eq.([av ] ) ; the particle s charge is @xmath33 . in the next step ,
we consider the quasiclassical approximation , i.e. we consider the difference @xmath34 \check{g}_0 [ \phi ] - \check{g}_0 [ \phi ] \check{g}_0^{-1 } [ \phi ] = 0 $ ] , keeping only the leading terms in the gradients with respect to the ( spatial ) center - of - mass coordinate ( which we again denote by @xmath28 ) .
the equation is then integrated with respect to the magnitude of the momentum .
the result is @xmath35 \check{g}_{tt^\prime}(\hat{p}x ) = i [ \check{\phi } , \check{g}]\ ] ] where @xmath36 is the direction of the center - of - mass momentum , and @xmath37 \equiv \check{\phi}(xt)\check{g}_{tt^\prime}(\hat{p}x ) -
\check{g}_{tt^\prime}(\hat{p}x ) \check{\phi}(xt^\prime ) \ ; .\ ] ] the quantity @xmath38 is the so - called quasiclassical green s function , @xmath39 a major advantage compared to the full green s function is the normalization , @xmath40 .
it should be noted that in the quasiclassical approximation , only long - wavelength contibutions of the fluctuating fields are taken into account which is adequate for the problem addressed below .
in the 1d case , @xmath41 , and @xmath42 . for the homogeneous equilibrium case , the solution is @xmath43 where the fourier transformed ( @xmath44 ) quantity is given , and @xmath45 , @xmath46 . a formal solution of eq . ( [ equ ] )
is found with the ansatz @xmath47 provided @xmath48 apparently , @xmath49 has the same matrix structure as @xmath50 , namely @xmath51 .
for example , we obtain explicitly from ( [ ansatz ] ) : @xmath52 @xmath53 where @xmath54 and @xmath55 for brevity .
we will discuss below that , for the luttinger model , it is sufficient to assume that the phases are gaussian distributed .
it is then straightforward to determine @xmath56 ; for example , we obtain , but we nevertheless use the notation @xmath57 for brevity . ]
@xmath58 \\ & = & - \langle \check{g}^a \rangle_\phi ( t^\prime t)\end{aligned}\ ] ] where @xmath59 note that @xmath60 and @xmath61 are even , and @xmath62 and @xmath63 are odd under time reversal , @xmath64
. then we define @xmath65 , such that @xmath66 \left [ b(\omega ) + 1 \right ] \left ( { \rm e}^{-i\omega t } -1 \right)\ ] ] where @xmath67 , and we used @xmath68 b ( \omega)$ ] . from the relation @xmath69 we obtain @xmath70 , implying that @xmath71 is odd in frequency . combining the above relations , the ( normalized ) dos is given by @xmath72\}_{t^\prime=0 } \nonumber \\ & = & 1 + \frac{1}{2 } \int d\omega \ , f(\epsilon - \omega ) [ p(\omega ) - p(-\omega)]\end{aligned}\ ] ] with @xmath73 note that @xmath74 , i.e. @xmath75 . in order to complete the argument ,
the phase fluctuations are easily related to the potential fluctuations , according to the relation ( compare eq .
( [ solution ] ) ) @xmath76_{x = x^\prime , t^\prime = 0 } \ ; , \ ] ] and the potential fluctuation can be related to the interaction .
this last step , however , requires some discussion .
first , note that expanding the exponent in eq .
( [ av ] ) up to second order , neglecting higher order terms , corresponds to the random phase approximation ( rpa ) . given this ( gaussian ) approximation , the procedure for performing the phase average , outlined above ,
is permitted . as an important point , however , _ the rpa is known to be exact for calculating the density response of the luttinger model _ , and hence the effective interaction , which concludes the argument : the potential fluctuations are given by @xmath77 where @xmath78 is the screened ( rpa ) interaction .
thus @xmath79 in an obvious short - hand notation .
the approach described in the preceding section is very similar to the theory of charge tunneling in ultrasmall junctions , as reviewed , e.g. , in @xcite , which is also known as @xmath0 theory .
( see also @xcite for related articles . ) in this context , the quantity @xmath0 characterizes the influence of the environment on the tunneling between the two electrodes .
for example , the forward tunneling rate is found to be given by @xmath80 \ , p(e - e^\prime ) \ ] ] where @xmath81 is the voltage , @xmath82 the tunneling resistance , and @xmath83 the fermi function .
note the detailed balance condition @xmath84 .
physically , @xmath0 describes the probability of exchanging the energy @xmath85 with the environment @xcite . on the other hand , in the present context , consider the tunneling rate from the interacting one - dimensional wire into a non - interacting lead , which clearly is given by @xmath86\ ] ] with @xmath87 the normalized dos as given in eq .
( [ dos ] ) . using this equation as well as the properties of @xmath88 as described in the previous section and the detailed balance condition , it is straightforward to confirm that eqs .
( [ forward ] ) and ( [ forwardb ] ) coincide , provided @xmath0 from the tunneling theory is identified with @xmath88 .
the reason for the complete correspondence between these two quantities is apparent on physical grounds , since the quantity @xmath88 characterizing the luttinger liquid arises from the auxiliary potential fluctuations due to the fermion - fermion interaction .
in the zero - temperature limit , one finds easily @xmath89 exploiting the analogy further , we identify @xmath90 where @xmath91 is the klitzing constant , and we may call @xmath92 effective impedance of the luttinger liquid . for the simple example of an ohmic impedance with a high - frequency cut - off @xmath93 the result is @xcite @xmath94^{2/g } \quad\quad ( t=0 , \ , 0<\omega < \omega_r)\ ] ] where @xmath95 is euler s constant .
thus the dos is found to vanish at the fermi energy , @xmath96 , where the exponent is given by @xmath97\right\}_{\omega\to 0 } = -i \left\ { \frac{\omega}{\pi } \int \frac{dq}{2\pi } \frac{v^r(q\omega)}{(-i\omega + 0 + i q v_f)^2 } \right\}_{\omega\to 0}\ ] ] where @xmath98 is the screened retarded interaction . in order to determine this quantity , consider the 2@xmath152 matrix structure of right- and left - moving particles , denoted by `` @xmath99 '' and `` @xmath100 '' , the bare interaction and the non - interacting response function : @xmath101 as well as the rpa equation @xmath102 .
this matrix structure implies that we have to introduce a branch index for the fluctuating fields , which we suppressed up to now : @xmath103 .
the density of states of the right - moving particles , for example , is determined from the @xmath104-@xmath104 correlation functions .
the relevant quantity , @xmath105 , is straightforwardly determined and inserted into eq .
( [ 2overg ] ) ; the @xmath106-integral can be done by contour integration , and we obtain the standard @xcite result : is denoted by @xmath107 , which we avoid here since we prefer to use the latter symbol for the dimensionless conductance , see eq .
( [ ohmic ] ) , in accordance with the @xmath0 theory @xcite . ]
near a boundary or an interface the quasiclassical approximation is insufficient , and the quasiclassical propagators pointing into or out of the boundary ( or interface ) have to be connected by boundary conditions @xcite .
the boundary condition is particularly simple for spinless fermions in one dimension at an impenetrable wall : obviously @xmath109 , to ensure that there is no current through the wall . in the following
we assume that the wall is located at @xmath110 , and that the particles move in the half - space @xmath111 . instead of considering two branches of fermions in this half - space ,
we find it more convenient to mirror the left movers at the boundary and to consider only right movers in the full space , i.e.@xmath112 this green s function solves eq .
( [ equ ] ) both for @xmath113 and @xmath114 .
the bare interaction is now given in real space by @xmath115 the fourier transfrom of @xmath17 is thus off - diagonal in momentum space , since the @xmath5-term couples @xmath106 with @xmath116 .
considering now the 2@xmath152 matrix structure of particles with momentum @xmath106 and @xmath116 , the bare interaction is @xmath117 i.e. identical to what was given in eq .
( [ v0 ] ) above with just a different meaning of the matrix index .
also the non - interacting response function and the rpa equation for the screended interaction are the same as before . due to the absence of translation symmetry
the effective impedance depends on the distance from the boundary , and is given by @xmath118 \\ & = & \frac{1}{4}\left [ k + k^{-1}-2 + \cos ( 2 \omega x/ v ) ( k^{-1}-k ) \right ] \ ; .\end{aligned}\ ] ] hence the density of states at the boundary is found to vanish as @xmath119 in agreement with the boundary exponent obtained in @xcite , and the impurity exponent given in @xcite .
in fact this result has been considerably debated @xcite ; compare also @xcite , as well as the detailed presentation by von delft and schoeller @xcite .
quasiclassical theory is known to be a useful tool in the theoretical description of superconductors and superfluids , including non - equilibrium states and interfaces . for normal - conducting electrons the theory mainly serves for the microscopic foundation of a boltzmann - like transport theory . however
, when taking into account the coulomb interaction in terms of a fluctuating hubbard - stratonovich field , the theory also captures important interaction effects , which are beyond the reach of the boltzmann equation . in this article we illustrated this by considering the one - dimensional version of the theory , and demonstrated that it is possible to describe the non - fermi liquid physics of luttinger liquids .
the mathematics involved is very similar to the @xmath0 theory of tunneling , which we made use of in this article , and also to the functional bosonization approach to luttinger liquids @xcite ( which we did not exploit here ) .
we concentrated on the spinless case and on thermal equilibrium ; the spin is straightforwardly included in the theory , and clearly since we use the keldysh formalism
non - equilibrium situations can be covered as well .
we are confident that in the future the quasiclassical theory will become a useful , complementary approach to transport in interacting one - dimensional systems .
m. sassetti and u. weiss , europhys . lett .
* 27 * , 311 ( 1994 ) ; j. rollbhler and h. grabert , phys .
lett . * 91 * , 166402 ( 2003 ) ; i. safi and h. saleur , phys .
* 93 * , 126602 ( 2004 ) ; k. le hur and m .-
li , phys .
b * 72 * , 073305 ( 2005 ) ; k. le hur , phys . rev .
b * 74 * , 165104 ( 2006 ) .
m. fabrizio and a. gogolin , phys .
b * 51 * , 17827 ( 1995 ) ; s. eggert , h. johannesson , and a. mattson , phys . rev
* 76 * ; 1505 ( 1996 ) ; v. meden , w. metzner , u. schollwock , o. schneider , t. stauber , and k. schonhammer , eur .
j. b * 16 * , 613 ( 2000 ) .
h. c. fogedby , j. phys .
c : solid state phys . * 9 * , 3757 ( 1976 ) ; d. k. k. lee and y. chen , j. phys . a : math
* 21 * , 4155 ( 1988 ) ; a. grishin , i. v. yurkevich , and i. v. lerner , phys .
b * 69 * , 165108 ( 2004 ) ; i. v. lerner and i. v. yurkevich , proceedings of lxxxi les houches school on `` nanoscopic quantum transport '' , les houches , france , june 28
july 30 , 2004 [ cond - mat/0508223 ] . | the method of the quasiclassical green s function is used to determine the equilibrium properties of one - dimensional ( 1d ) interacting fermi systems , in particular , the bulk and the local ( near a hard wall ) density of states .
while this is a novel approach to 1d systems , our findings do agree with standard results for luttinger liquids obtained with the bosonization method .
analogies to the so - called @xmath0 theory of tunneling through ultrasmall junctions are pointed out and are exploited .
further applications of the green s function method for 1d systems are discussed . |
as the number of extragalactic cepheids discovered with _ hst _ continues to increase and the value of @xmath3 is sought from distances based on these variables ( e.g. saha et al .
1999 ; mould et al .
2000 ) , it becomes even more important to understand various possible systematic errors which could affect the extragalactic distance scale .
currently , the most important systematic is a bias in the distance to the large magellanic cloud , which provides the zero - point calibration for the cepheid distance scale .
the lmc distance is very likely significantly shorter than usually assumed ( e.g. udalski 1998 ; stanek et al .
2000 ) , but it still might be considered uncertain at the @xmath4% percent level ( e.g. jha et al .
another possible systematic , the metallicity dependence of the cepheid period - luminosity ( pl ) relation , is also very much an open issue , with the empirical determinations ranging from 0 to @xmath5 ( freedman & madore 1990 ; sasselov et al .
1997 ; kochanek 1997 ; kennicutt et al .
1998 ) .
in this paper we investigate a much neglected systematic , that of the influence of blended stellar images on the derived cepheid distances .
although cepheids are very bright , @xmath6 at a period of @xmath7 , their images when viewed in distant galaxies are likely to be blended with other nearby , relatively bright stars .
we define _ blending _ as the close projected association of a cepheid with one or more intrinsically luminous stars , which can not be detected within the observed point - spread function ( psf ) by the photometric analysis ( e.g. , daophot , dophot ) .
such blended stars are mostly other young stars which are physically associated @xmath1 from actual binary and multiple systems to companions which are not gravitationally bound to the cepheid .
blending is thus a phenomenon different from _ crowding _ or _ confusion noise _ ; the latter occurs in stellar fields with a crowded and complex background due to the random superposition of stars of different luminosity . in this paper
we are concerned with blending due to wide unbound systems .
binary cepheid companions are well studied ( evans 1992 ) , and do not contribute enough flux to affect cepheid distances ( madore 1977 ) , due to the obvious constraints of coeval stellar evolution . on the other hand ,
the association of cepheids with other luminous stars in wide unbound systems is an unsolved problem in general .
while such association , i.e. a strong star - star correlation function , is expected and common to young stars ( harris & zaritsky 1999 ) , the specific case for cepheids is unknown . studies in our galaxy are very difficult due to the small sample and existing results , though tantalizing , are inconclusive ( evans & udalski 1994 ) .
this could explain the relative neglect of this issue in recent years , but blending had been of concern for early studies of magellanic cloud cepheids ( deyoreo & karp 1979 ; pel , van genderen & lub 1981 ) , because even faint b - star blends affect the optical colors of a cepheid significantly .
we investigate the effects of stellar blending on the cepheid distance scale by studying two local group spiral galaxies , m31 and m33 . in this paper
we concentrate on m31 ( andromeda galaxy ) , located at approximately @xmath8 ( e.g. holland 1998 ; stanek & garnavich 1998 ) from us . as part of the direct project ( e.g. kaluzny et al .
1998 ; stanek et al . 1998 ) we have collected for this galaxy an extensive data set , finding among other variables 206 cepheids .
we identify some of these cepheids on archival _ hst_-wfpc2 images and compare them to our ground - based data to estimate the impact of blending on our photometry , taking advantage of their superior resolution
the fwhm on the wfpc2 camera corresponds to @xmath9 at the distance of m31 , compared to @xmath10 for the ground - based data .
the average fwhm on the direct project ground - based images of m31 is about @xmath11 , or @xmath10 , which corresponds to the _ hst_-wfpc2 resolution of @xmath12 for a galaxy at a distance of @xmath2 .
any luminous star ( or several of them ) in a volume of that cross section through the disk ( at the inclination of the galaxy ) could be indistinguishable from the cepheid and would contribute to its measured flux .
as cepheids are relatively young stars , they reside strictly in the midplanes of the disks of spiral galaxies . the archival _
hst_-wfpc2 study of m31 cepheids from our project direct was undertaken to improve our distance determination to the galaxy .
the preliminary indirect ( via lmc ) cepheid distance we obtained ( kaluzny et al . 1998 ; sasselov et al .
1998 ) was practically the same as the cepheid distance by freedman & madore ( 1990 ) of @xmath13 .
our findings of blending now indicate that these distance estimates should be corrected upward .
we describe the ground - based and _ hst _ data and the applied reduction procedures in section 2 . in section 3
we discuss the task of identifying cepheids on _ hst _ wfpc2 images . in section 4 we present the blending catalog of cepheids and discuss it in section 5 . in section 6
we describe the effect of blending on the light and color curves of cepheids and how these curves can be used to detect blends . in section 7
we discuss the effectiveness of some of the blend analysis methods encountered in the literature .
the concluding remarks are to be found in section 8 .
the ground - based data were obtained as part of the direct project between september 1996 and october 1997 during 95 full / partial nights on the f. l. whipple observatory 1.2 m telescope and 36 full nights on the michigan - dartmouth - mit 1.3 m telescope .
six @xmath14 fields with a scale of 0.32 /pixel were monitored : four of them ( a d ) concentrated on the rich spiral arm in the northeast part of m31 , one ( e ) close to the bulge of m31 and one ( f ) containing the giant star formation region known as ngc 206 .
fields a d and f have been reduced and the @xmath15 photometry of cepheid variables published in stanek et al .
1998 ( hereafter , paper ii ) , kaluzny et al . 1998
( paper i ) , stanek et al .
1999 ( paper iii ) , kaluzny et al .
1999 ( paper iv ) and mochejska et al .
1999 ( paper v ) , respectively . the applied reduction , calibration and variable selection procedures are discussed therein , particularly in paper i , where full details are provided .
a total of 206 cepheids were found : 43 in field a , 38 in b , 35 in c , 38 in d and 52 in field f. the archival _ hst_-wfpc2
data used in this paper were retrieved from the hubble data archive .
we selected images overlapping our m31 ground - based data , taken in filters f336w ( roughly @xmath16 ) , f439w , f450w ( @xmath17 ) , f555w , f606w ( @xmath18 ) and f814w ( @xmath19 ) .
the pixel scales of the wide field ( wf ) and planetary camera ( pc ) chips are 0.0996 and 0.0455 /pixel , respectively .
the full list of exposures is provided in table [ tab : dat ] , along with the proposal i d , dataset name , equatorial coordinates of the frame centers , filter and exposure time information .
llcccc 5911 & u2y20103 t & 00 44 44.17 & 41 27 33.88 & f336w & 400 + 5911 & u2y20105 t & 00 44 44.23 & 41 27 33.86 & f439w & 160 + 5911 & u2y20106 t & 00 44 44.23 & 41 27 33.86 & f555w & 140 + 5911 & u2y20203 t & 00 44 49.28 & 41 28 59.04 & f336w & 400 + 5911 & u2y20205 t & 00 44 49.34 & 41 28 59.03 & f439w & 160 + 5911 & u2y20206 t & 00 44 49.34 & 41 28 59.03 & f555w & 140 + 5911 & u2y20303 t & 00 44 57.57 & 41 30 51.68 & f336w & 400 + 5911 & u2y20305 t & 00 44 57.63 & 41 30 51.65 & f439w & 160 + 5911 & u2y20306 t & 00 44 57.63 & 41 30 51.65 & f555w & 140 + 5911 & u2y20403 t & 00 45 09.20 & 41
34 30.56 & f336w & 400 + 5911 & u2y20405 t & 00 45 09.25 & 41 34 30.72 & f439w & 160 + 5911 & u2y20406 t & 00 45 09.25 & 41 34 30.72 & f555w & 140 + 5911 & u2y20503 t & 00 45 11.89 & 41 36 56.86 & f336w & 400 + 5911 & u2y20505 t & 00 45 11.95 & 41 36 57.02 & f439w & 160 + 5911 & u2y20506 t & 00 45 11.95 & 41 36 57.02 & f555w & 140 + 6038 & u2ye0603 t & 00 44 51.22 & 41 30 03.72 & f555w & 160 + 6038 & u2ye0605 t & 00 44 51.22 & 41 30 03.72 & f439w & 600 + 6038 & u2ye060bt & 00 44 51.22 & 41 30 03.72 & f336w & 900 + 6038 & u2ye060dt & 00 44 51.22 & 41 30 03.72 & f336w & 900 + 5237 & u2ab0101 t & 00 40 29.40 & 40 43 58.28 & f555w & 200 + 5237 & u2ab0102 t & 00 40 29.40 & 40 43 58.28 & f555w & 200 + 5237 & u2ab0103 t & 00 40 29.40 & 40 43 58.28 & f814w & 200 + 5237 & u2ab0104 t & 00 40 29.40 & 40 43 58.28 & f814w & 200 + 5237 & u2ab0105 t & 00 40 29.40 & 40 43 58.28 & f336w & 600 + 5237 & u2ab0106 t & 00 40 29.40 & 40 43 58.28 & f336w & 600 + 5494 & u2g20701 t & 00 40 33.17 & 40 45 38.97 & f606w & 350 + 8059 & u4woah02r & 00 40 10.11 & 40 46 08.91 & f606w & 400 + 8059 & u4woah04r & 00 40 10.11 & 40 46 08.91 & f450w & 600 + 8059 & u4woah05r & 00 40 10.11 & 40 46 08.91 & f814w & 100 + 8059 & u4woah06r & 00 40 10.11 & 40 46 08.91 & f814w & 300 + 8061 & u4x1of01r & 00 40 10.11 & 40 46 08.91 & f606w & 400 + [ tab : dat ] the data we obtained had already passed through the standard preliminary processing and calibration procedures prior to its placement in the archive .
the standard pipeline calibration is fully described in the _ hst _ data handbook .
the first two steps in our reduction procedure were to mark the bad pixels on the images and to compensate for the fact that pixels on the edges and corners of the ccd receive fewer photons due to the geometric distortion in the wfpc2 optics .
for each image a mask was created from the data quality file retrieved from the archive and a vignetting mask generated by stetson ( 1998 ) and then used to mark bad pixels and vignetted regions . to restore the integrity of the flux measurements
the images were multiplied by a pixel - area map , originally created by holtzman et al .
( 1995 ) and renormalized to the median pixel area on each chip by stetson ( 1998 ) .
these tasks were accomplished under iraf . in the next step of the reduction procedure pairs ( or multiplets ) of images
were selected , taken in the same filters , having identical center coordinates as well as similar exposure times .
the * crrej * task under iraf was used to combine the images and remove cosmic rays . when multiple images for a given field were not available , single images were used for photometry .
the photometry was extracted by means of the daophot / allstar package ( stetson 1987 , 1992 ) .
stars were identified using the find subroutine and aperture photometry was done on them with the phot subroutine .
we used the point - spread functions ( psf ) derived individually for each filter and chip of the wfpc2 camera , kindly provided to us by p. b. stetson ( private communication ) to obtain the allstar profile photometry . after the initial allstar run , find was ran again on the star subtracted images , to identify stars that were missed on the first pass .
aperture photometry was obtained for them with phot , followed by profile photometry with allstar .
the two star lists were merged and used as input to allstar to obtain the final photometry .
it should be noted that the _ hst _ photometry has not been calibrated to any standard system and therefore instrumental magnitudes are used throughout this paper .
this has , however , no bearing on the results presented in this paper , since they are strictly based on differential photometry .
the preliminary identification of direct cepheids on the _ hst _ frames was accomplished by a visual comparison of the _ hst _ data matched via world coordinate system ( wcs ) information to our ground - based @xmath0 template image for each field , using saoimage ds9 .
the wcs information for the ground - based template image header was obtained with a program written by mink ( 1997 ; 1999 ) by matching stars from the image to the usno a2.0 catalog stars ( monet et al .
1996 ) . we have matched a total of 22 cepheids to _ hst _ data : six in our field b , four in c ( two of them on two different overlapping fields ) and 12 in field f ( one found on two different fields ) .
this constitutes @xmath20 of the 125 cepheids in fields b , c and f. in some cases , when a cepheid which appeared to be a single star on the @xmath0 template was resolved into multiple stars on an _ hst _ image , it was difficult to distinguish which star was the cepheid on the basis of the template image alone .
two such cases are shown in figure [ fig : cfig ] .
the images plotted in the left panels are taken from the @xmath0 template , created by averaging together two images with the best seeing ( fwhm @xmath21 ) .
the image in the upper right panel was taken with the pc chip of the wfpc2 camera while the one in the lower right panel comes from a wf chip .
the inset has different iraf z - scale limits to show the two blended stars separately .
the circles , centered on the cepheids , have a radius of @xmath22 .
the cepheid v4125 d31f is shown in the upper panels of the figure [ fig : cfig ] .
on the ground - based template it appears as a single star , while on the pc chip it is resolved into two objects , one 3 magnitudes dimmer than the other .
a third star appears below and to the right , just outside the circle , fainter by almost 2 magnitudes from the brighter component of the doublet . that star was not identified on the ground - based template image , but after subtracting the cepheid a brighter spot is visible at that location .
the lower panels show the case of v2203 d31f , where the cepheid , appearing as a single star on the ground @xmath0 template image , is resolved by _ hst _ into two stars , differing by one magnitude in brightness .
to help confirm the cepheid nature of the selected objects , instrumental color - magnitude diagrams ( cmds ) were constructed from _ hst _
data , whenever photometry in two bands was available .
a few representative cmds are shown in figure [ fig : cmd ] , ( @xmath23 ) in the left panels and ( @xmath24 ) in the right panels .
the cepheids are denoted by circles and their companions by squares .
stars from the same chip are plotted in the background for reference .
the upper left panel shows a case where there is a substantial contribution of flux from a red giant companion ( @xmath25 , @xmath26 ) which is seen at a distance of @xmath27 from the cepheid v1893 d31f .
the other companion is a blue star at a distance of @xmath28 which appears to be located at the red edge of the main sequence ( also in the @xmath29 cmd which is not shown ) .
the lower left panel presents a case where the cepheid v2203 d31f has a luminous blue main sequence companion with @xmath30 and @xmath31 which are separated by @xmath32 .
the upper right panel shows the cepheid v7184 d31b with two blue main sequence companions , the more luminous one , located @xmath28 away , contributing @xmath33 as much light as the cepheid in the @xmath34 band and @xmath35 in @xmath0 . a typical situation where none of the companions ( in this case one ) had @xmath36-band photometry ( see table [ tab : cep ] )
is illustrated in the lower right panel of figure [ fig : cmd ] , where the cepheid v4954 d31b is shown .
the criterion for blending was determined empirically , by careful examination of ground - based and _ hst _ data for the same cepheids .
we consider a star to be blended with the cepheid if it is located at a distance of less than @xmath22 from it and is not detected by daophot in our ground - based images .
the choice of maximum distance was motivated by the typical full width at half maximum ( fwhm ) in our ground - based images ( @xmath37 ) . due to the relatively small number of cepheids having _ hst _ data , it was deemed worthwhile to examine them on the _ hst _ images in detail to see whether all of the companions had been identified by daophot and to check for the possibility of false detections ( cosmic rays in case of single images , bad columns , etc . ) . in the latter case
the object was removed from the list .
not much could be done in the former case , although that proved not to be a problem and only in a few of the 55 examined images there were very faint companions which eluded detection .
it must be noted that the fwhm of the psf on the wf chip is of the order of 1 pixel and so the probability of the detection of a faint star may depend on the way its light is distributed over the pixels .
one exception , however , is v1893 d31f , which daophot failed to recognize as a star on one of the single f606w filter images , along with its companion 2 pixels ( @xmath27 ) away .
no problem was encountered on the other three combined images .
we estimate that our data is fairly complete for companions contributing at least @xmath38 of the flux of the cepheid in the v band ( filters f555w and f606w ) .
we have used this somewhat arbitrary cutoff in evaluating the sum @xmath39 of all flux contributions in filter @xmath40 normalized to the flux of the cepheid : @xmath41 where @xmath42 is the flux of the i - th companion , @xmath43 the flux of the cepheid on the _ hst _ image and @xmath44 the total number of companions . in table
[ tab : cep ] we present the results for each of the 22 cepheids found in the _ hst _ images : the name , the mean @xmath0 , @xmath45 and @xmath34 magnitudes taken from papers ii , iii and v , the number of companions @xmath44 and their total flux contribution @xmath39 in the @xmath0 , @xmath45 , @xmath34 and @xmath16 bands respectively .
unless noted otherwise , the @xmath0 , as in @xmath46 and @xmath47 , refers to filter f555w , @xmath45 to f814w , @xmath34 to f439w and @xmath16 to f336w .
for the two cepheids identified on two _ hst _ images , the average values of @xmath39 are listed .
llllllllllll v7184 d31b & 19.15 & 18.45 & & 3 & 0.40 & & & 2 & 0.62 & 3 & 2.55 + v6379 d31b & 20.65 & 19.39 & & 3 & 0.27 & & & 0 & 0.00 & & + v7209 d31b & 20.07 & 19.14 & & 1 & 0.12 & & & 0 & 0.00 & 1 & 0.96 + v5646 d31b & 20.90 & 19.51 & & 1 & 0.05 & & & 0 & 0.00 & & + v4954 d31b & 20.86 & 19.98 & & 1 & 0.10 & & & 0 & 0.00 & 0 & 0.00 + v6872 d31b & 21.60 & 20.62 & & 1 & 0.05 & & & 0 & 0.00 & 0 & 0.00 + v14487 d31c & 19.95 & 18.95 & 20.75 & 2 & 0.12 & & & 0 & 0.00 & 1 & 0.60 + v13705 d31c & 21.50 & 19.98 & 22.83 & 0 & 0.00 & & & & & & + v14661 d31c & 21.39 & 19.70 & 22.76 & 2 & 0.30 & & & 0 & 0.00 & & + v14361 d31c & 21.21 & 19.32 & 22.36 & 3 & 0.27 & & & 0 & 0.00 & & + v4125 d31f & 20.46 & 19.63 & 21.11 & 1&0.06 & 3 & 0.25 & & & 0 & 0.00 + v3289 d31f & 20.89 & 19.76 & 20.77 & 2&0.11 & 3 & 0.27 & & & 1 & 0.53 + v3550 d31f & 20.55 & 19.85 & 21.03 & 2&0.47 & 0 & 0.00 & & & 1 & 6.33 + v1893 d31f & 18.79 & 17.47 & 19.54 & 2&0.41 & 1 & 0.66 & & & 1 & 0.33 + v2203 d31f & 17.94 & 17.17 & 18.32 & 1&0.39 & 1 & 0.17 & & & 3 & 2.88 + v3860 d31f & 21.24 & 19.65 & 22.08 & 0&0.00 & 3 & 0.63 & & & 0 & 0.00 + v3441 d31f & 21.28 & 20.52 & 21.18 & 2&0.26@xmath48 & & & & & & + v2320 d31f & 20.27 & 19.54 & 20.60 & 0&0.00@xmath48 & & & & & & + v1633 d31f & 21.30 & 19.85&&1&0.10@xmath48 & & & & & & + v1599 d31f & 20.94 & 19.86 & 21.78 & 4&0.34@xmath48 & & & & & & + v1549 d31f & 20.72 & 19.68 & 21.49 & 4&0.28@xmath48 & & & & & & + v7074 d31f & 20.78 & 19.70 & 21.51 & 0&0.00@xmath48 & 0 & 0.00 & 0&0.00@xmath49 & & + [ tab : cep ] the catalog of cepheid blending in m31 is illustrated in the following two figures .
figure [ fig : hist ] shows a histogram of the number of cepheids @xmath46 as a function of @xmath47 ( eq . [ eq : sv ] ) , the sum of flux contributions from its companions in the @xmath0 band ( filters f555w and f606w ) , normalized to the flux of the cepheid .
the upper left and right panels and the lower left panel show the histograms for fields b , c and f , respectively . in the lower right panel
the combined data is shown .
the width of each bin is 0.1 and the first one starts at 0 . for further discussion on figure [ fig : hist ] see 5 .
a diagram showing the flux contribution from companions @xmath47 as a function of the @xmath0 magnitude of the cepheid obtained from ground - based data ( papers ii , iii and v ) is presented in figure [ fig : vfv ] .
field b cepheids are denoted by squares , field c by triangles and field f by circles .
the first step in using the derived blending is to see what is the effect on the cepheid period - luminosity ( p - l ) relation .
figure [ fig : pl ] illustrates the influence of blending on the location of the cepheids on the p - l diagram .
the solid circles show the original locations of the cepheids , based on their mean @xmath0 magnitudes obtained from the ground - based data .
the arrows illustrate the shift in @xmath0 when the effects of blending are taken into account .
it has to be kept in mind that neither the differential reddening in m31 nor the random - phase nature of the _ hst _ data have been accounted for , thus there is significant scatter in the diagram .
however , the direction and the magnitude of the effect is obvious .
a pattern in the spatial distribution of cepheid blending in m31 could show us a correlation between blending and crowding .
we find no such correlation cepheids in environments of different surface brightness show roughly the same frequency and amount of blending , thus distinguishing the phenomenon of blending from crowding . in figure
[ fig : sky_sv ] the blending parameter @xmath47 is plotted as a function of the surface brightness around the cepheid for which it was determined .
the surface brightness was taken to be the mode within a 20 pixel radius on the direct template frames .
we used a bin width of 1 adu for the histogram to compute the value of the mode , smoothed with a flat - topped rectangular kernel ( boxcar ) filter , 11 units in length . after correcting for the sky level
, the instrumental surface brightness values were converted to mag / using 12 - 17 fairly bright isolated reference stars with known standard magnitudes by means of the following formula : @xmath50 = m_{ref } + 2.5 \log ( s^2 i_{ref } / i_{sb})\ ] ] where @xmath51 and @xmath52 are the flux in adu/ pixel and the corresponding magnitude of the reference star , @xmath53 is the surface brightness expressed in adu/ pixel and @xmath54 is the pixel scale of the chip .
the @xmath55 scatter around the value of the average ranged from 0.025 to 0.061 mag/. in order to give these numbers some perspective , we have computed surface brightness around cepheids in two galaxies which straddle their range , ngc 2541 and ngc 4535 ( fig . [
fig : sky_sv_cmp ] ) .
we find that the cepheids we discover in m31 reside in environments of surface brightness typical of spiral galaxies , despite the high inclination of m31 .
the _ hst _
data for these galaxies , observed as part of the hst key project on the extragalactic distance scale ( ferrarese et al .
1998 , macri et al . 1999 ) , were obtained from the hubble data archive . since we were unable to estimate the sky level for those images , we tried to choose epochs where the sky level would be the lowest .
we chose the 1995 nov 20 epoch for ngc 3541 and the 1996 may 24 epoch for ngc 4535 .
the surface brightness values were computed using the same method as for our m31 cepheids and were re - expressed in units of mag / using the zeropoints provided in the hst data handbook .
before attempting to draw any far - reaching conclusions from figs . [
fig : hist]-[fig : sky_sv_cmp ] or the catalog itself ( tab .
[ tab : cep ] ) , we must stress that they are based on a fairly small sample of cepheids and therefore subject to statistical uncertainties .
an illustration of this effect is the small number of counts in the second bin of the combined histogram ( fig .
[ fig : hist ] ) , compared to the third and , especially , the first bin , which could be an artifact of small number statistics .
an encouraging fact is that the distributions in separate fields , especially b and f , appear to be similar to a substantial degree in their overall shape and spanned range of @xmath47 . in the combined histogram there is a decreasing trend in the number of cases @xmath46 with increasing flux contribution from blending , @xmath47 .
the diagram in fig .
[ fig : vfv ] conveys some of the information contained in the previous figure a lack of cases with @xmath56 is readily apparent , which , as mentioned above , is believed to result from the fairly small size of the sample .
if we attribute the gap to statistical effects , then cepheids fainter than @xmath57 appear to populate a wide range of @xmath47 , from 0 up to approximately 0.35 , with one object at @xmath58 .
furthermore , it is observed that in none of the fields do cepheids show a tendency to favor any particular value of @xmath47 , as also apparent from figure [ fig : hist ]
. another striking feature of the diagram is that all of the three cepheids brighter than @xmath57 have @xmath59 , although it has to be emphasized that this sample is much too small to draw any definite conclusions .
one possible explanation may be ventured , however , which may hold true in these particular cases , but does not have to be the rule for the whole population of bright cepheids .
the most luminous cepheids are the most massive ones and , hence , the youngest .
the group or association of stars in which they had formed will have had less time to disperse .
this may manifest itself in a somewhat higher spatial density of stars and thus would increase the probability of blending .
figure [ fig : sky_sv ] shows the blending parameter @xmath47 of the m31 cepheids as a function of the local surface brightness . as there is a gap at @xmath60 , the separation into blended ( @xmath61 ) and unblended cepheids ( @xmath62 ) is readily discernible in the diagram .
no correlation between the blending parameter @xmath47 and the underlying surface brightness is apparent . to put the surface brightness values around the m31 cepheids into perspective , we plot the surface brightness values computed for the ngc 4535 and ngc 2541 cepheids ( fig . [
fig : sky_sv_cmp ] ) .
the range of surface brightness of ngc 2541 falls below the range of our m31 cepheids .
there is a 26% overlap between the ngc 4535 and m31 data 13 out of the 50 cepheids in ngc 4535 are located in regions with surface brightness in the range covered by our m31 cepheids .
it should be kept in mind that our sample may also be affected to some extent by selection effects : for example , 11 of our cepheids are located in the vicinity of ngc 206 , the giant star - forming region in m31 .
additionally , m31 is observed at a rather high inclination angle , which is not a typical situation for most galaxies searched for cepheids .
an opportunity to study the effects of blending with a larger sample of @xmath63 cepheids , in a more face - on system , will present itself in the next paper on cepheids in m33 ( mochejska et al .
we will also discuss there the influence of blending on the observed colors of cepheids .
the luminosity variations through the pulsation cycle of a cepheid seen in the optical are due primarily to changes in temperature .
the monochromatic surface brightness changes with temperature and is additionally quite dependent on wavelength .
of course , the cepheid also expands and contracts @xmath1 its area varies .
this contributes to the total light variation in a wavelength independent way .
the radius variation in a typical cepheid is out of phase by about 90@xmath64 with respect to the surface brightness variation .
therefore a comparison between color ( proxy for temperature variation ) and magnitude ( proxy for luminosity variation , due to _ both _ radius and temperature change ) would produce a loop in the cmd plane .
these loops , i.e. the tight correlation between color and luminosity , have had a limited use in the past as part of the cepheid plc calibration ( fernie 1964 ) , or to transform away the temperature - induced variation ( madore 1985 ; moffett & barnes 1986 ) .
the characterization of cepheid companions had been traditionally done by using the similar loops in the color - color plane of optical bands ( stobie 1970 ) .
the color - color loop arises from the phase shift introduced between the pair of color - magnitude relations when a companion of different temperature ( i.e. color ) is added .
blue or red companions distort the color - color loops in very different and quantifiable ways ( madore 1977 ; deyoreo & karp 1979 ) and can be very effective in deriving the amount of and color of contamination in a cepheids flux . in order to understand the wavelength - dependent effect of a blue or red companion on the light curves ( and loops ) of a cepheid , one should bear in mind that the contaminant flux is most prominent during the minimum in the cepheid lightcurve , and gradually becomes insignificant as the cepheid brightens up during its cycle .
the addition of a _ red companion _ to the flux of a cepheid has the following effects on its optical light curves : ( 1 ) the light curve exhibits a flatter minimum ( due to the added flux ) ; ( 2 ) the color curve has a deeper minimum ( due to the added red flux ) ; and ( 3 ) the asymmetry in the color curve decreases .
the addition of a _ blue companion _ to the flux of a cepheid has the following effects on its optical light curves : ( 1 ) the light curve exhibits a flatter minimum ( as above ) ; ( 2 ) the color curve has a flatter minimum ( due to the added blue flux ) ; and ( 3 ) the asymmetry in the color curve increases .
a typical example is the cepheid v7184 d31b ( paper i ) , which has blue blends ( table 2 ) and whose cm loop is shown in fig .
[ fig : blend ] .
the effect on its loop is predictable given the above discussion : as the cepheid fades and becomes naturally redder , the added constant blue flux of the blends becomes more prominent and diverts the lower part of the loop to the blue ( left in fig .
[ fig : blend ] ) .
this diversion could result eventually in a complete reversion of the naturally clockwise trajectory of the loop .
if the data is incomplete ( there are two gaps in the light curves of v7184 ) , the blue blending is still detectable by the steeper slope of the loop . these changes on the color - magnitude plane , together with a fit of the mean colors of the cepheid to extinction laws ,
can be a powerful means to find and characterize blended cepheids in a sample like our direct project one .
the necessary requirement is @xmath1 good , well - sampled light curves in at least two bands .
therefore we were able to find blended cepheids like v7184 in m31b and exclude them from our analysis without the benefit of _ hst _ images , but just using our light curves ( paper i ) .
however , the technique is limited to the cases with strongest blending a nonlinear fitting routine by rebel ( 1998 ) applied to the m31b data
still did not detect 2/3 of the blended cepheids we found in this paper .
in other words , the naive technique of visual inspection for `` flat - bottom '' light curves is completely inadequate for detecting 10%-20% cepheid blending , which is dominating our findings in m31 .
the influence of blending has been largely neglected in many recent galaxy distance determinations based on cepheids .
few attempts have been made to deal with this problem .
on one of the galaxies observed by the hst key project on the extragalactic distance scale , ngc 2541 ( ferrarese et al . 1998 ) , several criteria were used to reject cepheids which may suffer from crowding .
we have applied them to our sample of 22 cepheids to estimate their effectiveness in detecting blending . as we discussed already in 1 , blending and crowding are different phenomena , and judging from 5 they do not seem much correlated in m31 either , but it would be helpful if these criteria could detect blending .
one of the proposed tests ( ferrarese et al . 1998 ) involves identifying cepheids that have companions not resolved by profile photometry programs by looking at their photometric errors .
the underlying idea is that the photometric error would be higher than in case of an isolated star of the same magnitude due to a poorer psf fit resulting from the unresolved companion .
we have constructed diagrams showing the allstar photometric error @xmath65 as a function of the allstar @xmath0 magnitude , similar to the one presented in fig .
8 of ferrarese et al .
the diagram for field f is shown in figure [ fig : als2 ] .
the cepheids are plotted as solid dots surrounded by circles with their size proportional to the amount of blending @xmath47 ( no outside circle means @xmath66 ) .
the @xmath0 magnitudes and the errors are taken from the field f template image ( fwhm@xmath21 ) , as are the stars plotted in the background . upon examining figure [ fig : als2 ] it is apparent that the photometric errors @xmath65 do not show a correlation with the amount of light coming from companions @xmath47 .
one such case is v2320 d31f which is an unblended star with an error almost twice as large than most stars of similar magnitude .
another example is v3289 d31f with @xmath67 , which has a photometric error larger than two other , more blended cepheids of similar @xmath0 magnitude : v1549 d31f with @xmath68 and v3550 d31f with @xmath69 .
most of the cepheids lie within the most densely populated areas of the diagram , with v2320 d31f and v2203 d31f being the most clear outliers . a similar photometric error distribution , uncorrelated with @xmath47 , is also seen in fields b and c. in conclusion , the presence of close companions did not manifest itself in the form of a poorer psf fit and , hence , a larger photometric error in the studied sample .
another fact which should be kept in mind is that our ground - based psf is much better sampled than the _ hst _ psf ( comparing to the respective fwhms ) , thus it should be more sensitive to any deformations resulting from the superposition of two or more stars .
another proposed method of rejecting blended cepheids is to remove all stars having companions contributing more than @xmath70 of the total light within a radius of two pixels .
taking into account the fact that the two pixels on the _ hst _ wf chips correspond to twice the fwhm , we have conducted a similar test on our ground - based @xmath0 template data for the cepheids .
only one of our cepheids , v5646 d31b with @xmath71 would be rejected from our sample on the basis of this criterion .
this test also failed to identify the blended cepheids in our sample .
a third proposed test for blending is to check if the cepheid lies on the instability strip in the cmd .
the usual color range spanned by the final sample of cepheids discovered with _
hst _ is about one magnitude , for example @xmath72 in ngc 2541 ( ferrarese et al .
1998 ) , @xmath73 in ngc 4639 ( saha et al .
1997 ) , @xmath74 in ngc 4535 ( macri et al .
the examination of a ground - based cmd for the 12 field f cepheids yields a @xmath75 color span of also about one magnitude ( @xmath76 ) .
no clear correlation of the color with the @xmath47 of the cepheid is seen in the diagram , as unblended cepheids also exhibit a large color scatter , from @xmath77 for v2320 d31f to 1.59 for v3860 d31f , most likely due to differential reddening . in the case of our sample of 22 cepheids , the tests were not successful in rejecting the blends .
these tests may be helpful in some cases of blending , when the two stars are almost resolved or when the companion has a larger flux contribution than the maximum value of @xmath69 present in our sample .
but a contamination of @xmath69 has a large impact on the photometry of the cepheid , which actually would be dimmer by almost one third from the measured brightness .
for our sample of 22 cepheids with both ground - based and _ hst _ data , we find that the mean @xmath0-band flux contribution from companions unresolved in the ground - based images , @xmath78 , is about @xmath79 of the flux of the cepheid , while the median @xmath47 is 12% .
this shows that blending could potentially be a substantial source of error in the cepheid distance scale , as the distance derived from our ground - based photometry for this admittedly small cepheid sample would be systematically underestimated by @xmath80 ( for the mean @xmath47 ) or @xmath81 ( for the median @xmath47 ) .
this is to be compared to the current cepheid distance to m31 ( which is subject to the lmc distance uncertainty ) of @xmath13 ( freedman & madore 1990 ; kaluzny et al .
1998 ; sasselov et al . 1998 ) .
the cepheid photometry by freedman & madore ( 1990 ) was from the canada - france - hawaii telescope ( average seeing 1.0 , freedman , wilson , & madore 1991 ) .
our findings of blending here require that these distance estimates to m31 be corrected upward by about 9% .
blending becomes even more severe when we consider galaxies at a distance of @xmath82 , i.e. at the edge of what can be currently observed with _
the @xmath12 fwhm on the wf chips of the wfpc2 camera would span a linear distance of @xmath83 in such a galaxy , which corresponds to @xmath84 at the distance of m31 . in order to obtain an estimate as to the degree of contamination caused by blending at a distance of @xmath82
, we have summed the contributions of all cepheid companions within a diameter of @xmath84 in our _ hst _ data for m31 .
the result is presented in figure [ fig : ks ] in the form of a cumulative probability distribution of @xmath47 .
the diagram shows that @xmath70 of our cepheids have @xmath85 , though some of the cepheids with a very high degree of contamination would probably elude detection .
this indicates that blending will very likely introduce a significant contamination to cepheid photometry at such distances and resolution . as the result of blending with other unresolved stars ,
the cepheids appear brighter than they really are when observed in distant galaxies with _ hst_. as we compare them with mostly unblended lmc cepheids , this leads to systematically low distances to galaxies observed with the _ hst _ , and therefore to systematically high estimates of @xmath3 .
the sign of the blending effect on the @xmath3 is opposite to that caused by the lower lmc distance ( e.g. udalski 1998 ; stanek et al .
2000 ) and might be of comparable value , as discussed in this paper .
it should be stressed that blending is a factor which contributes in only one direction , and therefore it will not average out when a large sample of galaxies is considered . one obvious solution to the problem of blending would be to obtain data with better resolution .
such an opportunity will be available after the launch of the next generation space telescope , scheduled for 2008 .
an alternative approach would be to determine the amount and color of flux contamination for the cepheids by the analysis of their light curves and/or color - color loops .
the requirement , however , is to have good quality , well - sampled light curves in at least two bands , which is not the case for _ hst _ cepheids .
however , even with our good light curves it is difficult to detect cepheids with blending of @xmath86 .
it must be stressed that the images taken with the _ hst _ wfpc2 camera , despite having a resolution @xmath87 times better than our ground - based images ( fwhm @xmath88 vs. @xmath89 ) , will still leave some blends unresolved , including physical companions ( e.g. evans 1992 ) .
thus , the @xmath0-band flux contribution from cepheid companions derived in this paper sets only the lower limit on the true influence of blending on the cepheid photometry . | we investigate the influence of blending on the cepheid distance scale .
blending is the close association of a cepheid with one or more intrinsically luminous stars .
high - resolution _ hst _ images are compared to our ground - based data , obtained as part of the direct project , for a sample of 22 cepheids in the m31 galaxy . the average ( median ) @xmath0-band flux contribution from luminous companions which are not resolved on the ground - based images is about 19% ( 12% ) of the flux of the cepheid .
this is a large effect @xmath1 at the 10% level for distances .
the current cepheid distance estimates to m31 are all ground - based , and are thus affected ( underestimated ) .
we discuss indirect methods to find which cepheids are blended , e.g. by the use of well - sampled light curves in at least two optical bands . more generally , our ground - based resolution in m31 corresponds to the _ hst _ resolution at about @xmath2 . blending leads to systematically low distances in the observed galaxies , and therefore to systematically high estimates of @xmath3
; we discuss the issue and the implications . |
brownian motion in periodic and random potentials has been a very active area of research for many decades .
apart from the well established applications to electronics @xcite and to solid state physics such as superionic conductors , the josephson tunneling junction @xcite and surface diffusion @xcite , new and exciting applications to physics ( self - assembled molecular film growth , catalysis , surface - bound nanostructures ) and to biology ( stochastic modeling of molecular and brownian motors @xcite ) keep the subject of brownian motion at the forefront of current research , both theoretical and experimental . despite the fact that brownian motion in periodic potentials has been studied extensively ( * ? ? ?
11 ) , @xcite and many analytical and numerical results have been obtained , there are still many open questions , in particular in the underdamped , multidimensional case .
the main purpose of this paper is to develop a general method for calculating the diffusion tensor @xmath0 of the brownian particle in arbitrary dimensions , and to then use this method for setting up an efficient numerical method for computing @xmath0 .
furthermore , we will show that our method for calculating @xmath0 will enable us to study various asymptotic limits of physical interest in a systematic and rigorous fashion .
the dynamics of a brownian particle moving in a periodic potential is governed by the langevin equation @xmath1 where @xmath2 denotes the particle position , @xmath3 is a smooth periodic potential , @xmath4 denotes the friction coefficient and @xmath5 is a white noise gaussian process with correlation function @xmath6 where @xmath7 is boltzmann s constant and @xmath8 is the absolute temperature , in accordance with the fluctuation - dissipation theorem .
we use the notation @xmath9 to denote ensemble average . we will also write @xmath10 and @xmath11 , where @xmath12 is a standard brownian motion in @xmath13 . throughout this work
we will assume that the diffusing particle is of unit mass , @xmath14 .
the langevin equation has been studied extensively as a theoretical model for the diffusion of adsorbates on crystal surfaces @xcite . in this
setting , @xmath15 represents the position of the diffusing particle , @xmath16 the substrate potential and the friction and noise terms represent the interaction of the diffusing particle with the phonon heat bath @xcite it is well known that at low temperatures ( which is usually the regime of physical interest ) , the diffusing particle performs a hopping motion ( random walk ) between the local minima of the potential .
this hopping motion is characterized by the mean square jump length @xmath17 and the hopping rate @xmath18 ( or , equivalently , the mean escape time @xmath19 with @xmath20 ) .
these two quantities are related to the diffusion coefficient through the formula ( in one dimension ) @xmath21 knowledge of two of the three quantities @xmath22 is sufficient for the calculation of the third .
of course , the diffusion tensor can also be defined either in terms of the ensemble average of the second moment , i.e. @xmath23 or in terms of the time integral of the velocity autocorrelation function , i.e. through the green - kubo formula @xmath24 with @xmath25 . in the above formulas @xmath26 stands for the tensor product between two vectors in @xmath13 .
we remark that , although formulas and are valid in arbitrary dimensions , it is not clear how to interpret formula in dimensions higher than @xmath27 . of particular interest
is the dependence of the diffusion coefficient @xmath0 as well as the jump rate @xmath18 and ( mean square ) jump length @xmath28 on the friction coefficient . in particular , it is well known that in the underdamped regime the particle diffusion is dominated by the occurrence of long jumps @xcite .
apart from theoretical investigations and numerical simulations based on the langevin dynamics , the occurrence of long jumps in the underdamped regime is also verified by means of molecular dynamics simulations of a model for co / ni@xmath29 @xcite and is also , by now , a well established experimental result @xcite .
indeed , in the case of weak adsorbate - substrate interaction i.e. in the case of weak coupling between the diffusing particle and the heat bath which corresponds to the underdamped dynamics regime , the diffusion mechanism is controlled by long jumps , spanning multiple lattice spacings @xcite .
the rigorous and systematic study of the diffusion process in the weak dissipation regime is still a major challenge for theoreticians , in particular in dimensions higher than @xmath27 . the problem of diffusion in periodic potentials is well studied in one dimension @xcite or for separable potentials in two and three dimensions @xcite
kramers theory @xcite applied to brownian motion in a periodic potential or the mean first passage time method @xcite enables us to calculate the rate of escape from a local minimum of the potential ( hopping rate ) . in the overdamped limit
this is sufficient to calculate the diffusion coefficient , since only single jumps occur and consequently @xmath30 where @xmath31 is the period of the potential .
this leads to the well known lifson - jackson formula for the diffusion coefficient @xcite ( * ? ? ?
* ch . 13 ) @xmath32 where @xmath33 and @xmath34 denotes the diffusion coefficient of the free particle @xmath35 kramers formula for the rate of escape enables us to obtain a formula for the diffusion coefficient which is valid in the moderate - to - strong friction regime , for small temperatures @xmath36
@xcite : @xmath37 where @xmath38 , small temperature @xmath39 , limit this formula reduces to the small temperature asymptotics of equation : @xmath40 the calculation of the diffusion coefficient in the underdamped limit requires two calculations , that of the hopping rate and that of the mean squared jump length @xmath41 .
such a calculation was presented in @xcite for the case of a cosine potential . for this potential , a formula for the diffusion coefficient which is valid in the regime @xmath42
was also obtained by risken and presented in his monograph @xcite : @xmath43 in contrast to the one dimensional problem , a similar theory in higher dimensions is still lacking , except for the overdamped limit . in this limit
approximate analytical results for certain two - dimensional potentials have been derived in the literature @xcite .
furthermore , it is possible to prove , using rigorous mathematical analysis , that the diffusion tensor scales like @xmath44 when @xmath45 @xcite . on the contrary , it is still not clear what the scaling of the ( trace of the ) diffusion tensor with the friction constant is in the underdamped , multidimensional case .
numerical experiments @xcite suggest that this scaling depends crucially on the detailed properties of the periodic potential ; however , a rigorous and systematic theory for explaining the dependence of the diffusion coefficient on the strength of the dissipation in arbitrary dimensions is still lacking .
the situation becomes even more unclear when an external driving force ( either constant or periodic in time ) is present .
brownian motion in tilted periodic potentials has only been studied in one dimension @xcite and explicit formulas are only valid in the overdamped limit @xcite .
furthermore , numerical experiments @xcite seem to indicate that stochastic resonance in periodic potentials in only possible in dimensions higher than one .
in view of the ubiquity of diffusive motion in periodic potentials in applications , it seems to be important to develop the multidimensional theory in a systematic and rigorous way , taking into account external driving forces .
this paper is a contribution towards this goal , and it is a part of our research program on the study of brownian motion in periodic and random potentials @xcite .
all studies on the problem of brownian motion in a periodic potential that have been reported in the physics literature rely heavily on the analysis of the fokker - planck ( kramers - chandrashekhar ) equation which governs the evolution of the transition probability density for the brownian particle .
for example , the continued fraction expansion method is used in order to solve the fokker - planck equation @xcite in a semi - analytic fashion and to calculate quantities such as the mobility , the intermediate structure function and the dynamic structure function @xcite . or , kramers theory is being used , which again relies on the study of the fokker - planck equation .
on the other hand , many tools from stochastic analysis @xcite , the theory of limit theorems for markov processes @xcite and the emerging field of multiscale analysis @xcite are appropriate for the study of this problem and , yet , they have received very little attention in the physics community .
the purpose of this paper is to use multiscale methods such as homogenization theory and singular perturbation theory in order to offer a new insight into the problem of brownian motion in periodic potentials . in particular , we derive in a rigorous and systematic way a formula for the diffusion tensor @xmath0 of a brownian particle moving in a periodic potential in arbitrary dimensions and then we use in order to develop an efficient numerical method for calculating @xmath0 .
this numerical method is related to the continued fraction expansion method , but is easier to implement and to analyze . as a byproduct of our analysis ,
we derive rigorously a formula for the diffusion coefficient which is valid in the weak friction limit ; furthermore , we also calculate higher order corrections to the large @xmath4 asymptotics of the diffusion coefficient .
the rest of the paper is organized as follows . in section [ sec : multiscale ] we present the multiscale analysis and we derive a formula for the diffusion coefficient .
we also show the equivalence between our formula and the green - kubo formula . in section
we derive formulas which are valid in the @xmath46 and @xmath47 limits . in section [
sec : numerical ] we develop the numerical method and we compare the numerical results obtained using our method with results obtained from monte carlo simulations , from approximate analytical formulas and from the numerical implementation of formula .
section [ sec : conclusions ] is reserved for conclusions .
in this section we use multiscale analysis @xcite to derive a formula for the diffusion tensor of a brownian particle moving in a periodic potential in arbitrary dimensions .
we then show the equivalence between this formula and the green - kubo formula for the diffusion tensor .
we start by rescaling the langevin equation @xmath48 where we have set @xmath49 .
we will assume that the potential is periodic with period @xmath50 in every direction .
since we expect that at sufficiently long length and time scales the particle performs a purely diffusive motion , we perform a diffusive rescaling to the equations of motion : @xmath51 , @xmath52 . using the fact that @xmath53 in law we obtain : @xmath54 introducing @xmath55 and @xmath56 we write this equation as a first order system : @xmath57 with the understanding that @xmath58^d$ ] and @xmath59 our goal now is to eliminate the fast variables @xmath60 and to obtain an equation for the slow variable @xmath61 .
we shall accomplish this by studying the corresponding backward kolmogorov equation using singular perturbation theory for partial differential equations .
let @xmath62 where @xmath63 denotes the expectation with respect to the brownian motion @xmath12 in the langevin equation and @xmath64 is a smooth function .
] where @xmath65 is the solution of the fokker - planck equation and @xmath66 is the initial distribution . ]
the evolution of the function @xmath67 is governed by the backward kolmogorov equation associated to equations is @xcite @xmath68 where : @xmath69 the invariant distribution of the fast process @xmath70 in @xmath71}^d \times { { \mathbb r } } ^d$ ] is the maxwell - boltzmann distribution @xmath72}^d \times { { \mathbb r } } ^d } e^{-\beta h(q , p ) } \ , dq dp,\ ] ] where @xmath73 .
indeed , we can readily check that @xmath74 where @xmath75 denotes the fokker - planck operator which is the @xmath76-adjoint of the generator of the process @xmath77 : @xmath78 the null space of the generator @xmath77 consists of constants in @xmath79 .
moreover , the equation @xmath80 has a unique ( up to constants ) solution if and only if @xmath81}^d \times { { \mathbb r } } ^d } g(q , p ) \rho_{\beta}(q , p ) \ , dq dp = 0.\ ] ] equation is equipped with periodic boundary conditions with respect to @xmath82 and is such that @xmath83}^d \times { { \mathbb r } } ^d } |f|^2 \mu_{\beta } \ , dq dp < \infty.\ ] ] these two conditions are sufficient to ensure existence and uniqueness of solutions ( up to constants ) of equation @xcite .
we assume that the following ansatz for the solution @xmath84 holds : @xmath85 with @xmath86 being @xmath50 periodic in @xmath87 and satisfying condition .
we substitute ( [ eq : expansion ] ) into ( [ eq : backw_kolmog_2 ] ) and equate equal powers in @xmath88 to obtain the following sequence of equations : @xmath89 [ eq : oep ] from the first equation in ( [ eq : oep ] ) we deduce that @xmath90 , since the null space of @xmath77 consists of functions which are constants in @xmath91 and @xmath87 . now the second equation in ( [ eq : oep ] ) becomes : @xmath92 since @xmath93 , the right hand side of the above equation is mean - zero with respect to the maxwell - boltzmann distribution .
hence , the above equation is well - posed .
we solve it using separation of variables : @xmath94 with @xmath95 this poisson equation is posed on @xmath71}^d \times { { \mathbb r } } ^d$ ] . the solution is periodic in @xmath87 and satisfies condition .
now we proceed with the third equation in .
we apply the solvability condition to obtain : @xmath96}^d \times { { \mathbb r } } ^d } \mathcal{l}_1 u_1 \rho_{\beta}(p , q ) \ , dp dq \\ & = & \sum_{i , j=1}^d \left ( \int_{{[-\pi , \pi]}^d \times { { \mathbb r } } ^d } p_i \phi_j \rho_{\beta}(p , q ) \ , dp dq \right ) { \frac{\partial^2 u_0}{\partial { x_i } \partial{x_j}}}.\end{aligned}\ ] ] this is the backward kolmogorov equation which governs the dynamics on large scales .
we write it in the form @xmath97 where the effective diffusion tensor is @xmath98}^d \times { { \mathbb r } } ^d } p_i \phi_j \rho_{\beta}(p , q ) \ , dp dq , \quad i , j=1 , \dots d. \label{eqn : eff_diff}\ ] ] the calculation of the effective diffusion tensor requires the solution of the boundary value problem and the calculation of the integral in .
the limiting backward kolmogorov equation is well posed since the diffusion tensor is nonnegative .
indeed , let @xmath99 be a unit vector in @xmath13 .
we calculate ( we use the notation @xmath100 and @xmath101 for the euclidean inner product . ) @xmath102 where an integration by parts was used .
thus , from the multiscale analysis we conclude that at large lenght / time scales the particle which diffuses in a periodic potential performs and effective brownian motion with a nonnegative diffusion tensor which is given by formula .
we mention in passing that the analysis presented above can also be applied to the problem of brownian motion in a tilted periodic potential .
the langevin equation becomes @xmath103 where @xmath3 is periodic with period @xmath50 and @xmath104 is a constant force field .
the formulas for the effective drift and the effective diffusion tensor are @xmath105}^d } p \rho(q , p ) \ , dq dp , \quad d = \int_{{{\mathbb r } } ^d \times { [ -\pi , \pi]}^d } ( p - v ) \otimes \phi \rho(p , q ) \ , dp dq,\ ] ] where [ e : cell_fp ] @xmath106 @xmath107}^d } \rho(p , q ) \ , dp dq = 1.\ ] ] with @xmath108 we have used @xmath26 to denote the tensor product between two vectors ; @xmath109 denotes the @xmath76-adjoint of the operator @xmath110 , i.e. the fokker - planck operator .
equations are equipped with periodic boundary conditions in @xmath87 .
the solution of the poisson equation is also taken to be square integrable with respect to the invariant density @xmath111 : @xmath112}^d } |\phi(q , p)|^2 \rho(p , q ) \ , dp dq < + \infty.\ ] ] the diffusion tensor is nonnegative definite .
a calculation similar to the one used to derive shows the positive definiteness of the diffusion tensor : @xmath113 for every vector @xmath99 in @xmath13 .
the study of diffusion in a tilted periodic potential , in the underdamped regime and in high dimensions , based on the above formulas for @xmath114 and @xmath0 , will be the subject of a separate publication .
let us now show that the formula for the diffusion tensor obtained in the previous section , equation , is equivalent to the green - kubo formula . to simplify the notation
we will prove the equivalence of the two formulas in one dimension .
the generalization to arbitrary dimensions is immediate .
let @xmath115 with @xmath116 and initial conditions @xmath117 be the solution of the langevin equation @xmath118 where @xmath5 stands for gaussian white noise in one dimension with correlation function @xmath119 we assume that the @xmath120 process is stationary , i.e. that the initial conditions are distributed according to the maxwell - boltzmann distribution @xmath121 the velocity autocorrelation function is ( * ? ? ?
2.10 ) @xmath122 and @xmath65 is the solution of the fokker - planck equation @xmath123 where @xmath124 we rewrite in the form @xmath125 the function @xmath126 satisfies the backward kolmogorov equation which governs the evolution of observables ( * ? ? ?
6 ) @xmath127 we can write , formally , the solution of as @xmath128 we combine now equations and to obtain the following formula for the velocity autocorrelation function @xmath129 we substitute this into the green - kubo formula to obtain @xmath130 where @xmath131 is the solution of the poisson equation . in the above derivation we have used the formula @xmath132 , whose proof can be found in ( * ? ? ?
in this section we derive approximate formulas for the diffusion coefficient which are valid in the overdamped @xmath133 and underdampled @xmath42 limits .
the derivation of these formulas is based on the asymptotic analysis of the poisson equation . in this section
we will take the period of the potential is @xmath134 . in this subsection
we solve the poisson equation in one dimension perturbatively for small @xmath4 .
we shall use singular perturbation theory for partial differential equations .
the operator @xmath135 that appears in can be written in the form @xmath136 where @xmath137 stands for the ( backward ) liouville operator associated with the hamiltonian @xmath138 and @xmath139 for the generator of the ou process , respectively : @xmath140 we expect that the solution of the poisson equation scales like @xmath141 when @xmath42 .
thus , we look for a solution of the form @xmath142 we substitute this ansatz in to obtain the sequence of equations @xmath143 from equation we deduce that , since the @xmath144 is in the null space of the liouville operator , the first term in the expansion is a function of the hamiltonian @xmath145 : @xmath146 now we want to obtain an equation for @xmath144 by using the solvability condition for . to this end , we multiply this equation by an arbitrary function of @xmath82 , @xmath147 and integrate over @xmath91 and @xmath87 to obtain @xmath148 we change now from @xmath149 coordinates to @xmath150 , so that the above integral becomes @xmath151 where @xmath152 is the jacobian of the transformation .
operator @xmath135 , when applied to functions of the hamiltonian , becomes : @xmath153 hence , the integral equation for @xmath154 becomes @xmath155 \frac{1}{p(z , q ) } \ , dz dq = 0.\ ] ] let @xmath156 denote the critical energy , i.e. the energy along the separatrix ( homoclinic orbit ) .
we set @xmath157 where risken s notation @xcite has been used for @xmath158 and @xmath159 .
we need to consider the cases @xmath160 and @xmath161 separately .
we consider first the case @xmath162 . in this case @xmath163 .
we can perform the integration with respect to @xmath87 to obtain @xmath164 \
, dz = 0,\ ] ] this equation is valid for every test function @xmath165 , from which we obtain the following differential equation for @xmath144 : @xmath166 where primes denote differentiation with respect to @xmath82 and where the subscript @xmath167 has been dropped for notational simplicity .
a similar calculation shows that in the regions @xmath168 and @xmath169 the equation for @xmath144 is @xmath170 and @xmath171 equations , , are augmented with condition and a continuity condition at the critical energy @xcite @xmath172 where @xmath173 are the solutions of equations , and , respectively .
the average of a function @xmath174 can be written in the form @xcite @xmath175 where the partition function is @xmath176 from equation we deduce that @xmath177 .
furthermore , we have that @xmath178 . these facts , together with the above formula for the averaging with respect to the boltzmann distribution , yield : @xmath179 to leading order in @xmath4 , and where @xmath154 is the solution of the two point boundary value problem .
we remark that if we start with formula @xmath180 for the diffusion coefficient , we obtain the following formula , which is equivalent to : @xmath181 now we solve the equation for @xmath154 ( for notational simplicity , we will drop the subscript @xmath167 ) .
using the fact that @xmath182 , we rewrite as @xmath183 this equation can be rewritten as @xmath184 condition implies that the derivative of the unique solution of is @xmath185 we use this in , together with an integration by parts , to obtain the following formula for the diffusion coefficient : @xmath186 we emphasize the fact that this formula is exact in the limit as @xmath187 and is valid for all periodic potentials and for all values of the temperature .
consider now the case of the nonlinear pendulum @xmath188 .
the partition function is @xmath189 where @xmath190 is the modified bessel function of the first kind .
furthermore , a simple calculation yields @xmath191 where @xmath192 is the complete elliptic integral of the second kind .
the formula for the diffusion coefficient becomes @xmath193 we use now the asymptotic formula @xmath194 and the fact that @xmath195 to obtain the small temperature asymptotics for the diffusion coefficient : @xmath196 which is precisely formula , obtained by risken . unlike the overdamped limit which is treated in the next section , it is not straightforward to obtain the next order correction in the formula for the effective diffusivity
this is because , due to the discontinuity of the solution of the poisson equation along the separatrix .
in particular , the next order correction to @xmath131 when @xmath42 is of @xmath197 , rather than @xmath198 as suggested by ansatz . upon combining the formula for the diffusion coefficient and the formula for the hopping rate from kramers theory (
4.48(a ) ) we can obtain a formula for the mean square jump length at low friction .
for the cosine potential , and for @xmath39 , this formula is @xmath199 in this subsection we study the large @xmath4 asymptotics of the diffusion coefficient . as in the previous case ,
we use singular perturbation theory , e.g. ( * ? ? ?
the regularity of the solution of when @xmath45 will enable us to obtain the first two terms in the @xmath44 expansion without any difficulty .
we set @xmath200 .
the differential operator @xmath135 becomes @xmath201 we look for a solution of in the form of a power series expansion in @xmath4 : @xmath202 we substitute this into and obtain the following sequence of equations : [ e : gamma_large ] @xmath203 the null space of the ornstein - uhlenbeck operator @xmath135 consists of constants in @xmath91 .
consequently , from the first equation in we deduce that the first term in the expansion in independent of @xmath91 , @xmath204 .
the second equation becomes @xmath205 let @xmath206 be the invariant distribution of the ou process ( i.e. @xmath207 ) .
the solvability condition for an equation of the form @xmath208 requires that the right hand side averages to @xmath167 with respect to @xmath209 , i.e. that the right hand side of the equation is orthogonal to the null space of the adjoint of @xmath139 . this condition is clearly satisfied for the equation for @xmath210
thus , by fredholm alternative , this equation has a solution which is @xmath211 where the function @xmath212 of is to be determined .
we substitute this into the right hand side of the third equation to obtain @xmath213 from the solvability condition for this we obtain an equation for @xmath214 : @xmath215 together with the periodic boundary conditions .
the derivative of the solution of this two - point boundary value problem is @xmath216 the first two terms in the large @xmath4 expansion of the solution of equation are @xmath217 where @xmath214 is the solution of . substituting this in the formula for the diffusion coefficient and using
we obtain @xmath218 where @xmath219 .
this is , of course , the lifson - jackson formula which gives the diffusion coefficient in the overdamped limit @xcite .
continuing in the same fashion , we can also calculate the next two terms in the expansion . from this
, we can compute the next order correction to the diffusion coefficient .
the final result is @xmath220 where @xmath221 . in the case of the nonlinear pendulum
, @xmath222 , formula gives @xmath223 where @xmath224 is the modified bessel function of the first kind .
in the multidimensional case , a similar analysis leads to the large gamma asymptotics : @xmath225 where @xmath99 is an arbitrary unit vector in @xmath13 and @xmath34 is the diffusion coefficient for the smoluchowski ( overdamped ) dynamics : @xmath226 where @xmath227 and @xmath228 is the solution of the pde @xmath229 with periodic boundary conditions .
in order to calculate the diffusion coefficient @xmath0 we have to solve equation and calculate the integral in .
it is possible to do this by means of a spectral method . in particular , by expanding the solution of into hermite polynomials in @xmath91 and a standard fourier series in @xmath87 , we can convert the boundary value problem into an infinite system of linear equations ; upon truncating this system we can obtain a finite dimensional , sparse system of linear equations which we can easily solve . for simplicity we will consider the problem in one dimension , though our numerical method works in arbitrary dimensions .
we look for a solution of in the form @xmath230 where @xmath231 denotes the @xmath232th eigenfunction of the ornstein - uhlenbeck operator @xmath233 ; the corresponding eigenvalue is @xmath234 .
the eigenfunctions of the ou processes are related to the hermite polynomials through the formula ( * ? ? ?
5 ) @xmath235 where @xmath236 we remark that the solution of the poisson equation is defined up to a constant which we have taken to be @xmath167 ( i.e. we assume that the solution averages to zero with respect to the maxwell - boltzmann distribution ) .
it is elementary to check that the choice of this constant does not effect the value of the diffusion coefficient .
notice also that the boundary conditions ( the solution is periodic in @xmath87 and square integrable with respect to the maxwell - boltzmann distribution ) have already been taken into account when writing . using the fact that the @xmath237 , together with the orthonormality of the eigenfunctions of the ornstein - uhlenbeck operator , we obtain the following formula for the diffusion coefficient : @xmath238 for the cosine potential @xmath239 the above formula becomes @xmath240 now we shall obtain a linear system of equations for the coefficients @xmath241 for the cosine potential .
the poisson equation in one dimension is @xmath242 we substitute the expansion into and use the properties of hermite polynomials and of trigonometric functions to obtain the following linear system of equations [ e : lin_syst ]
@xmath243 @xmath244 @xmath245 [ fig : deff ] we truncate by taking into account the first @xmath246 terms of the hermite expansion and the first @xmath247 terms of the fourier expansion .
this leads to a sparse linear system of @xmath248 equations and @xmath249 variables ( note the absence of the term @xmath250 ) .
this implies that one of the equations in is linearly dependent on the others ; to obtain a nonsingular system we remove the equation for @xmath251 . with respect to the canonical ensemble . ] in this way we obtain a nonsingular system of @xmath252 equations and unknowns .
this system can be written in the form @xmath253 with @xmath254 and @xmath255 is the coefficient of @xmath256 in the @xmath257 equation , which is taken from equations for values of @xmath232 and @xmath234 @xmath258 i=\ { ll k + k + 1 & n=0 , k
< 0 + n ( 2k+1 ) + k+k & n=0 , k > 0@xmath259n0 + .
@xmath258 the sparsity of this system implies that we can solve it very efficiently . in particular
, we can calculate accurately the diffusion coefficient with a minimal computational cost , even for very small values of the friction constant @xmath4 . to illustrate the efficiency of our numerical method
, we calculate the diffusion coefficient for the cosine potential as a function of the dissipation @xmath4 , at a fixed temperature @xmath260 .
we compare the results obtained through our numerical method with the approximate analytical expressions and @xmath261 , results from monte calro simulations using and results obtained through numerical calculation of the hopping rate and the jump length distribution , equation . for the calculation of @xmath262 and @xmath18 we generate a long path ( for every @xmath4 ) of the langevin dynamics using the milstein scheme .
the results of the numerical simulations are presented in figure [ fig : deff ] .
there agreement between the approximate analytical formulas and the calculation of the diffusion coefficient using the method described in this section are excellent .
our method is far superior in comparison to monte carlo simulations or the calculation of the mean square jump length and the hopping rate , since even for very small @xmath4 the solution of a rather small linear system of equations is required
. decreases .
however , even for @xmath4 very small , the resulting linear system of equations is small enough so that it can be solved in a few seconds in matlab . ]
on the other hand , contrast , the path of integration over which we have to solve the langevin equation in order to compute accurate statistics increases as @xmath4 decreases and the calculation of @xmath0 using monte carlo becomes computationally expensive .
the problem of brownian motion in a periodic potential in arbitrary dimensions was studied in this paper .
using multiscale techniques @xcite we derived a formula for the effective diffusion tensor which is valid for all values of the friction coefficient and the temperature , and in arbitrary dimensions .
we also showed the equivalence between our formula and the green - kubo formula .
the calculation of the diffusion tensor using our approach requires the solution of a poisson equation together with the calculation of the average of an appropriate function with respect to the canonical distribution .
furthermore , the overdamped and underdamped limits where studied and approximate analytical formulas for the diffusion coefficient in these two limits where obtained .
in addition , a very efficient numerical method for the calculation of the diffusion coefficient was developed ; this numerical method is based on the solution of the poisson equation via a spectral method and it leads to the accurate and very efficient calculation of the diffusion coefficient even for very low values of the friction coefficient . the approach developed in the paper for the study of the problem of diffusion in periodic potentials offers various advantages over other analytical and numerical methods .
first , all the results reported in this paper can be justified rigorously and they can also lead to a rigorous analysis of the dependence of the diffusion tensor on the friction coefficient and on the temperature . a first step in this direction was taken in @xcite .
second , our method enables us to study various distinguished limits of physical interest ( such as the overdamped and underdamped limits ) in a systematic fashion through asymptotic analysis of the poisson equation .
third , it leads to an efficient numerical method for calculating the diffusion tensor through the numerical solution of the poisson equation .
the effectiveness of our method was shown in this paper for the one dimensional problem , for which analytical approximate formulas are can be derived .
the method is also very efficient in two and three dimensions and can offer insight into the problem of brownian motion in periodic potentials in higher dimensions .
a thorough numerical investigation of the multidimensional problem will be presented elsewhere .
the authors are grateful to igor goychuk for useful suggestions and comments . | in this paper we present a systematic and rigorous method for calculating the diffusion tensor for a brownian particle moving in a periodic potential which is valid in arbitrary dimensions and for all values of the dissipation .
we use this method to obtain an explicit formula for the diffusion coefficient in one dimension which is valid in the underdamped limit , and we also obtain higher order corrections to the lifson - jackson formula for the diffusion coefficient in the overdamped limit . a numerical method for calculating the diffusion coefficient
is also developed and is shown to perform extremely well for all values of the dissipation . |
stannic oxide ( sno@xmath0 ) is a wide - gap semiconductor ( @xmath3 = 3.6 ev @xcite ) which is used in gas - sensing devices @xcite , and is also of interest as an oxidation catalyst @xcite .
these applications have stimulated intensive study of its surface properties @xcite , concentrating mainly on the ( 110 ) surface ( see fig .
1 ) , which is the most stable . because of the variable valence of sn , the material readily loses surface oxygen , and there have been a number of studies of the atomic structure , electronic structure and electrical properties of the ( 110 ) surface as a function of temperature and oxygen partial pressure .
the existence of a rather complicated series of surface reconstructions @xcite , marked changes of electronic structure in the band gap @xcite , and strong variation of the surface electrical conductivity with deviation from stoichiometry @xcite are all well documented . in spite of this work , many questions remain unresolved .
for example , there is no experimental information about the energetics of the surface , or about the nature of surface relaxations . nor is there a clear picture at present of the electronic charge density at the surface , and the way this is modified by removal of oxygen .
the detailed interpretation of the observed gap states also remains controversial .
the aim of the present work is to use first - principles calculations to build up a detailed picture of the energetics , atomic structure and electronic structure of both the stoichiometric and the reduced sno@xmath0 ( 110 ) surface .
we report calculated results for the energy of the relaxed and unrelaxed surfaces , the relaxed atomic positions , the valence electron distribution and the surface electronic density of states .
the calculations have been performed using density functional theory ( dft ) and the pseudopotential technique @xcite .
these methods have a growing record of success in studying a wide range of oxides , including mgo @xcite , li@xmath0o @xcite , al@xmath0o@xmath4 @xcite , sio@xmath0 @xcite and tio@xmath0 @xcite , and have been used to investigate the structure and energetics of lattice defects and surfaces of oxide materials , and the process of molecular adsorption on their surfaces @xcite . in a previous paper @xcite , we have reported a detailed investigation of the sno@xmath0 and sno perfect crystals using the present methods .
we showed there that the calculations account satisfactorily for the equilibrium crystal structure , phonon frequencies and electronic structure of the two materials , and those investigations form the basis for the present work . when considering the structure of oxide surfaces , it is important to recall the very great differences that exist between different surfaces .
at one extreme , there is the almost complete lack of relaxation effects revealed both by experiments @xcite and by ab initio calculations @xcite for the case of the mgo ( 001 ) surface ; at the other , there are the very large atomic relaxations leading to a major reduction in surface energy found on the @xmath5-alumina basal - plane surface @xcite . in this context
, we shall see that sno@xmath0 ( 110 ) is an intermediate case , showing only moderate deviations from bulk termination .
even the gross disturbance associated with loss of surface oxygen will be shown to lead to only rather minor displacements .
a particularly important aspect of the present work concerns the spatial distribution of valence electrons at the surface . on a simple ionic picture , removal of neutral oxygen would lead to the reduction of tin from the sn@xmath1 to the sn@xmath2 state , and spectroscopic studies have given support to this @xcite .
alternatively , we might regard the removal of neutral oxygen as creating surface f - centres . on the first view
, we would expect the electrons left on the surface to be concentrated on surface sn ions .
the second view might lead us to expect electrons to be left on the site vacated by the removed oxygen .
the present first - principles calculations will give an answer to this question , and will also allow us to investigate the states occupied by these electrons .
there have been many spectroscopic measurements of the surface electronic structure of sno@xmath0 ( 110 ) .
the experimental evidence shows that on the stoichiometric surface there are no states in the gap @xcite .
however , removal of surface oxygen by heating in high vacuum leads to a broad distribution of gap states @xcite .
the surprising feature is that these gap states appear to extend up in energy from the valence band maximum ( vbm ) . at moderate temperatures of @xmath6 600 k
, there is a weak distribution of states extending to ca .
1.6 ev above the vbm . as the temperature is raised to @xmath6 1075 k , and more oxygen is lost , a continuous distribution of states spreads up to the fermi level , which lies just below the conduction band minimum ( cbm ) .
this implies that gap states extend almost across the entire gap .
this behaviour is surprising because in most materials it is appropriate to consider the f - center states as pulled down from the cbm by the coulomb attraction of conduction electrons to the anion vacancy .
indeed , the electron states associated with f - centers in _ bulk _ sno@xmath0 are known to lie only @xmath6 0.15 ev below the cbm @xcite .
an instructive comparison here is with the electronic structure of the ( 110 ) surface of tio@xmath0 , which has the same crystal structure as sno@xmath0 .
when tio@xmath0 ( 110 ) loses oxygen , gap states appear below the cbm , as expected from the simple arguments , the depth of these states below the cbm being @xmath6 0.8 ev @xcite .
an important purpose of the present work is to shed light on the seemingly anomalous behaviour of sno@xmath0 .
as we shall see , our calculations do show a broad distribution of states in the gap for the reduced surface of sno@xmath0 , and we shall be able to identify these surface states as the states occupied by electrons left by removed oxygen
. we shall also be able to confirm a conjecture by p. a. cox _
_ @xcite that the energy of the surface states is strongly influenced by the high polarizability of the sn@xmath2 ion , and we shall suggest a connection with the unusual structure of the sno crystal , which is also stabilized by this polarizability , as pointed out in our recent paper @xcite .
so far as we are aware , no previous first - principles work on sno@xmath0 surfaces has been reported , although there have been two recent papers on the tio@xmath0 surface @xcite , and we shall discuss the relation between that work and the work reported here .
there have been previous attempts to treat the surface properties of sno@xmath0 ( 110 ) using non - self - consistent tight - binding methods @xcite , which will also be discussed later .
the rest of the present paper is organized as follows .
the following section gives a brief summary of the techniques used in the work .
3 then presents our results for the equilibrium structure , energy , valence charge distribution and density of states of the stoichiometric ( 110 ) surface .
results for the reduced ( 110 ) surface are described in sec .
finally , the significance of our results and some pointers to future work are discussed in sec .
the calculations are based on density functional theory and the pseudopotential approximation , with electron correlation described by the local density approximation ( lda ) .
these well established and commonly used techniques have been extensively reviewed in the literature ( see e.g. @xcite ) . in the pseudopotential approach , it is assumed that the core orbitals have exactly the same form as in free atoms , and only valence electrons are represented explicitly in the calculations .
the valence - core interactions are described by non - local pseudopotentials , which are generated by _ ab initio _
calculations on isolated atoms .
the solid - state calculations are performed on periodically repeating cells , with occupied valence orbitals represented by a plane - wave expansion .
this expansion includes all plane waves whose kinetic energy @xmath7 ( @xmath8 the wavevector , @xmath9 the electron mass ) satisfies @xmath10 , where @xmath11 is a chosen plane - wave cut - off energy .
convergence of the calculations with respect to the size of basis set is achieved by systematic increase of the plane - wave cut - off @xmath11 .
the present calculations were performed partly using the sequential code castep ( _ c_ambridge _ s_equential _ t_otal _ e_nergy _ p_ackage ) @xcite running on the cray y - mp at the rutherford - appleton laboratory , and partly with its parallel version cetep ( _ c_ambridge-_e_dinburgh _ t_otal _ e_nergy _ p_ackage ) @xcite on the 64-node intel ipsc/860 machine at the daresbury laboratory .
the pseudopotentials we use are exactly the same as those used in our earlier work on the sno@xmath0 and sno perfect crystals @xcite . briefly
, the sn pseudopotential was generated using the standard kerker scheme @xcite , treating the 5@xmath12 and 5@xmath13 levels as valence states and the 4@xmath14 levels and all lower levels as core states . for oxygen
, we use a pseudopotential that was carefully optimized using the technique of lin _ et al .
_ @xcite in order to reduce the plane - wave cut - off as much as possible .
both pseudopotentials are treated in the kleinman - bylander ( kb ) representation @xcite , with the @xmath13 wave treated as local for both elements .
the lda exchange - correlation energy is represented by the ceperley - alder formula @xcite , and brillouin zone sampling is performed using the monkhorst - pack scheme @xcite .
the pseudopotentials , and other aspects of the current techniques , have been thoroughly tested in our earlier work on the sno@xmath0 and sno perfect crystals @xcite , where we made detailed comparisons with experiment for the equilibrium crystal structures , the zone - center phonon frequencies and the electronic densities of states .
we showed there that with the pseudopotentials we are using a plane - wave cut - off of 1000 ev is needed to achieve good convergence , and this is the cut - off used throughout the present work .
a general view of the stoichiometric ( 110 ) surface of sno@xmath0 is shown in fig .
1 . the structure we are assuming here is the one deduced from a wide range of measurements ( see e.g. @xcite ) .
the terminology we shall use when discussing the surface structure is as follows .
the outermost ions , which are oxygen ions forming ridges parallel to the @xmath15-axis , will be referred to as bridging oxygens .
below these is a plane containing both tin and oxygen ions ; the oxygens in this plane will be called in - plane oxygens .
the tin ions in this plane occupy two kinds of site , one of which lies below the ridge of bridging oxygens and will be call the bridging tin site .
the second kind of tin site is five - fold coordinated on the stoichiometric surface , and we call it the five - fold tin site . finally , we shall need to refer to the oxygens below this plane , lying vertically beneath bridging oxygens ; we call these sub - bridging oxygens . as in our previous work on oxide surfaces
@xcite , the calculations are done in slab geometry .
the use of this geometry is an artifact which we are obliged to use because our electronic structure techniques are based on periodic boundary conditions .
our calculations are performed on a collection of ions in a box having the shape of a parallelepiped , which is periodically repeated in all three directions .
the repeating box contains a single unit cell of a slab of ions , and repetition in the plane of the slab has the effect of creating an infinite slab .
repetition in the third direction creates an infinite stack of slabs , each slab being separated from its neighbours by a vacuum layer .
our real interest is , of course , in the properties of a single surface of a semi - infinite crystal . in order to be sure that these properties can be obtained by studying repeated slabs , one has to choose the width of the vacuum between slabs , and the thickness of the slabs themselves large enough so that the surfaces have only a negligible effect on each other .
slabs of stoichiometric sno@xmath0 having ( 110 ) surfaces consist of layers containing sn and o ions ( referred to as sn - o layers ) , and layers containing only o ions .
it is convenient to specify the thickness of such a slab by giving the number of sn - o layers .
the thinnest slab for which we can expect meaningful results consists of two sn - o layers .
we report results both for this system , whose slab unit cell contains 12 ions , and the four - layer system having a unit cell of 24 ions .
we have investigated the effect of vacuum width by performing a series of calculations on the two - layer system . for the purpose of these tests ,
the ions in each slab are at their unrelaxed perfect - crystal positions .
brillouin zone sampling is performed using the lowest - order monkhorst - pack scheme @xcite , so that the sampling wavevectors are ( @xmath16 , @xmath16 , @xmath16 ) in units of the primitive reciprocal lattice vectors of the repeating cell . we specify the vacuum width @xmath17 by the distance between neighbouring slabs measured along the surface normal , defined so that for zero width the stack of slabs becomes identical to the infinite perfect crystal . for these tests
, we wish to know how large the vacuum width has to be for the surface energy @xmath18 to be independent of this width .
table 1 shows our calculated values of the unrelaxed surface energy of the two - layer system as a function of @xmath17 .
the surface energy @xmath18 is obtained by subtracting the energy per cell for the case @xmath19 from the energy per cell for the @xmath17 of interest and dividing by the total surface area per cell .
judging by the dependence of @xmath18 on @xmath17 , one sees that slab interactions across the vacuum width are completely negligible for @xmath20 3.28 , and this is the value of @xmath17 we have adopted for all subsequent calculations .
the effects of slab _ thickness _ will become clear as we proceed .
it is useful to note that for thick vacuum layers the energy dispersion of electron states propagating normal to the surface must vanish .
this means that there is nothing to be gained from @xmath8-point sampling normal to the surface .
the monkhorst - pack sampling set ( @xmath16 , @xmath16 , @xmath16 ) can therefore be replaced by the set ( 0 , @xmath16 , @xmath16 ) without making any difference .
we have checked that this is indeed the case for the vacuum width of 3.28 .
we have performed calculations of the equilibrium relaxed structure for both the two - layer and four - layer systems .
relaxation has been performed using the steepest descents method , the criterion for equilibrium being that the forces on all atoms should be less than 0.4 ev @xmath21 .
this criterion is enough to ensure that further reduction would change the surface energy by less than 0.01 j m@xmath22 and the ionic positions by less than 0.01 ; these changes are negligible for practical purposes
. relaxation of the two - layer system causes the surface energy to decrease from 1.96 to 1.35 j m@xmath22 . for the four - layer system ,
the relaxed surface energy is 1.50 j m@xmath22 .
the difference between these two values is not entirely negligible , and it is clear that interactions between the surfaces are significant for the two - layer slab .
so far as we are aware , there are no experimental values for the surface energy of sno@xmath0 . however , it is interesting to note that the calculations of mulheran and harding @xcite based on an empirical interaction model for sno@xmath0 gave the value of 1.38 j m@xmath23 for the ( 110 ) surface .
table 2 shows the displacements of the surface sn and o ions away from their perfect - crystal positions for both the two - layer and four - layer slabs .
for both the five - fold and bridging sn ions and for the bridging o ion , the relaxations are along the surface normal , i.e. the ( 110 ) direction , by symmetry . for the in - plane o ion , relaxations along both the ( 110 ) and ( @xmath24 ) directions are possible .
relaxations are counted positive if they are along the outward normal , and for in - plane o , displacements along ( @xmath24 ) are positive if the ion moves away from the neighbouring bridging sn .
two conclusions emerge from this table .
firstly , the displacements of surface ions are substantial , compared with the relaxations at inert surfaces like mgo ( 001 ) .
secondly , there are significant differences between the results for the two slabs . in all cases ,
the displacements are smaller for the thicker slab , and this explains why the relaxed surface energy is slightly higher for the 4-layer slab .
the indication is that relaxations at the two surfaces are interacting significantly with each other in the 2-layer slab , and that this slab is not fully adequate for quantitative calculations .
however , the sign and general magnitude of the displacements are the same for both slabs .
more discussion of the displacements will be given in sec .
we show in fig .
2 the valence electron density on two ( 001 ) planes which both pass through sn and o sites .
the remarkable feature of these plots is the almost total absence of disturbance of the density in the region away from the surface .
we shall refer to these plots again when we discuss the non - stoichiometric surface .
we have calculated the electronic density of states ( dos ) of the relaxed two - layer and four - layer systems .
as with our calculations on the dos of the perfect crystal @xcite , we have used the tetrahedron method to perform the integration over the brillouin zone .
( in fact , since @xmath8-point sampling is unnecessary along the surface normal , we sample only over the _ surface _
brillouin zone , so that the tetrahedron method reduces to the triangle method . )
it should be stressed that for systems of the size treated here the calculations are very demanding , and the brillouin zone sampling we have been able to achieve is rather restricted .
we have needed to limit the sampling to only nine points in the irreducible wedge of the surface brillouin zone , and the energy resolution that this gives is limited .
the calculated dos for the two sizes of slab are compared with the bulk dos obtained in our earlier work in fig .
as expected , there is a close similarity between the bulk and slab dos .
however , there are three important differences .
firstly , the slabs exhibit new structure at the top of the valence band , in the form of a pair of sharp peaks .
secondly , there is a slight reduction in the width of the gap .
the calculated gap for the bulk is 2.2 ev ( experimental value is 3.6 ev @xcite ) , while for the thin and thick slab systems we find gaps of 1.6 and 1.5 ev respectively .
the new peaks at the vbm appear to be mainly responsible for the narrowing of the gap .
the third difference is that in the slab system states split off from the top of the o(2@xmath12 ) band .
we have studied the states at the top of the valence band in some detail .
examination of the eigenvalues shows that there are two states per repeating cell associated with each of the sharp peaks at the vbm . to analyze the nature of these states ,
we have calculated the electronic density @xmath25 of each of these states separately .
this density is defined as : @xmath26 where the sum goes over sampling vectors in the irreducible wedge , @xmath27 is a sampling weight , and @xmath28 is the normalized wave function of state @xmath29 at sampling vector @xmath30 .
we find that all four states are quite strongly localized on bridging oxygens , with some weight also on sub - bridging oxygens ; the weight on in - plane atoms and elsewhere in the slab is small .
one of the states associated with each dos peak can be regarded as being localized at one of the surfaces of the slab . to illustrate the localization of the states , we show in figure 4 a plot of the sum of the densities @xmath25 associated with the two states responsible for the first peak at the vbm .
this shows the rather rapid decrease of density as one passes into the slab along the surface normal through the bridging oxgyen .
detailed examination of density contour plots ( not shown here ) indicates that the states associated with the two peaks have different symmetries .
the states of lower energy have mainly @xmath13-like character , with the axis along the surface normal .
the higher states are also @xmath13-like , but with the axis along the @xmath31 direction , in other words perpendicular to the rows of bridging oxygens .
although most of the reduction of the band gap is due to the localized states at the vbm , there is some indication of a small contribution to the reduction from states at the cbm .
this indication comes from examination of @xmath25 for the two lowest states of the conduction band , which show some localization in the surface region .
the peak in the dos split off from the o(2@xmath12 ) band contains two states per cell .
plots of @xmath25 show that these states are strongly localized on the bridging oxygens , with some weight on the sub - bridging oxygens .
the form of @xmath25 confirms the expected @xmath12-like character of these states .
all the experiments indicate that reduction of the ( 110 ) surface occurs by removal of bridging oxygens . however , it is also clear that the fraction of bridging oxygens removed depends on the conditions .
we have investigated two cases : first , the case where all bridging oxygens are removed ( referred to as fully reduced ) , and second , the case where only half of them are removed ( half reduced ) .
we have studied the fully reduced surface using both the two - layer and four - layer slabs . to study the half - reduced case , we have treated the situation where every other bridging oxygen is removed , so that we have an ordered array of surface vacancies .
for this case , the surface unit cell has to be doubled along the @xmath15-axis , and at present we can only handle this with the two - layer slab .
views of the two reduced surfaces are shown in fig .
it is important to stress that we create the reduced surface by removing oxygen from one surface of the slab only ; the other surface remains stoichiometric .
the ionic relaxations for the fully relaxed surface are reported in table 3 .
it is extremely remarkable that the displacements of the ions near the surface are almost identical to those we have found at the stoichiometric surface ( see table 2 ) .
this strongly suggests that the electrons left at the surface by the removal of oxygen are exerting roughly the same forces as the original bridging oxygens . in table 4
, we show the ionic relaxations for the half - reduced surface . in this case , because of the lower symmetry , the ionic displacements can have components in directions for which they were zero by symmetry at the fully reduced surface .
comparison of these results with those for the 2-layer stoichiometric slab ( table 2 ) shows that the displacements are almost identical in the two cases .
this indicates once again that removal of bridging oxygen leads to surprisingly small adjustments of the ionic positions .
the valence charge density of the relaxed fully - reduced four - layer slab is shown in fig .
2 on the same two ( 001 ) planes for which we showed the charge density of the stoichiometric surface . comparing the two pairs of plots , we see a very large change of density distribution in the immediate neighborhood of the removed bridging oxygen and the bridging tin directly below it .
the extreme localization of this disturbance is very striking .
not only is the density completely unaltered on atomic layers immediately below the surface , but even around the in - plane oxygen and five - fold tin sites there is no visible change .
it appears that the two electrons per bridging oxygen left on the surface are localized in channels running along the ( 001 ) direction just above the bridging tin sites and passing through the bridging oxygen positions .
this will be confirmed in greater detail when we discuss the electronic structure below . the valence distribution at the half - reduced surface is also shown in fig .
the main differences with the plots for the fully reduced surface are that the electron density is now lower near the bridging sn site and greater near the empty bridging oxygen site .
this is what we should expect , since the surface electrons are now more strongly confined between the remaining bridging oxygens .
the calculated densities of states for the two - layer and four - layer fully reduced systems and for the two - layer half - reduced system are shown in fig .
the major change from the stoichiometric system is the appearance of broad distributions of states in the band gap . for the fully reduced case ,
the states span the whole gap , but for the half - reduced system there is a band of gap states roughly in the middle of the gap .
we also note the two sharp peaks at the vbm and the peak split off from the o(2@xmath12 ) band that we found for the stoichiometric surface .
since bridging oxygens have been removed only from one side of the slab , we clearly expect these states to remain unchanged on the other surface .
this implies that there should be one state in each of these peaks , and we have verified that this is the case .
we find that there is a single state responsible for the part of the dos spanning the gap .
the variation of the energy of this state with @xmath30-vector is responsible for the spread in energy .
density plots of @xmath25 for this state show that it is strongly localized in the surface region near the bridging tin and oxygen sites .
contour plots of @xmath32 on ( 001 ) planes passing through this site and through the bridging tin site are shown in fig .
the plots show that there are substantial concentrations of density in the region vertically above the bridging tin site , on in - plane oxygens and on the sub - bridging oxygens .
integration of the density in these features shows that roughly 30 % is in the region above the tin site , 10 % on each of the in - plane oxygens , 10 % on sub - bridging oxygen , and the remainder in small features elsewhere .
in assessing our results , it is important to bear in mind the strengths and weaknesses of density functional theory , which forms the basis of all our calculations .
dft is a theory of the ground state , and we expect it to give an accurate description of the energetics of the system in the ground state , which in the present case includes surface energies and the equilibrium positions of the atoms in the relaxed surface . the theory is also designed to give accurate results for the valence electron distribution , and for related quantities such as the distribution of electrostatic potential . on the other hand , the principles of dft give us no right to expect accurate results for excited - state properties .
it is well known , for example , that dft generally underestimates band gaps . as we reported in our previous paper @xcite ,
our calculated band gap for bulk sno@xmath0 is 2.2 ev , which is @xmath6 40 % smaller than the experimental value of 3.6 ev .
one therefore needs to be more cautious about the interpretation of excited - state results .
nevertheless , it has been found that dft does usually give a reliable semiquantitative guide to electronic densities of states , and we shall certainly draw important conclusions from our results for such quantities .
we have shown that atomic relaxations at both the stoichiometric and reduced ( 110 ) surfaces are rather moderate , being on the order of 0.1 - 0.3 .
this is large compared with the very tiny displacements on the mgo ( 001 ) surface @xcite , but small compared with the giant displacements at the @xmath5-al@xmath0o@xmath4 basal - plane surface @xcite .
we find that five - fold coordinated sn moves into the surface , while bridging sn and in - plane o move out .
the separation between bridging o and bridging sn decreases markedly .
these qualitative features are the same as found by ramamoorthy _
@xcite in their dft - pseudopotential calculations on the tio@xmath0 ( 110 ) surface , and the magnitudes of their displacements are also very similar to ours .
the only significant difference is that the absolute displacement of bridging o in tio@xmath0 is inwards , whereas we find it be outwards for sno@xmath0 .
independent dft calculations on tio@xmath33 ( 110 ) have also been performed by vogtenhuber _
_ @xcite using the flapw technique .
these authors find qualitatively similar results .
relaxation has a significant effect on the surface energy , reducing it from 1.96 to 1.50 j m@xmath22 for the stoichiometric surface .
this relaxed surface energy is quite close to the value of 1.38 j m@xmath22 calculated by mulheran and harding @xcite using an empirical interaction model .
this provides support for the realism of the ionic model they use @xcite .
unfortunately , mulheran and harding do not report results for the surface relaxations they obtain with the model .
it is also worth noting that our calculated surface energy is rather similar to the value of 0.9 j m@xmath22 found by ramamoorthy _
et al . _ for the relaxed tio@xmath0 ( 110 ) surface . on both the stoichiometric and reduced surfaces , we find marked changes of electronic structure compared with the bulk .
the rather well localized surface states on bridging oxygen that we have found on the stoichiometric surface are not unexpected .
these states , both above the o(2@xmath12 ) band and at the vbm are localized largely on bridging oxygen , which is in an exposed position where one would expect a reduced madelung potential and hence a reduction in the binding of states .
so far as we are aware , there is at present no experimental evidence for the existence of surface o(2@xmath12 ) states .
detailed ups measurement have been performed on the stoichiometric sno@xmath0 ( 110 ) surface , and no evidence has been found for surface states at the vbm .
because of this rather surprising disagreement , we have tested the robustness of this aspect of our calculations by repeating them with an independently generated pseudopotential , but this makes no difference .
it is conceivable that the problem might arise from our treatment of the sn(4@xmath14 ) states as core states , or even from deficiencies of the lda , but it would require a substantial effort to test either of these possibilities . at the reduced surface , we have shown that the two electrons left by the removal of neutral oxygen occupy surface states localized in the region of the empty bridging oxygen sites and the adjacent bridging tin sites . more accurately ,
their density is concentrated immediately above the bridging tin atoms , with some weight on in - plane oxygens and elsewhere .
this finding provides direct support for the common assumption that oxygen loss leads to reduction of tin from sn@xmath1 to sn@xmath2 .
however , our results make it clear that if one does wish to use the sn@xmath2 description , then it is crucial to recognize the extreme distortion of this ion .
the change of electron density caused by reduction occurs entirely on the vacuum side of the bridging sn ions , which , if considered as sn@xmath2 ions , must be regarded as maximally polarized .
a high degree of polarization of these ions is expected , because they are very polarizable , and oxygen removal makes their invironment very unsymmetrical .
their high polarizability arises from the rather small energy difference between 5@xmath12 and 5@xmath13 states and the large spatial overlap of these states .
as emphasized in our earlier paper @xcite , the sn@xmath2 ion in the sno perfect crystal is also in a very unsymmetrical environment , and we showed that there also the sn ion exhibits maximal polarization .
it would be interesting to know the valence charge distribution at the reduced tio@xmath0 surface , but this appears not to have been examined in the first - principles calculations of ramamoorthy et al . @xcite .
the lowest - lying conduction - band states in that case have ti(3@xmath14 ) character , and the weak spatial overlap between 3@xmath14 and 4@xmath13 states will make the reduced ti ion much less polarizable
. the surface charge distribution of tio@xmath0 should therefore be rather different from that of reduced sno@xmath0 . on the fully reduced surface
, we find a broad band of surface states spanning the entire gap , and we have demonstrated that occupation of these states is responsible for the surface electron density above bridging sn sites . on the half - reduced surface , the band of gap states contracts to have a width of @xmath6 1.5 ev , and lies @xmath6 1.1 ev above the vbm . in comparing with experiment , it must be remembered that the real surface is likely to be significantly disordered .
experimentally @xcite , a weak distribution of states near the bottom of the gap is observed for small degrees of reduction .
as further oxygen is removed , the distribution spreads upwards to the fermi level , which is just below the top of the gap .
it seems clear that these must be essentially the same as the gap states we find .
the large width of the band of gap states is not surprising , since the electrons in these states are able to propagate almost freely along the @xmath15-axis .
the width of @xmath6 1.5 ev we find for the half - reduced surface is also reasonable .
it is interesting to note that recent calculations on periodic arrays of f - centers in bulk mgo @xcite find a band width of gap states of 1.6 ev when the spacing between f - centers is 5.9 , and the f - center states have been found to have a substantial amplitude on neighboring oxygens .
the energy of the gap state is also understandable , following an argument originally due to p. a. cox _
we have seen that the sn@xmath2 ions are highly polarized .
this must arise from a strong mixing of the 5@xmath12 and 5@xmath13 states caused by surface fields , and this mixing will displace one of the hybrid states downwards in energy . alternatively , we can adopt the f - center viewpoint and note that f - center levels are commonly found rather deep in the gap .
for example , in bulk mgo the f - center levels are roughly 5 ev below the cbm . in very recent dft calculations
, we have found that surface f - centers in mgo have their levels at almost the same energy as as those in the bulk @xcite .
this still leaves an unresolved question , however , concerning the relation between bulk and surface f - centers in sno@xmath0 .
it is rather well established that the donor states involved in @xmath34-type sno@xmath0 are oxygen vacancies , and that the binding energy of electrons to these vacancies is no more than @xmath6 0.15 ev @xcite .
this implies that bulk and surface f - centers in sno@xmath0 must have very different properties , and the reasons for this need to be explored .
we hope to return to this question in the future .
the insights we have gained allow us to comment on the predictions of tight - binding calculations .
the recent tight - binding work of godin and lafemina @xcite made a thorough investigation of the stoichiometric and fully reduced sno@xmath0 ( 110 ) surfaces .
no states were found in the gap for either surface .
this was consistent with previous tight - binding work of munnix and schmeits @xcite , who studied oxygen vacancies on the sno@xmath0 ( 110 ) surface and found no gap states .
these results are in conflict with what we have found .
it is also noteworthy that tight - binding calculations on the tio@xmath0 ( 110 ) surface @xcite failed to find any gap states associated with oxygen vacancies , even though the existence of such states has been well established for many years @xcite . in our view , the tight - binding models used in the above work lack an essential piece of physics .
we have argued from our results that if one adopts an lcao viewpoint then sn@xmath2 states are crucially influenced by cation @xmath12-@xmath13 mixing caused by surface fields .
but such a mechanism is completely lacking from the tight - binding models used in the above work .
there is no term in the hamiltonians which accounts for this mechanism of @xmath12-@xmath13 mixing , and no inclusion of any kind of effect arising from surface fields .
very recently , self - consistent tight - binding calculations on tio@xmath0 ( 110 ) and other oxide surfaces have been reported , in which coulombic effects are approximately included @xcite .
although only the stoichiometric surface has been studied so far by these methods , it is already clear from these papers that self - consistency has a major effect on the surface electronic states .
finally , we note that much of the practical interest in sno@xmath0 surfaces centers on the adsorption and reactions of molecules .
we are currently using the results of the present work as a basis for investigating the dissociation of h@xmath0o at the sno@xmath0 ( 110 ) surface @xcite , and we hope to report on this soon .
i.m . thanks the serc for a research studentship and aea technology for partial financial support .
the calculations were performed partly on the intel ipsc/860 machine at daresbury laboratory and partly on the cray y - mp at the rutherford - appleton laboratory under serc grant gr / j59098 .
subsidiary analysis employed local computer equipment provided by serc grants gr / h31783 and gr / j36266 . useful discussions with a. h. harker , g. thornton and p. a. cox are gratefully acknowledged .
we are also indebted to j. holender , who has been an unfailing source of help during this work .
99 v. t. agekyan , _ phys .
a _ , * 43 * ( 1975 ) 11 j. f. mcaleer , p. t. moseley , j. o. w. norris and d. e. williams , _
faraday trans .
i _ , * 83 * ( 1987 ) 1323 ; _ ibid .
_ , * 84 * ( 1988 ) 441 i. brown and w. r. patterson ,
_ j. chem .
faraday trans .
i _ , * 79 * ( 1983 ) 1431 e. de frsart , j. darville and j. m. gilles , _ solid st
_ , * 37 * ( 1980 ) 13 p. a. cox , r. g. egdell , c. harding , w. r. patterson and p. j. tavener , _ surf .
_ , * 123 * ( 1982 ) 179 r. g. egdell , s. eriksen and w. r. flavell , _ solid st .
_ , * 60 * ( 1986 ) 835 r. g. egdell , s. eriksen and w. r. flavell , _ surf .
_ , * 192 * ( 1987 ) 265 j. w. erickson and s. semancik , _ surf .
_ , * 187 * ( 1987 )
l658 d. f. cox , t. b. fryberger and s. semancik , _ phys .
b _ , * 38 * ( 1988 ) 2072 d. f. cox , t. b. fryberger and s. semancik , _ surf .
_ , * 224 * ( 1989 ) 121 d. f. cox and t. b. fryberger , _ surf .
_ , * 227 * ( 1990 ) l105 j. m. themlin , r. sporken , j. darville , r. caudano , j. m. gilles and r. l. johnson , _ phys . rev .
b _ , * 42 * ( 1990 ) 11914 g. l. shen , r. casanova , g. thornton and i. colera , _ j. phys . : condens .
, * 3 * ( 1991 ) s291 r. g. egdell , in _ the science of ceramic interfaces ii _ , ed
. j. nowotny ( elsevier , new york , 1994 ) , p. 527
s. munnix and m. schmeits , _ phys
b _ , * 27 * ( 1983 ) 7624 ; _ ibid . _ , * 33 * ( 1986 ) 4136 t. j. godin and j. p. lafemina , _ phys . rev .
b _ , * 47 * ( 1993 ) 6518 p. a. mulheran and j. h. harding , _ mod . sim . mat .
_ , * 1 * ( 1992 ) 39 j. a. marley and r. c. dockerty , _ phys .
_ , * 140 * ( 1965 ) a304 s. samson and c. g. fonstad , _ j. appl .
_ , * 44 * ( 1973 ) 4618 g. p.
srivastava and d. weaire , _ adv .
_ , * 36 * ( 1987 ) 463 m. j. gillan , in _ proc .
nato asi on computer simulation in materials science , aussois , march 1991 _ , ed . m. meyer and v. pontikis , p. 257
( dordrecht , kluwer , 1991 ) m. c. payne , m. p. teter , d. c. allan , t. a. arias and j. d. joannopoulos , _ rev . mod .
_ , * 64 * ( 1992 ) 1045 a. de vita , m. j. gillan , j .- s .
lin , m. c. payne , i. tich and l. j. clarke , _ phys .
_ , * 68 * ( 1992 ) 3319 a. de vita , m. j. gillan , j .- s .
lin , m. c. payne , i. tich and l. j. clarke , _ phys .
b _ , * 46 * ( 1992 ) 12964 i. manassidis , a. de vita , j .- s .
lin and m. j. gillan , _ europhys .
_ , * 19 * ( 1992 ) 605 i. manassidis , a. de vita and m. j. gillan , _ surf .
_ , * 285 * ( 1993 ) l517 i. manassidis and m. j. gillan , _
_ , * 77 * ( 1994 ) 335 n. binggeli , j. r. chelikowsky and r. m. wentzcovitch , _ phys . rev .
b _ , * 49 * ( 1994 ) 9336 m. ramamoorthy , r. d. king - smith and d. vanderbilt , _ phys . rev .
b _ , * 49 * ( 1994 ) 7709 m. ramamoorthy , d. vanderbilt and r. d. king - smith , _ phys . rev .
b _ , * 49 * ( 1994 ) 16721 s. pugh and m. j. gillan , _ surf .
_ , * 320 * , 331 ( 1994 ) i. manassidis and m. j. gillan , _ phys
b _ , submitted m. r. welton - cook and w. berndt , _ j. phys
* 15 * ( 1982 ) 5691 d. l. blanchard , d. l. lessor , j. p. lafemina , d. r. baer , w. k. ford and t. guo , _ j. vac .
_ , * a9 * ( 1991 ) 1814 t. urano , t. kanaji and m. kaboragi , _ surf .
_ , * 134 * ( 1983 ) 109 m. caus , r. dovesi , c. pisani and c. roetti , _ surf .
_ , * 175 * ( 1986 ) 551 v. e. henrich , g. dresselhaus and h. j. zeiger , _ phys .
_ , * 36 * ( 1976 ) 1335 k. e. smith , j. l. mackay and v. e. henrich , _ phys .
b _ , * 35 * ( 1987 ) 5822 r. l. kurtz , r. stockbauer , t. e. madey , e. romn and j. l. de segovia , _ surf .
_ , * 218 * ( 1989 ) 178 d. vogtenhuber , r. podloucky , a. neckel , s. g. steinmann and a. j. freeman , _ phys . rev .
b _ , * 49 * ( 1994 ) 2099 l. j. clarke , i. tich and m. c. payne , _ comput . phys .
_ , * 72 * ( 1992 ) 14 g. p. kerker , _ j. phys .
c _ , * 13 * ( 1980 ) l189 j .- s .
lin , a. qteish , m. c. payne and v. heine , _ phys .
b _ , * 47 * ( 1993 ) 4174 l. kleinman and d. m. bylander , _ phys . rev .
_ , * 48 * ( 1980 ) 566 d. m. ceperley and b. j. alder , _ phys .
_ , * 45 * ( 1980 ) 566 j. perdew and a. zunger , _ phys . rev .
b _ , * 23 * ( 1981 ) 5048 h. j. monkhorst and j. d. pack , _ phys . rev .
b _ , * 13 * ( 1976 ) 5188 a. m. rappe , j. d. joannopoulos and p. a. bash , _
_ , * 114 * ( 1992 ) 6466 c. m. freeman and c. r. a. catlow , _ j. solid st .
_ , * 85 * ( 1990 ) 65 q. s. wang and n. a. w. holzwarth , _ phys .
b _ , * 41 * ( 1990 ) 3211 l. n. kantorovich and m. j. gillan , surf . sci .
, submitted s. munnix and m. schmeits , _ phys .
b _ , * 31 * ( 1985 ) 3369 ; _ j. vac .
a _ , * 5 * ( 1987 ) 910 j. goniakowski and c. noguera , _ surf .
_ , * 323 * ( 1995 ) 129 n. yu and j. w. halley , university of minnesota supercomputer institute research report umsi 94/129 ( 1994 ) j. goniakowski and m. j. gillan , unpublished @xmath35 @xmath36 @xmath37 @xmath38
1 : perspective view of the stoichiometric ( upper panel ) , fully reduced ( middle ) and half - reduced ( lower ) surfaces of sno@xmath0 .
small dark spheres and large light spheres represent sn and o respectively .
2 : contour plots of valence electron density ( units : 10@xmath22 electron / @xmath39 ) on ( 001 ) planes passing through the slabs used for studying the stoichiometric ( panels a and b ) , fully reduced ( c and d ) and half - reduced ( e and f ) ( 110 ) surfaces . for each surface , two planes are shown , passing through the in - plane oxygen and bridging tin sites ( a , c , e ) and the bridging oxygen and fivefold tin sites ( b , d , f ) .
regular sn sites are marked by the symbol @xmath40 .
fig 3 : electronic dos for the perfect crystal ( a ) , the two - layer stoichiometric ( b ) and four - layer stoichiometric ( c ) slabs , the two - layer fully reduced ( d ) and four - layer fully reduced ( e ) slabs , and the two - layer half - reduced slab ( f ) . in all cases , the dashed line marks the energy of the highest occupied state .
4 : plot of electronic density associated with the two states in the first peak at the top of the valence band in the four - layer stoichiometric slab .
the line on which the density is plotted passes through the bridging and sub - bridging oxygen sites , which are marked by arrows .
fig . 5 : contour plots of electron density ( units : electrons / @xmath39 ) associated with the single state ( per removed oxygen ) spanning the band - gap for the four - layer fully reduced slab .
plots are on ( 001 ) planes passing through the bridging tin and in - plane oxygen sites ( panel a ) and bridging and sub - bridging oxygen sites ( b )
. intersections of the guide lines mark the significant sites : in panel ( a ) the intersection below the centre is the bridging tin , and the intersections on either side are in - plane oxygens ; in panel ( b ) , the intersection at the centre is the bridging oxygen site and the lower intersection is sub - bridging oxygen .
distances are in . | first - principles calculations based on density functional theory ( dft ) and the pseudopotential method have been used to study the stoichiometric and reduced sno@xmath0 ( 110 ) surface .
the ionic relaxations are found to be moderate for both the stoichiometric and reduced surfaces , and are very similar to those found in recent dft - pseudopotential work on tio@xmath0 .
removal of neutral oxygen leaves two electrons per oxygen on the surface , which are distributed in channels passing through bridging oxygen sites .
the associated electron density can be attributed to reduction of tin from sn@xmath1 to sn@xmath2 , but only if the charge distribution on sn@xmath2 is recognized to be highly asymmetric .
reduction of the surface gives rise to a broad distribution of gap states , in qualitative agreement with spectroscopic measurements . |
medical scanner is the most used application of tomographic reconstruction .
it allows to explore the interior of a human body . in the same way
, industrial tomography explores the interior of an object and is often used for non - destructive testing .
we are interested here in a very specific application of tomographic reconstruction for a physical experiment described later .
the goal of this experiment is to study the behavior of a material under a shock .
we obtain during the deformation of the object an x - ray radiography by high speed image capture .
we suppose this object is radially symmetric , so that one radiograph is enough to reconstruct the 3d object .
several authors have proposed techniques ( standard in medical tomography ) for tomographic reconstruction when enough projections ( from different points of view ) are available : this allows to get an analytic formula for the solution ( see for instance @xcite or @xcite ) .
these methods can not be used directly when only a few number of projections is known .
some alternative methods have been proposed in order to partially reconstruct the densities ( see for instance @xcite ) .
we are interesting here in single view tomographic reconstruction for radially symmetric object ( see for instance @xcite for a more complete presentation of the subject ) . as any tomographic reconstruction
, this problem leads to an ill - posed inverse problem .
as we only have one radiograph , data are not very redundant and the ill - posed character is even more accurate .
we present here a tomographic method adapted to this specific problem , originally developed in @xcite , and based on a curve evolution approach .
the main idea is to add some a priori knowledge on the object we are studying in order to improve the reconstruction .
the object may then be described by a small set of characters ( in this case , they will be curves ) which are estimated by the minimization of an energy functional .
this work is very close to another work by feng and al @xcite .
the main difference is the purpose of the work : whereas they are seeking recovering textures , we are looking for accurate edges .
it is also close to the results of bruandet and al @xcite .
however , the present work handles very noisy data and highly unstable inverse problems , and shows how this method is powerful despite these perturbations .
further , we take here into account the effects of blur ( which may be non - linear ) and try to deconvolve the image during the reconstruction .
let us mention at this point that our framework if completely different from the usual tomographic point of view , and usual techniques ( such as filtered back - projection ) are not adapted to our case .
indeed , usually , as the x - rays are supposed to be parallel ( this is also the case here ) , the `` horizontal '' slices of the object are supposed to be independent and studied separately .
usual regularization techniques deal with one slice and regularize this particular slice . here , because of the radial symmetry , the slices are composed of concentric annulus and do not need any regularization .
the goal of this work is to add some consistency between the slices in order to improve the reconstruction .
the paper is organized as follows .
first we present the physical experiment whose data are extracted and explain what are the motivations of the work .
next , we introduce the projection operator . in section 4 ,
we present a continuous model with the suitable functional framework and prove existence result .
section 5 is devoted to formal computation of the energy derivative in order to state some optimality conditions . in section 6 , a front propagation point of view
is adopted and the level set method leads to a non local hamilton - jacobi equation . in the last section , we present some numerical results and give hints for numerical schemes improvement .
this work is part of some physical experiments whose goal is the study of the behavior of shocked material .
the present experiment consists in making a hull of well known material implode using surrounding explosives .
the whole initial physical setup ( the hull , the explosives ... ) are radially symmetric .
a reasonable assumption is to suppose that during the implosion , everything remains radially symmetric .
physicists are looking for the shape of the interior at some fixed time of interest . at that time
, the interior may be composed of several holes which also may be very irregular .
figure [ fig : object ] is a synthetic object that contains all the standard difficulties that may appear .
these difficulties are characterized by : * several disconnected holes . * a small hole located on the symmetry axis ( which is the area where the details are difficult to recover ) .
* smaller and smaller details on the boundary of the top hole in order to determine a lower bound detection . to achieve this goal ,
a x - rays radiograph is obtained . in order to extract the desired informations
, a tomographic reconstruction must be performed .
let us note here that , as the object is radially symmetric , a single radiography is enough to compute the reconstruction .
a radiography measures the attenuation of x - rays through the object . a point on the radiography
will be determined by its coordinates @xmath0 in a cartesian coordinates system where the @xmath1-axis will be the projection of the symmetry axis .
if @xmath2 is the intensity of the incident x - rays flux , the measured flux @xmath3 at a point @xmath0 is given by @xmath4 where the integral operates along the ray that reaches the point @xmath0 of the detector , @xmath5 is the infinitesimal element of length along the ray and @xmath6 is the local attenuation coefficient . for simplicity
, we will consider that this coefficient is proportional to the material density . to deal with linear operators ,
we take the neperian logarithm of this attenuation and will call the transformation @xmath7 the projection operator . through the rest of the paper , in order to simplify the expression of the projection operator , we will suppose that the x - ray source is far enough away from the object so that we may consider that the rays are parallel , and orthogonal to the symmetry axis . as a consequence , the horizontal slices of the object may be considered separately to perform the projection .
as the studied object is radially symmetric , we will work in a system of cylinder coordinates @xmath8 where the @xmath9-axis is the symmetry axis .
the object is then described by the density at the point @xmath8 , which is given by a function @xmath10 which depends only on @xmath11 by symmetry . in the text
, the notation @xmath10 will always refer to the density of the object .
a typical function @xmath10 is given in figure [ fig : whole - object ] .
it represents an object composed of concentric shells of homogeneous materials ( called the `` exterior '' in what follows ) surrounding a ball ( called the `` interior '' ) of another homogeneous material that contains some empty holes .
this figure may be viewed as a slice of the object by a plane that contains the symmetry axis . to recover the 3d - object
, it suffices to perform a rotation of this image around the @xmath9 axis .
for instance , the two round white holes in the center are in fact the slice of a torus . as the looked - after characteristic of the object is the shape of the holes
, we will focus only on the interior of the object ( see figure [ fig : object ] ) .
we here handle only binary objects composed of one homogeneous material ( in black ) and some holes ( in white ) .
-axis ) . , width=188 ] .
the homogeneous material is in black whereas the holes are in white.,width=188 ]
we first explicit the projection operator and its adjoint . in the case of a radially symmetric object ,
the projection operator , denoted by @xmath12 , is given , for every function @xmath13 with compact support , by @xmath14 _ proof - _ consider a 3d - object which is described by a function @xmath15 ( cartesian coordinates system ) .
the projection operator @xmath12 is @xmath16 in the case of radially symmetric object , we parametrize the object by a function @xmath17 with cylinder coordinates .
therefore @xmath18 then , we have , for @xmath19 @xmath20 we perform the following change of variable @xmath21 to get @xmath22 for @xmath23 , we have @xmath24 using the change of variable @xmath25 and the fact that @xmath26 is even , by symmetry , we get @xmath27 @xmath28 operator @xmath12 may be defined by density on measurable functions @xmath10 such that all the partial applications @xmath29 belong to @xmath30 .
then , all the functions @xmath31 belong to the space @xmath32 of bounded mean oscillations functions : @xmath33 where @xmath34 . for more details , one can refer to @xcite . in the sequel , we will need to handle functions @xmath10 that are defined on @xmath35 ( instead of on @xmath36 ) .
we thus define the operator @xmath12 for function @xmath37 with compact support by @xmath38 although this has no more physical meaning . here , the function @xmath39 is defined by @xmath40 we shall also need the back - projection that is the adjoint operator @xmath41 of @xmath12 ; it can be computed in a similar way .
[ propadj]the adjoint operator ( in @xmath42 ) @xmath41 of the projection operator is given , for every function @xmath43 with compact support by : @xmath44 _ proof - _ the adjoint operator @xmath41 of @xmath12 is the unique operator such that , for every @xmath10 and @xmath45 in @xmath46 with compact support , @xmath47 using ( [ h ] ) and fubini s theorem , we get @xmath48 so we obtain the following expression for the back projection : @xmath49 @xmath28 thanks to the symmetry , this operator characterizes the radon transform of the object and so is invertible ; one radiograph is enough to reconstruct the object .
the inverse operator is given , for an almost everywhere differentiable function @xmath45 with compact support , and for every @xmath50 , by @xmath51 because of the derivative term , the operator @xmath52 is not continuous . consequently , a small variation on the measure @xmath45 leads to significant errors on the reconstruction .
as our radiographs are strongly perturbed , applying @xmath52 to our data leads to a poor reconstruction . due to the experimental setup
they are also two main perturbations : * a blur , due to the detector response and the x - ray source spot size . * a noise .
others perturbations such as scattered field , motion blur ... also exist but are neglected in this study .
+ we denote by @xmath53 the effect of blurs .
we will consider the following simplified case where @xmath53 is supposed to be linear @xmath54 where @xmath55 is the usual convolution operation , @xmath56 is the projected image and @xmath57 is a positive symmetric kernel .
a more realistic case stands when the convolution operates on the intensity : then @xmath53 is of the form @xmath58 where @xmath59 is the multiplicative coefficient between the density and the attenuation coefficient .
some specific experiments have been carried out to measure the blur effect .
consequently , we will suppose that , in both cases , the kernel @xmath57 is known .
the linear blur is not realistic but is treated here to make the computations simpler for the presentation .
the noise is supposed for simplicity to be an additive gaussian white noise of mean 0 , denoted by @xmath60 .
consequently , the projection of the object @xmath10 will be @xmath61 the comparison between the theoretical projection @xmath62 and the perturbed one is shown on figure [ fig : real_proj ] . the reconstruction using the inverse operator @xmath52 applied to @xmath45
is given by figure [ fig : inverse ] .
the purpose of the experiment is to separate the material from the empty holes and consequently to precisely determine the frontier between the two areas , which is difficult to perform on the reconstruction of figure [ fig : inverse ] . of the object of figure [ fig : object ] .
right - hand side : real projection of the same object with realistic noise and blur.,title="fig:",width=188 ] of the object of figure [ fig : object ] .
right - hand side : real projection of the same object with realistic noise and blur.,title="fig:",width=188 ] applied to the real projection on the right - hand side.,title="fig:",width=188 ] applied to the real projection on the right - hand side.,title="fig:",width=188 ] it is clear from figure [ fig : inverse ] that the use of the inverse operator is not suitable . in order to improve the reconstruction
, we must add some a priori knowledge on the object to be reconstructed .
indeed the object that we reconstruct must satisfy some physical property .
we chose to stress on two points : * the center of the object is composed of one homogeneous known material s density with some holes inside .
* there can not be any material inside a hole . in a previous work @xcite , j.m .
lagrange reconstructed the exterior of the object .
in this reconstruction , the density of the material at the center of the object is known , only the holes are not reconstructed .
in other words , we can reconstruct an object without holes and we can compute ( as @xmath12 and @xmath53 are known ) the theoretical projection of this reconstruction .
we then act as if the blurred projection was linear and subtract the projection of the non - holed object to the data . in
what follows , we will call experimental data this subtracted image which corresponds to the blurred projection of a `` fictive '' object of density 0 with some holes of known `` density '' @xmath63 .
consequently , the space of admitted objects will be the set of functions @xmath10 that take values in @xmath64 .
this space of functions will be denoted @xmath65 in the sequel .
the second hypothesis is more difficult to take into account .
we chose in this work to tackle the problem via an energy minimization method where the energy functional is composed of two terms : the first one is a matching term , the second one is a penalization term which tries to handle the second assumption . the matching term will be a @xmath42-norm which is justified by the gaussian white noise . in the case where @xmath53 is not linear , the exact method to remove
the exterior is to operate the blur function on the addition of the known exterior and the center . for the sake of simplicity
, we will not use this method and will consider that the errors are negligible when subtracting the projections .
let us first describe more precisely the set @xmath65 .
this time , the functions @xmath66 will be defined on @xmath35 , with values still in @xmath67 , with compact support .
therefore , such a function @xmath10 will be characterized by the knowledge of the curves that limit the two areas where @xmath10 is equal to @xmath68 and to @xmath69 . indeed , as the support of the function @xmath10 is bounded , these curves are disjoint jordan curves and the density of the inside is @xmath68 whereas the density of the outside is 0 .
consequently , the energy that we will consider will be a function of @xmath70 where @xmath70 is a set of disjoint jordan curves . for mathematical reasons
, we must add an extra - assumption : the curves @xmath70 are @xmath71 so that the normal vector of the curves is well - defined ( as an orthogonal vector to the tangent one ) . in this continuous framework ,
the matching term is just the usual @xmath42-norm between @xmath72 and the data @xmath45 ( where @xmath12 is given by ( [ h ] ) ) .
so , the first term is @xmath73 for the penalization term , we choose @xmath74 where @xmath75 denotes the length of the curves @xmath70 .
let us remark that this penalization term may be also viewed as the total variation ( up to a multiplicative constant ) of the function @xmath10 because of the binarity . eventually , the total energy functional is @xmath76 which is an adaptation of the well - known mumford - shah energy functional introduced in @xcite .
the `` optimal '' value of @xmath77 may depend on the data .
the previous analysis gives the mains ideas for the modelization .
now , we make it precise using an appropriate functional framework .
let @xmath78 be a bounded open subset @xmath35 with lipschitz boundary .
we shall consider bounded variation functions . recall that the space of such functions is @xmath79 where @xmath80 where @xmath81 denotes the space of @xmath71 functions with compact support in @xmath78 .
the space @xmath82 endowed with the norm @xmath83 is a banach space .
+ if @xmath84 its derivative in @xmath85 ( distributions ) is a bounded radon measure denoted @xmath86 and @xmath87 is the total variation of @xmath88 on @xmath78 .
let us recall useful properties of bv - functions ( @xcite ) : let @xmath78 be an open subset @xmath35 with lipschitz boundary .
+ 1 . if @xmath89 , we get the following decomposition for @xmath86 : @xmath90 where @xmath91 is the absolutely continuous part of @xmath86 with respect of the lebesgue measure and @xmath92 is the singular part .
the map @xmath93 from @xmath82 to @xmath94 is lower semi - continuous ( lsc ) for the @xmath95 topology .
@xmath96 with compact embedding .
@xmath97 with compact embedding .
we precise hereafter an important continuity property of the projection operator @xmath12 .
[ h1 ] the projection operator @xmath12 is continuous from @xmath98 to @xmath99 for every @xmath100 $ ] and @xmath101 .
_ proof - _ it is a direct consequence of hlder inequality .
let be @xmath101 and @xmath102 .
its support is included in @xmath103 which is included in some @xmath104 \times [ -m,+m ] $ ] where @xmath105 and only depends on @xmath106 .
it is clear that @xmath72 is defined everywhere on @xmath106 and , for every @xmath19 @xmath107 & \le 2\displaystyle { \left [ \int_{u}^{m}|f(r , v)|^{2+s } \ , dr \right]^{\frac{1}{2+s } } \ , \left[\int_{u}^{m } \frac{r^q}{(r + u)^{\frac{q}{2 } } } \ , \frac{1 } { ( r -u)^{\frac{q}{2 } } } \ , dr\right ] ^{\frac{1}{q } } } \end{array}\ ] ] where @xmath108 . therefore @xmath109
\le 2 \ , m^{\frac{q}{2 } } \|f\|_{l^{2+s } } \,\left[\int_{u}^{m } ( r -u)^{-\frac{q}{2 } } dr \right]\ ] ] the computations in the case @xmath110 are similare and lead to the same inequality ( with some additional absolute values ) .
as @xmath111 we get @xmath112^\frac{s}{2(1+s)}\le c ( { \omega},s ) \|f\|_{l^{2+s}}~;\ ] ] here and in the sequel @xmath113 denotes a generic constant depending on @xmath114 and @xmath106 .
so @xmath115 as @xmath106 is bounded , this yields @xmath116 \quad \|hf \|_{l^p ( { \omega } ) } \le c ( { \omega},s ) \|f\|_{l^{2+s}}~.\ ] ] @xmath28 as we consider the length of curves , the most suitable functional space to set a variational formulation of the reconstruction problem is @xmath82 .
therefore , we consider the following minimization problem @xmath117 stands for the @xmath118- norm , @xmath119 and @xmath120 * the operator @xmath53 is given by ( [ blur ] ) . without loss of generality , we may assume ( for simplicity ) that @xmath121 * at last , `` @xmath122 '' , is the binarity constraint .
we have mentionned that the image @xmath123 takes its values in @xmath124 where @xmath125 . with the change of variable @xmath126
, we may assume that the image values belong to @xmath127 .
a similar problem has been studied in @xcite with smoother projection operator and * convex * constraints .
this is not our case .
the pointwise constraint `` @xmath128 '' is a very hard constraint .
the constraint set is not convex and its interior is empty for most usual topologies .
now we may give the main result of this section : problem @xmath129 admits at least a solution .
_ proof - _ let @xmath130 be a minimizing sequence .
it satisfies @xmath131 ; so @xmath132 therefore the sequence @xmath133 is @xmath134- bounded .
as @xmath135 is bounded as well , the sequence is bounded in @xmath82 .
thus it converges ( extracting a subsequence ) to some @xmath136 for the weak - star topology .
+ estimate ( [ estim1 ] ) implies the weak convergence of @xmath133 to @xmath137 in @xmath138 for every @xmath101 .
thanks to the @xmath12 continuity property of proposition [ h1 ] , we assert that @xmath139 weakly converges to @xmath140 in @xmath118 .
we get @xmath141 with the lower semi - continuity of the norm .
+ moreover @xmath82 is compactly embedded in @xmath134 .
this yields that @xmath133 strongly converges to @xmath137 in @xmath134 .
as @xmath142 is lsc with respect to @xmath134- topology , we get @xmath143 finally @xmath144 as the pointwise constraint is obviously satisfied , @xmath137 is a solution to @xmath129 . @xmath28
now we look for optimality conditions .
unfortunately we can not compute easily the derivative of the energy in the @xmath82 framework .
indeed we need regular curves and we do not know if the @xmath82 minimizer provides a curve with the required regularity . moreover , the set of constraints is not convex and it is not easy to compute the gteaux- derivative ( no admissible test functions ) .
so we have few hope to get classical optimality conditions and we rather compute minimizing sequences .
we focus on particular ones that are given via the gradient descent method inspired by @xcite .
formally , we look for a family of curves @xmath145 such that @xmath146 so that @xmath147 decreases as @xmath148 .
let us compute the energy variation when we operate a small deformation on the curves @xmath149 .
in other word , we will compute the gteau derivative of the energy for a small deformation @xmath150 : @xmath151 we will first focus on local deformations @xmath152 . let @xmath153 be a point @xmath154 of @xmath149 .
we consider a local reference system which center is @xmath154 and axis are given by the tangent and normal vectors at @xmath154 and we denote by @xmath155 the new generic coordinates in this reference system . with an abuse of notation , we still denote @xmath156 .
we apply the implicit functions theorem to parametrize our curve : there exist a neighborhood @xmath157 of @xmath154 and a @xmath71 function @xmath158 such that , for every @xmath159 , @xmath160 eventually , we get a neighborhood @xmath157 of @xmath154 , a neighborhood @xmath3 of @xmath161 and a @xmath71 function @xmath158 such that @xmath162 the local parametrization is oriented along the outward normal @xmath163 to the curve @xmath149 at point @xmath154 ( see figure [ fig : deformation ] ) .
more precisely , we define the local coordinate system @xmath164 where @xmath165 is the usual tangent vector , @xmath163 is the direct orthonormal vector ; we set the curve orientation so that @xmath163 is the outward normal .
the function @xmath10 if then defined on @xmath157 by @xmath166 this parametrization is described on figure [ fig : deformation ] .
we then consider a local ( limited to @xmath157 ) deformation @xmath152 along the normal vector .
this is equivalent to handling a @xmath71 function @xmath167 whose support is included in @xmath3 . the new curve @xmath168 obtained after the deformation @xmath169 is then parametrized by @xmath170 this defines a new function @xmath171 : @xmath172 we will also set @xmath173 . this deformation is described on figure [ fig : deformation ] . .
@xmath154 is the current point , @xmath157 is the neighborhood of @xmath154 in which the deformation is restricted to and @xmath174 is the new curve after deformation .
the interior of the curve is the set where @xmath175 the gteau derivative for the energy @xmath176 has already been computed in @xcite and is @xmath177 where @xmath178 denotes the curvature of the curve and @xmath179 is the parametrization of @xmath150 .
+ it remains to compute the derivative for the matching term .
first we estimate @xmath180 : a simple computation shows that @xmath181 and @xmath182 now we compute @xmath183 where @xmath149 ( resp .
@xmath168 ) is the curve associated to the function @xmath10 ( resp .
@xmath171 ) : @xmath184 to simplify the notations , we denote by @xmath185 so that we need to compute @xmath186 as @xmath180 is zero out of the neighbourhood @xmath157 , we have @xmath187 in the case @xmath188 , we have , @xmath189 as the function @xmath190 is continuous ( and thus bounded on @xmath157 ) , we may pass to the limit by dominated convergence and get @xmath191 in the case @xmath192 , we have @xmath193 and we obtain the same limit as in the nonnegative case . finally , the energy derivative is @xmath194 if we set @xmath195 , we get @xmath196 as @xmath197 formula ( [ derivebis ] ) may be written @xmath198 where @xmath163 denotes the outward pointing normal unit vector of the curve @xmath149 , @xmath199 denotes the usual scalar product in @xmath35 and @xmath200 is a positive coefficient that depends on the curvilinear abscissa @xmath114 .
the latter expression is linear and continuous in @xmath150 , this formula is also true for a non - local deformation ( which can be achieved by summing local deformations ) .
the goal of the present section is to consider a family of curves @xmath201 that will converge toward a local minimum of the functional energy . from equation ( [ eq : derivee ] ) , it is clear that if the curves @xmath202 evolve according to the differential equation @xmath203 the total energy will decrease . to implement a numerical scheme that discretizes equation ( [ eq : edp ] )
, it is easier to use a level set method ( see @xcite for a complete exposition of the level set method ) . indeed ,
equation ( [ eq : edp ] ) may present some instabilities , in particular when two curves collide during the evolution or when a curve must disappear .
all these evolutions are handled easily via the level set method .
the level set method consists in viewing the curves @xmath149 as the 0-level set of a smooth real function @xmath204 defined on @xmath35 .
the function @xmath10 that we are seeking is then just given by the formula @xmath205 we must then write an evolution pde for the functions @xmath206 that corresponds to the curves @xmath168 .
let @xmath207 be a point of the curve @xmath168 and let us follow that point during the evolution .
we know that this point evolves according to equation [ eq : edp ] @xmath208 we can re - write this equation in terms of the function @xmath204 recognizing that @xmath209 where @xmath210 stands for the gradient of @xmath204 with respect to @xmath211 , @xmath212 denotes the euclidean norm .
the evolution equation becomes @xmath213 then , as the point @xmath207 remains on the curve @xmath168 , it satisfies @xmath214 . by differentiating this expression
, we obtain @xmath215 which leads to the following evolution equation for @xmath204 : @xmath216 that is @xmath217 the above equation is an hamilton - jacobi equation which involves a non local term ( through @xmath12 and @xmath53 ) .
such equations are difficult to handle especially when it is not monotone ( which is the case here ) .
in particular , existence and/or uniqueness of solutions ( even in the viscosity sense ) are not clear .
the approximation process is not easy as well and the numerical realization remains a challenge though this equation is a scalar equation which is easier to discretize than the vectorial one .
here we used the discrete scheme described in @xcite to get numerical results .
we briefly present the numerical scheme .
we used an explicit scheme in time and the spatial discretization has been performed following @xcite .
we set @xmath218 so that the equation ( [ eq : levelset ] ) is @xmath219 we set @xmath220 with @xmath221 et @xmath222 .
the explicit euler scheme gives : @xmath223 the curvature term is computed as @xmath224 where @xmath225 stands for the partial derivative with respect to @xmath211 .
the discrete approximation of the gradient is standard : @xmath226 @xmath227 are @xmath228 defined in the same way .
a usual approximation for @xmath229 is given by : @xmath230^{1/2}~.\ ] ] the non local term @xmath231 is exactly computed .
the previous scheme has been implemented on a 3.6 ghz pc .
a classical reinitialization process has been used each 500 iterations .
the test image size was 256 @xmath232 256 pixels .
the other parameters of the computation were set to @xmath233 and the blur kernel is a gaussian kernel of standard deviation 5 pixels .
the computed image is quite satisfying ( see figure 7 . )
however , we note a bad reconstruction along the symmetry axis due to the problem geometry and a lack of information
. moreover we have to improve the algorithme behavior .
indeed , we observe numerical instability ( in spite of the regularization process ) that leads to a very small time step choice .
therefore the computational time is quite long ( about 2.5 hours ) .
in addition , classical stopping criteria are not useful here : the expected solution corresponds to a `` flat '' level of function @xmath234 and the difference between two consecutive iterates means no sense .
an estimate of the cost function decrease is not appropriate as well ( we observe oscillations ) .
we decided to stop after a large enough number of iterations ( here 20 000 ) . in spite of all these disadvantages ,
this method is satisfactory considering the low signal to noise ratio of the radiograph .
these good results can be explained by the strong assumptions that we add ( in particular the binary hypothesis ) which are verified by our synthetic object .
anyway , the method has beeen successfully tested on `` real '' images as well , that is imgaes of objets with the same kind of properties ( `` almost '' binary ) but we can not report them here ( confidential data ) . a semi - implicit version of the algorithm is actually tested to improve stability . | this paper deals with a method of tomographic reconstruction of radially symmetric objects from a single radiograph , in order to study the behavior of shocked material . the usual tomographic reconstruction algorithms such as generalized inverse or filtered back - projection can not be applied here because data are very noisy and the inverse problem associated to single view tomographic reconstruction is highly unstable . in order to improve the reconstruction , we propose here to add some a priori assumptions on the looked after object .
one of these assumptions is that the object is binary and consequently , the object may be described by the curves that separate the two materials .
we present a model that lives in bv space and leads to a non local hamilton - jacobi equation , via a level set strategy .
numerical experiments are performed ( using level sets methods ) on synthetic objects . |
the coalescence of gases by turbulence and gravity intricately controls star formation within giant molecular clouds @xcite . on one hand , turbulence has the ability to hinder star formation by providing kinetic energy that can oppose gravity . on the other ,
the supersonic turbulence ubiquitously observed in the molecular phase of the interstellar medium ( ism ) produces local shocks and compressions , which lead to enhanced gas densities that are key for triggering star formation @xcite .
understanding the complex effects of turbulence in the ism is therefore crucial to understanding the process of galaxy evolution .
the cold turbulent gas that provides the fuel for star formation is only visible in the millimetre / sub - millimetre to radio wavelengths , and is often faint , making it difficult to detect at high spatial resolutions .
a standard method to measure the mean column density of molecular gas ( @xmath1 ) is to use rotational lines of co. a severe problem with this method is that , because co is about @xmath8 times less abundant than the main mass carrier , @xmath0 , one requires a co - to-@xmath0 conversion factor , which is typically calibrated based on measurements in our own galaxy .
however , the co - to-@xmath0 conversion factor may depend on metallicity , environment and redshift , introducing high uncertainties in the reconstruction of the total gas surface densities from measurements of co @xcite .
another method is to measure dust emission or dust extinction and assuming a gas - to - dust ratio to infer the molecular gas masses and surface densities .
these methods can suffer from uncertainties in the gas - to - dust ratio , especially for low - metallicity galaxies where this ratio becomes increasingly uncertain .
both co and dust observations require telescopes and instruments that work at millimetre / sub - millimetre wavelengths , which may not always be available and/or may have relatively low spatial resolution . here
we present a new method to estimate @xmath1 based on the star formation rate ( sfr ) , which can be obtained with optical spectroscopy .
large optical integral field spectroscopy ( ifs ) surveys have started to provide us with details regarding the chemical distribution and kinematics of extragalactic sources at a size and uniformity unprecedented until recent times .
large galaxy surveys such as the sloan digital sky survey ( sdss ; * ? ? ?
* ; * ? ? ?
* ) , 2-degreefield galaxy redshift survey ( 2dfgrs ; * ? ? ?
* ) , the cosmic evolution survey ( cosmos ; * ? ? ?
* ) , the vimos vlt deep survey ( vvds ; * ? ? ?
* ) , and the galaxy and mass assembly survey ( gama ; * ? ? ?
* ; * ? ? ?
* ) have contributed more than 3.5 million spectra that have been of extraordinary aid to our understanding of galaxy evolution .
however , those spectra have been taken with a single fibre or slit , and provide only a single , global spectrum per galaxy @xcite .
these spectra are therefore susceptible to aperture effects because differing parts or fractions of the galaxies are recorded for each source , thus making each observation dependent on the size and distance of the galaxy , as well as the positioning of the fibre @xcite .
conversely , ifs can spatially resolve each galaxy observed , thus assigning individual spectra at many locations across the galaxy .
here we utilise data from the sami galaxy survey , an ifs survey with the aim to observe 3400 galaxies over a broad range of environments and stellar masses .
we use the sfrs measured in sami in order to provide a tool for estimating @xmath1 .
the basis of our @xmath1 reconstruction method is a recent star formation relation developed in the multi - freefall framework of turbulent gas @xcite .
there have been many ongoing efforts to find an intrinsic relation between the amount of gas and the rate at which stars form in a molecular cloud .
initiated by @xcite , hereafter k98 , @xmath3 correlates with @xmath1 @xcite , which can be approximated by an empirical power law with exponent @xmath9 , @xmath10 for a sample of low - redshift disc and starburst galaxies , k98 found an exponent of @xmath11 .
however , significant scatter and discrepancies between different sets of data exist within this framework , commonly referred to as the kennicutt - schmidt relation .
these discrepancies suggest that @xmath3 does not only depend on @xmath1 , but also on factors such as the turbulence and the freefall time of the dense gas on small scales . motivated by the fact that dense gas forms stars at a higher rate , a new star formation correlator was derived in @xcite , hereafter sfk15 .
this descriptor , denoted by @xmath12 and called the ` maximum or multi - freefall gas consumption rate ' ( mgcr ) , is dependent on the probability density function ( pdf ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) of molecular gas , @xmath13 where @xmath4 is the mach number of the turbulence , @xmath14 is the turbulence driving parameter @xcite and @xmath15 is the ratio of thermal to magnetic pressure @xcite in the molecular gas . the sfk15 model for @xmath3 given by eq .
( [ eq : newsflaw ] ) is built upon foundational concepts laid out by @xcite , hereafter kdm12 , which had parameterised @xmath3 by the ratio between @xmath1 and the average ( single ) freefall time @xmath16 , a correlator hereon denoted by @xmath17 @xcite .
our new correlator instead uses the concept of a multi - freefall time , which was pioneered by @xcite , tested with numerical simulations in @xcite , and used in sfk15 as a stepping stone to expand upon the kdm12 model .
sfk15 found that @xmath3 is equal to @xmath18 of the mgcr by placing observations of milky way clouds and the small magellanic cloud ( smc ) in the k98 , kdm12 and sfk15 frameworks , confirming the measured low efficiency of star formation @xcite .
statistical tests in sfk15 showed that a significantly better correlation between @xmath3 and @xmath12 was achieved than that which could be attained between either the @xmath1 or @xmath17 parameterisations of the previous star formation relations by k98 and kdm12 , respectively .
the scatter in the sfk15 relation was found to be a factor of 34 lower than in the k98 and kdm12 relations , suggesting that it provides a better physical model for @xmath3 compared to the empirical relation by k98 and compared to the single - freefall relation by kdm12 .
the aim of the current work is to formulate a method to predict the distribution of @xmath1 by inverting eq .
( [ eq : newsflaw ] ) and using optical observations , which will be plentiful in the coming few years . here
we use the h@xmath2 luminosities and velocity dispersions provided by the sami galaxy survey to estimate @xmath1 from measurements of @xmath3 and @xmath4 . in sec .
[ sec : obs ] we describe the observations of our sami galaxy sample .
[ sec : theory ] introduces our new method to derive @xmath1 by inverting the sfk15 relation . in sec .
[ sec : results ] we present our results and compare purely star - forming with composite / agn / shock galaxies in our sample . in sec .
[ sec : comp ] we compare our own and other observations and predictions to previous star formation relations within the kennicutt - schmidt framework . in sec .
[ sec : k98vssfk15 ] we demonstrate that our new method for predicting @xmath1 is superior to inverting the k98 relation .
our conclusions are summarised in sec .
[ sec : conc ] .
the new data products for each sami galaxy in our sample derived here ( average turbulent mach number , cold gas density , freefall time , etc . , and finally @xmath1 ) are listed in tab .
[ tab : galaxies ] in appendix [ app : online ] and are available for download in the online version of the journal or by contacting the authors .
we selected a sample of 260 galaxies from the sami galaxy survey internal data release version 0.9 .
the sydney - aao multi - object integral field spectrograph ( sami ; * ? ? ?
* ) is a front - end fibre feed system for the aaomega spectrograph @xcite , consisting of 13 bundles of 61 fibres each ( hexabundles ; * ? ? ?
* ; * ? ? ?
* ) that can be deployed over a 1 degree diameter field of view .
sami therefore enables simultaneous spatially - resolved spectroscopy of twelve galaxies and one calibration star with a 15 diameter field - of - view on each object .
the aaomega spectrograph can be configured to provide different resolutions and wavelength ranges ; the sami galaxy survey employs the 570v grating to obtain a resolution of @xmath19 ( @xmath20 ) at and the 1000r grating to obtain @xmath21 ( @xmath22 ) at .
sami datacubes are reduced and re - gridded to a spatial scale of @xmath23 @xcite and the spatial resolution is about @xmath24 ( green et al .
, in prep ) .
the sami galaxy survey plans to include more than 3000 galaxies at redshift @xmath25 covering a wide range of stellar masses and environments .
the sample is drawn from gama @xcite with additional entries from eight nearby clusters to cover denser environments ( * ? ? ?
* owers et al .
, in prep )
. reduced datacubes and a variety of emission line based higher - level data products are included in the first public data release ( * ? ? ?
* green et al . , in prep ) .
the emission lines of sami galaxies have been analysed using the spectral fitting pipeline lzifu @xcite to extract emission line fluxes and kinematics for each spectrum .
the spectrum associated with each spectral / spatial pixel ( ` spaxel ' ) is first fit with a stellar template using the ` penalized pixel - fitting ' ( ppxf ) routine @xcite before fitting up to three gaussian line profiles to each of eleven strong emission lines simultaneously . for this paper , we choose to use the single gaussian fits , and make use of the emission line flux maps , gas velocity maps , and gas velocity dispersion maps below . also available in the sami galaxy survey database
are maps of sfr and @xmath3 ( in units of @xmath26 ) .
these maps are made using extinction - corrected h@xmath2 fluxes converted to sfrs following the relation derived in @xcite .
the sfr maps are fully described in medling et al .
( in prep ) . from the pool of sami galaxies , we select a sub - sample of galaxies according to the criteria described below .
we only consider spaxels with a sufficiently high signal - to - noise ( s / n ) ratio .
the s / n was defined to be the ratio of the total emission line flux to the statistical one - sigma error in the line flux .
this error was inferred using the levenberg - marquardt technique of chi - squared minimisation @xcite . in the following , we list the selection criteria : 1 .
source extractor ( sextractor ) ellipticity values are available .
these values were obtained from the gama database @xcite .
we require the ellipticity for each galaxy to estimate the physical volume of gas within each spaxel ( explained in detail in sec .
[ sec : tff ] below ) .
2 . the s / n ratio must be @xmath27 in the h@xmath2 , h@xmath15 , [ n`ii ` ] , [ s`ii ` ] , [ o`i ` ] and [ o`iii ` ] emission lines .
this allows reliable classification of the emission mechanism .
however , in order to measure velocity dispersions down to about @xmath28 , we require and impose an s / n ratio of @xmath29 in the measured velocity dispersion ( explained in detail in sec .
[ subsec : veldisp ] below ) .
we also require that beam - smearing ( see sec .
[ subsec : veldisp ] ) did not have a significant effect on the measured velocity dispersion .
3 . after removing spaxels that have low s / n and/or
are affected by beam - smearing , the galaxy must have more than ten star - forming spaxels remaining .
the star - forming spaxels were filtered using the optical classification criteria given in @xcite , an example of which is shown in fig .
[ fig : bpt ] .
this classification scheme uses optical emission line ratios ( bpt / vo diagrams ; * ? ?
* ; * ? ? ?
* ) , in order to distinguish between star - forming galaxies and galaxies that are dominated by an active galactic nucleus ( agn ) or by shocks .
the h@xmath2-to - sfr conversion factor used in this work is only valid for star - forming regions , because agn / shock - dominated spaxels are contaminated with emission from agn / shock regions @xcite .
emission line fluxes of each spaxel were corrected for extinction using the balmer decrement and the @xcite reddening curve .
standard extinction for the diffuse ism was assumed , with an @xmath30 value of 3.1 being utilised throughout the analysis @xcite .
each galaxy was classified as either a ` star - forming ' or ` composite / agn / shock ' galaxy .
to be classified as star - forming , the galaxy had to have at least @xmath31 of all valid spaxels lying below and to the left - hand side of the @xcite classification line in the [ o`iii`]/h@xmath15 versus [ n`ii`]/h@xmath2 diagram , and below and to the left - hand side of the @xcite line in the [ s`ii`]/h@xmath2 and [ o`i`]/h@xmath2 diagrams , as described in @xcite ( see fig .
[ fig : bpt ] ) .
a galaxy was classified as composite / agn / shock , if at least @xmath32 of all valid spaxels lie above the @xcite classification line on the [ o`iii`]/h@xmath15 versus [ n`ii`]/h@xmath2 diagram and above the @xcite classification line on the [ s`ii`]/h@xmath2 and [ o`i`]/h@xmath2 diagnostic diagrams .
thus , composite / agn / shock galaxies may include composite , agn , or shock @xcite galaxies according to the classification in @xcite .
these classifications resulted in a sample of 219 star - forming and 41 composite / agn / shock classified galaxies .
here we exploit the spatially resolved sami h@xmath2 flux and @xmath3 maps in combination with the h@xmath2 velocity dispersion maps to derive predictions of @xmath1 across each galaxy in our sample .
examples of @xmath3 maps are shown in the left - hand panels of fig .
[ fig : maps ] for the star - forming and composite / agn / shock classified galaxies from fig .
[ fig : bpt ] .
we further derive spaxel - averaged values of the physical parameters for each galaxy in our subsample .
the sfk15 model relies on the availability of the sonic mach number @xmath33 , the ratio between the gas velocity dispersion and the local speed of sound .
this was prompted by the findings of @xcite , showing that the observed scatter within the k98 and kdm12 relations may be primarily attributed to the physical variations in @xmath33 . as direct measurements of @xmath33
are unavailable for the sami sample , for every pixel we estimate a value using the method described in the following sections .
the result of these procedures is shown in the middle panels of fig .
[ fig : maps ] , for the two example galaxies , gama 492414 ( top ) and gama 69740 ( bottom ) .
molecular clouds in galactic spiral arms exhibit a gas temperature range of , while those in the galactic centre can have temperatures up to @xmath34 @xcite .
all @xmath0 gas should lie within this temperature range , otherwise it will cease to be molecular under typical conditions in the ism @xcite .
the local sound speed ( @xmath35 ) of the gas is given by @xmath36^{1/2},\ ] ] with the boltzmann constant @xmath37 , the mass of a hydrogen atom @xmath38 , and the mean particle weight @xmath39 .
the latter is @xmath40 for molecular gas and @xmath41 for ionised gas , assuming standard cosmic abundances @xcite
. therefore temperatures of @xmath42 and @xmath43 correspond to molecular sound speeds of @xmath44 and @xmath45 , respectively .
we hence estimate the mach number of the gas in each spaxel by dividing the velocity dispersion by the molecular sound speed of @xmath46 , which is appropriate for the dense , cold star - forming phase of the ism in the temperature range . in order to apply our star formation relation , eq .
( [ eq : newsflaw ] ) , we need an estimate of the turbulent velocity dispersion of the molecular gas in order to construct the turbulent mach number . here
we use the h@xmath2 velocity dispersion to approximate the velocity dispersion of the cold gas .
the h@xmath2 velocity dispersion is similar ( to within a factor of ) to the molecular gas velocity dispersion , because of the coupling of turbulent gas flows between the hot , warm and cold phases of the ism .
for instance , it has been found that for m33 , the second - most luminous spiral galaxy in our local group , the atomic h`i ` dispersions are a fair estimator of the co dispersions @xcite . in m33 , h@xmath2 velocities have been found to trace h`i ` velocities reasonably well @xcite .
however , the h@xmath2 velocity dispersion is expected to be somewhat higher than the h@xmath47 velocity dispersion , because the ionised emission comes from h`ii ` regions close to massive stars , which directly contribute to driving turbulence .
we thus expect the velocity dispersion in the direct vicinity of massive stars to be overestimated . in order to take this effect into account
, we crudely approximate the h@xmath47 velocity dispersion with half the h@xmath2 velocity dispersion , @xmath48 . while this provides only a rough estimate of the molecular velocity dispersion ( trustworthy only to within a factor of 23 ) , we show below that the uncertainties that this introduces into our @xmath1 estimate are only of the order of @xmath49 .
this is because of the relatively weak dependence of @xmath1 on @xmath4 , as we will derive in sec .
[ sec : uncertainties ] below . to demonstrate this , we investigate a case below , where we assume that the molecular gas velocity dispersion is equal to the h@xmath2 velocity dispersion , @xmath50 , which yields @xmath51 lower @xmath1 .
thus , even though our velocity dispersion estimate is uncertain by factors of , the final uncertainty in @xmath1 is @xmath52 . [
[ sn - requirements ] ] s / n requirements : + + + + + + + + + + + + + + + + + the sami / aaomega spectrograph setup has an instrumental velocity resolution of @xmath53 at the wavelength of h@xmath2 ( see sec .
[ sec : sami ] ) .
velocity dispersions below this resolution limit can still be reliably measured if the s / n in the observed ( instrument - convolved ) velocity dispersion is sufficiently high . in the following , we estimate the required s / n in order to reconstruct intrinsic velocity dispersions down to @xmath54 .
we choose this cutoff of @xmath28 , because it is the sound speed of the ionised gas , eq .
( [ eq : cs ] ) with @xmath55 and @xmath41 , and thus represents a physical lower limit for @xmath56 . the intrinsic ( true ) velocity dispersion ( @xmath57 )
can be obtained by subtracting the instrumental velocity resolution ( @xmath58 ) from the observed ( instrument - convolved ) velocity dispersion ( @xmath59 ) in quadrature , with @xmath60 the same relation holds for the uncertainties ( noise ) in the velocity dispersion , @xmath61 assuming that the instrumental velocity resolution is fixed , we can use @xmath62 and simplify the last equation to @xmath63 dividing both sides by @xmath57 and substituting eq .
( [ eq : s_n_cuts ] ) yields @xmath64 since @xmath65 and @xmath66 are the observed ( instrument - convolved ) and intrinsic s / n ratios , respectively , we can estimate the required @xmath67 for the target intrinsic @xmath68 and the target intrinsic velocity dispersion that we want to resolve , @xmath54 , by evaluating @xmath69 = 34.\end{aligned}\ ] ] thus , for spaxels with observed ( instrument - convolved ) velocity dispersion s / n ratios greater or equal to 34 , we can reliably reconstruct the intrinsic ( instrument - corrected ) velocity dispersion down to @xmath28 , with an intrinsic s / n ratio of at least 5 .
we note that the sami database provides the instrument - subtracted velocity dispersion @xmath70 ( `` vdisp ` ' ) and its error @xmath71 ( `` vdisp_err ` ' ) based on the lzifu fits @xcite .
thus , in order to apply the s / n cut of 34 derived in eq .
( [ eq : sncutobs ] ) , we first reconstruct @xmath72 and its error @xmath73 , using error propagation . this criterion is functionally equivalent to setting a s / n cut on the instrument - subtracted velocity dispersion , @xmath74 after applying our s / n cuts of 34 to the observed ( instrument - convolved ) velocity dispersion , any spaxels with velocity dispersions less than @xmath28 are disregarded .
we note that this final cut only removes 1% of the spaxels with @xmath75 .
[ [ beam - smearing ] ] beam smearing : + + + + + + + + + + + + + + we also have to account for ` beam smearing ' , a phenomenon that occurs because of the limitation in spatial resolution of the instrument .
beam smearing occurs for a physical velocity field that changes on spatial scales smaller than the spatial resolution of the observation .
if there is a steep velocity gradient across neighbouring pixels , such as near the centre of a galaxy , beam smearing leads to an artificial increase in the measured velocity dispersion at such spatial locations . to account for beam smearing ,
we follow the method in @xcite and estimate the local velocity gradient @xmath76 for a given spaxel with coordinate indices @xmath77 as the magnitude of the vector sum of the difference in the velocities in the adjacent pixels , @xmath78 ^ 2 + [ v(i , j+1)-v(i , j-1)]^2}. \end{aligned}\ ] ] note that the differencing to compute @xmath76 occurs over a linear scale of three sami pixels along @xmath79 and @xmath80 and thus covers roughly the spatial resolution of the seeing - limited sami observations with @xmath81 ( see sec . [
sec : obs ] ) .
if a pixel has a neighbour that is undefined ( e.g. , because of low s / n ) , the gradient in that direction is not taken into account .
as our standard criterion to account for beam smearing , we cut any pixels in which the velocity dispersion is less than twice that of the velocity gradient ( @xmath82 ) and disregard such pixels in further analyses , leaving only spaxels that are largely unaffected by beam smearing .
in addition to our fiducial beam - smearing criterion ( @xmath82 ) , we test a case with a relaxed beam - smearing cut of @xmath83 , and find nearly identical results ( see tab .
[ tab : sum ] below ) .
we note that our standard beam - smearing cut with @xmath82 tends to remove spaxels near the centre of some of the galaxies ( see e.g. , fig .
[ fig : maps ] ) .
however , using the relaxed beam - smearing cut with @xmath83 yields global ( galaxy - averaged ) mach numbers and global @xmath1 estimates that agree to within 4% with our standard beam - smearing cut ( see tab .
[ tab : sum ] ) , demonstrating that our results are largely unaffected by beam smearing .
[ [ turbulent - velocity - dispersion - versus - systematic - motions ] ] turbulent velocity dispersion versus systematic motions : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + beam - smearing is the result of un - resolved velocity gradients in the plane - of - the - sky .
however , systematic velocity gradients ( such as resulting from rotation or large - scale shear ) along the line - of - sight ( los ) also increase the velocity dispersion ( even for arbitrarily high spatial resolution ) by los - blending .
these large - scale systematic motions do not represent turbulent gas flows ( see e.g. , the recent study of turbulent motions in the galactic - centre cloud brick , which is subject to large - scale shear * ? ? ? * ) . as we have not subtracted or accounted for these factors , our values of the turbulent velocity dispersion may be overestimated . in summary , we emphasise that the turbulent velocity dispersion has large uncertainties and is only accurate to within a factor of @xmath84@xmath85 .
however , the uncertainties that this introduces into our final product ( @xmath1 ) are @xmath52 , because of the relatively weak dependence of @xmath1 on @xmath4 ( derived in detail in sec .
[ sec : uncertainties ] below ) . to find the mgcr @xmath12 ,
we divide @xmath3 ( left - hand panels of fig .
[ fig : maps ] ) by the sfr efficiency of 0.45@xmath86 found in sfk15 .
that is , we invert eq .
( [ eq : newsflaw ] ) , @xmath87 = \frac{{\sigma_\mathrm{sfr}}}{0.0045 } .\end{aligned}\ ] ] in order to find the ratio between the gas column density and the freefall time at the average gas density , @xmath17 , we take the mach number calculated in sec . [
sec : mach ] and convert @xmath12 to @xmath17 , @xmath88 = \frac{{({\sigma_\mathrm{gas}}/t)_\mathrm{multi - ff}}}{\left(1 + b^2 \mathcal{m}^2 \frac{\beta}{\beta+1}\right)^{3/8}}.\end{aligned}\ ] ] in the following , we will assume a fixed turbulence driving parameter @xmath89 , representing a natural mixture @xcite , and assume an absence of magnetic fields such that @xmath90 .
although both of these are strong assumptions , we emphasise that , in the absence of constraints on @xmath14 or @xmath15 in these galaxies , we have to assume fixed , typical values for them and allow that these assumptions contribute to the uncertainties of the @xmath1 estimation .
however , if these parameters will be measured in the future , they can be used in eqs .
( [ eq : newsflaw ] ) and ( [ eq : gastff ] ) to obtain a more accurate prediction of @xmath1 . for simplicity ,
here we fix @xmath14 and @xmath15 , and only consider the remaining dependence on @xmath4 . now that we have @xmath92 from eq .
( [ eq : gastff ] ) , we need an estimate of the average freefall time @xmath93 to obtain @xmath1 from @xmath17 .
thus , we need an estimate of the local gas density @xmath91 , which requires some geometrical considerations and assumptions similar to the ones outlined in kdm12 . ) . in this diagram
the green rectangle represents a section of a galaxy with scale height @xmath94 , which is inclined an angle @xmath95 away from the vertical .
@xmath96 is the column length used to estimate the local gas density and freefall time . ]
first , we make the assumption that the galaxy has a uniform gas disc geometry with a scale height of @xcite .
however , depending on the viewing angle with respect to the orientation of the galactic disc in the plane of the sky , the los length through the gas may be greater than the scale height ( see fig . [
fig : geometry ] ) .
this angle can be estimated from the observed ellipticity of the galaxy . to correct for the viewing angle ,
we obtain sextractor ellipticity values , @xmath97 , for each galaxy from the gama database , from which we obtain the inclination angle , @xmath95 , of the galaxy : @xmath98 the column length , @xmath96 , can then be inferred by dividing the scale height , @xmath94 , by the cosine of the inclination angle , as pictured in fig .
[ fig : geometry ] , @xmath99 = \frac{h}{\cos(\theta ) } = \frac{100}{\cos(\theta ) } = \frac{100}{1 - \varepsilon}. \label{eq : l}\end{aligned}\ ] ] we caution that the assumed cylindrical geometry is a drastic simplification , as there has been much evidence to suggest that the scale height of a galaxy follows a relation dependent upon its radius from the galactic centre @xcite .
this may cause our predicted gas maps to underestimate the gas density towards the centre and overestimate the gas density towards the outskirts of the galaxy .
however , the general shape of the predicted distribution of gas and especially the galaxy - averaged gas surface density should not be affected significantly by this geometrical simplification .
a refinement in the geometry is relatively straightforward to implement , if one requires more accurate maps .
we estimate that the relative uncertainties in @xmath96 may be up to @xmath100 .
however , our final result ( @xmath1 ) , does not depend significantly on @xmath96 ( see detailed discussion in sec . [
sec : uncertainties ] ) . given the column length @xmath96 ,
we can write the gas density @xmath101 since we do not have @xmath1 because it is our final product , we now substitute a rearrangement of the definition of @xmath17 , @xmath102 as well as the definition of the freefall time in terms of @xmath91 , @xmath103 where @xmath104 is the gravitational constant .
we combine the three previous equations and solve for the gas density , @xmath105 we substitute @xmath91 back into eq .
( [ eq : tff ] ) to obtain the freefall time @xmath106 for the average gas density @xmath91 .
finally , we obtain our prediction for @xmath1 either by multiplying the freefall time from sec .
[ sec : tff ] by @xmath17 calculated in sec .
[ sec : gastff ] , i.e. , using eq .
( [ eq : gas_single_tff ] ) , or by multiplying the volume density @xmath91 from eq .
( [ eq : rho ] ) by the column length @xmath96 from eq .
( [ eq : l ] ) . in terms of the principle observables , @xmath3 and @xmath107 , as well as our assumptions for the parameters @xmath108 , @xmath14 and @xmath15 ,
this corresponds to the final expression for @xmath1 given by @xmath109^{2/3}.\end{aligned}\ ] ] two examples of the spatially resolved maps of estimated @xmath1 based on the new method provided by eq .
( [ eq : gas ] ) are shown in the right - hand panels of fig .
[ fig : maps ] . here we estimate the uncertainties in our @xmath1 prediction based on eq .
( [ eq : gas ] ) .
we derive the uncertainties by error propagation of all variables in eq .
( [ eq : gas ] ) .
first , we note that the dependence of @xmath1 on @xmath96 is weak ( @xmath110 ) and the dependence on @xmath4 is also relatively weak ( @xmath111 ) , which means that the uncertainties in @xmath96 and @xmath4 enter the final uncertainty in @xmath1 with a weight of 1/3 and 1/2 , respectively .
the strongest dependence of @xmath1 is on the sfr , i.e. , @xmath112 , so the uncertainties in @xmath3 are weighted by 2/3 , and we thus expect these to dominate the final uncertainties .
rigorously , the relative uncertainty @xmath113 from eq .
( [ eq : gas ] ) is given by @xmath114^{1/2},\end{aligned}\ ] ] where we approximated the denominator @xmath115 in eq .
( [ eq : gas ] ) as @xmath116 for the uncertainty propagation ( recall that we also assumed @xmath117 ) , because @xmath118 , based on our velocity dispersion cut and sound speed ( see sec .
[ subsec : veldisp ] ) . with typical relative uncertainties of 70% in @xmath96 ,
@xmath100 in @xmath4 ( see sec . [
sec : mach ] ) and @xmath119 in @xmath3 ( based on our s / n cuts of 5 on the h@xmath2 flux ; see sec . [
sec : oursubsample ] ) , we find a relative uncertainty of @xmath120 , which is dominated by the uncertainty in @xmath3 .
even if the uncertainties in both @xmath96 and @xmath4 were @xmath100 and @xmath121 respectively , we would still be able to estimate @xmath1 with an uncertainty of @xmath122 . in summary , despite the large uncertainties in @xmath4 and @xmath96 ( see sec . [
sec : mach ] and [ sec : tff ] ) , our final uncertainties in @xmath1 are less than a factor of 2 .
our main objective is to estimate @xmath1 from @xmath3 and the turbulence properties ( @xmath4 ) in our sami galaxy sample .
we do this by applying the new method introduced in the previous section ( sec .
[ sec : theory ] ) , going step - by - step from @xmath3 to @xmath1 .
[ fig : derivation ] shows each of the @xmath3 parameterisations explored in sfk15 , presented in the same order as the computations of our @xmath1 derivations ( sec .
[ sec : theory ] ) .
the framework of the first panel assumes a direct correlation between @xmath3 and @xmath12 .
that is , it assumes the star formation relation of eq .
( [ eq : newsflaw ] ) to hold , thus by construction the sami data points in this framework lie along the same line .
the data points from sfk15 which were used to obtain this relation are also shown .
we note that in the sfk15 derivation of eq .
( [ eq : newsflaw ] ) , the k98 galaxies were omitted because they did not have @xmath12 values assigned to them due to their lack of @xmath4 measurements .
they are thus similarly excluded in this panel .
compared to the observational data published in sfk15 , we updated and corrected some of the previous data , and added new observations in figure [ fig : derivation ] . first , we replace the @xcite data for the smc by the most recent @xmath123 resolution data provided in @xcite ( j16 ) .
we also add the large magellanic cloud ( lmc ) data from @xcite and assume that the smc and lmc data have mach numbers in the range , i.e. , we basically treat the mach number as unconstrained , i.e. , varying in a plausible range , but we currently do not have direct measurements of @xmath4 in the smc or lmc . resolution data from @xcite and @xcite are more consistent with velocity dispersions that correspond to for the smc and lmc . however , without a direct measurement of the velocity dispersion and gas temperature , the mach number remains rather unconstrained for the smc and lmc . ]
second , we replace the global cmz data from @xcite by the local cmz cloud g0.253 + 0.016 ` brick ' @xcite for which significantly more information is available .
we take the values of @xmath1 , @xmath4 , @xmath14 , and @xmath15 measured in @xcite ( f16 ) and use the sfr per freefall time estimate of 2% from @xcite to obtain @xmath3 for the ` brick ' .
the other cloud data are identical to those published in kdm12 , f13 and sfk15 , which were taken from @xcite ( h10 ) , @xcite ( g11 ) , @xcite ( w10 ) , and @xcite ( h10 ) .
however , we corrected the error bar on the l10 clouds , which showed the standard deviation instead of the standard deviation of the mean in sfk15 .
we further propagated the uncertainties in @xmath3 , @xmath4 and @xmath16 between @xmath12 , @xmath17 and @xmath1 .
finally , we note that the observational data included in fig .
[ fig : derivation ] cover a wide range in spatial and spectral resolution ( for details we refer the reader to the source publications of these data ) , which allowed us to test the universality of the sfk15 relation . in the future , when turbulence estimates become available for high - redshift data ,
those need to be included as well , to revisit the question of universality of the star formation relation derived in sfk15 .
the second panel of fig .
[ fig : derivation ] depicts the kdm12 parameterisation , @xmath3 versus @xmath17 .
the derivation of this value for the sami galaxies required inputs from both the h@xmath2 flux and velocity dispersion , with @xmath17 computed from eq .
( [ eq : gastff ] ) .
in addition to the observational data shown in the left - hand panel , we added the individual k98 disc and starburst galaxies from kdm12 ( with corrections based on * ? ? ? * ; * ? ? ? * ) .
the third panel of fig .
[ fig : derivation ] shows the final product of our @xmath1 predictions ; the average gas column density estimate for each of the sami galaxies in our sample . these predictions span a range of .
we note that the estimated @xmath1 values for the sami galaxies are close to the @xmath1 values of the k98 low - redshift disc galaxies .
this is encouraging , because they are the most similar in type to our sample of sami galaxies .
the offset in @xmath1 and @xmath3 by @xmath124 between the sami and k98 galaxies can be understood as a consequence of spatial resolution .
in contrast to our spaxel - resolved analysis of the sami galaxies ( with spatial resolution of @xmath125 ; see sec .
[ sec : sami ] ) , the k98 galaxies are unresolved , which reduces the inferred @xmath3 @xcite .
the reason is that , although the total h@xmath2 flux ( @xmath126 ) remains similar even at lower resolutions , the area @xmath127 over which h@xmath2 is emitted tends to be overestimated and hence the @xmath3 tends to be underestimated ( @xmath128 ) for the global k98 data .
similar holds for @xmath1 , because it depends on @xmath127 in the same way as @xmath3 , and indeed , we find that the resolved sami galaxies tend to lie at somewhat higher @xmath1 compared to the unresolved k98 disc galaxy sample . for sami galaxies classified as pure star - forming galaxies ( blue ) and composite / agn / shock galaxies ( orange ) .
] lcc physical parameter & average ( standard deviation ) & average ( standard deviation ) + & for star - forming galaxies & for composite / agn / shock galaxies + @xmath129 $ ] & @xmath130 & @xmath131 + + mach number ( @xmath33 ) & @xmath132 & @xmath133 + @xmath134 $ ] & @xmath135 & @xmath136 + + mach number ( @xmath33 ) & @xmath137 & @xmath138 + @xmath134 $ ] & @xmath139 & @xmath140 + + mach number ( @xmath33 ) & @xmath141 & @xmath142 + @xmath134 $ ] & @xmath143 & @xmath144 + [ tab : sum ] * notes . *
note that the values in brackets denote the standard deviation ( galaxy - to - galaxy variations ) of each physical parameter ; not the uncertainty in the parameter .
uncertainties are discussed in sec .
[ sec : uncertainties ] .
[ fig : derivation ] suggests that the distributions of @xmath1 and @xmath3 are similar between the star - forming and composite / agn / shock galaxies . to quantify any statistical differences in @xmath1 between these two sub - samples , we investigate the distribution functions of @xmath1 .
[ fig : gas_histograms ] shows the histograms of @xmath1 .
we see that @xmath1 is enhanced in composite / agn / shock galaxies compared to star - forming galaxies .
this difference in @xmath1 between star - forming and composite / agn / shock type galaxies is primarily a consequence of the differences in @xmath3 , and secondarily a consequence of the differences in @xmath4 between the two classes , i.e. , the dependences of @xmath1 on @xmath3 and @xmath4 ( see eqs .
[ eq : gas ] and [ eq : gas_uncertainty ] )
. other dependences are realtively insignificant , such as the dependence on the assumed scale height of the galaxies . in this context
, we have checked that any differences in the ellipticity distributions between the star - forming and composite / agn / shock galaxies are statistically insignificant . the measured mean and standard deviation of @xmath3 , @xmath4 and @xmath1 in the star - forming and composite / agn / shock galaxy samples
are listed in table [ tab : sum ] ( the full list of physical parameters derived for each galaxy is provided in table [ tab : galaxies ] ) .
the sfr surface densities and mach numbers of the star - forming and composite / agn / shock sample are @xmath145 and @xmath146 , and @xmath147 and @xmath148 , respectively .
the resulting average @xmath1 values are @xmath149 and @xmath150 for star - forming and composite / agn / shock galaxies , respectively .
table [ tab : sum ] further shows that changing the beam - smearing cutoff from the fiducial @xmath82 to a less strict cutoff ( @xmath83 ) yields nearly identical results .
finally , the last two rows of table [ tab : sum ] show that using the velocity dispersion of the ionized gas ( @xmath50 ) instead of the approximate velocity dispersion of the molecular gas ( @xmath48 ) , reduces the derived @xmath1 by @xmath151 .
thus , even with the large uncertainties in @xmath152 and hence in @xmath4 ( see sec . [
sec : mach ] ) , our final estimates in @xmath1 can be considered accurate to within a factor 2 . ) as a function of the derived @xmath1 for star - forming ( blue ) and composite / agn / shock ( orange ) galaxies in our sami sample .
the large open circles show the same data , but binned in steps of 0.1 dex in @xmath1 .
the error bars show the standard deviation of the mach number in each bin . despite the significant variations in mach number per individual @xmath1 bin , composite / agn / shock galaxies tend to have on average @xmath153 higher @xmath4 than star - forming galaxies . ] here we investigate global differences in the gas kinematics between star - forming and composite / agn / shock galaxies in our sami sample .
[ fig : mach_gas ] shows our measurements of the mach number ( c.f .
[ sec : mach ] ) as a function of derived @xmath1 for the two galaxy classes .
we see that overall and also for fixed @xmath1 , composite / agn / shock galaxies have higher mach number by a factor of @xmath154 compared to star - forming galaxies .
this may be a consequence of agn and/or shocks raising the velocity dispersion over turbulence driven by pure star - formation feedback .
[ fig : mach_gas ] further reveals a significant scatter in mach number for fixed @xmath1 , which is somewhat more pronounced in the case of composite / agn / shock galaxies compared to purely star - forming ones .
this may indicate that different driving sources of the turbulence act together and possibly dominate at different times in different galaxies .
such driving sources can be divided into two main categories : i ) stellar feedback ( such as supernova explosions , stellar jets , and/or radiation pressure ) and ii ) galaxy dynamics ( such as galactic shear , magneto - rotational instability , gravitational instabilities , and/or accretion onto the galaxy ) @xcite . our results here suggest that agn feedback may be another important , potentially highly variable source of the turbulent gas velocity dispersion in galaxies .
many studies in the literature have attempted to measure the correlation between @xmath3 and @xmath1 within different sets of data .
however , there is no clear consensus on the coefficients and scaling exponents , due to the intrinsic scatter in the kennicutt - schmidt relation ( kdm12 , * ? ? ?
* sfk15 ) .
some studies find breaks in the power - law relations , which can be interpreted as thresholds @xcite , while other studies do not find evidence for such thresholds @xcite . here
we explore how our @xmath1 predictions for the sami galaxies compare to the relations published in the literature . in fig .
[ fig : gas_comp ] , we show an enhanced version of the right - hand panel of fig . [
fig : derivation ] , in order to compare our @xmath1 estimates to previously derived star formation relations in the @xmath3 versus @xmath1 framework .
the relations we investigate are described in k98 , @xcite ( b08 ) , @xcite ( w10 ) , @xcite ( h10 ) and @xcite ( b11 ) .
these relations are shown as lines in fig .
[ fig : gas_comp ] .
we see that different sets of data follow different relations , which often show significant deviations from one another .
the sami galaxies have higher @xmath3 than described by the k98 , b08 , h10 , or b11 relations , but lower @xmath3 than described by the w10 relation .
a power - law fit to all the sami galaxies yields a power - law exponent of @xmath155 instead of @xmath156 ( k98 ) ; see eq .
( [ eq : k98 ] ) . in summary
, we find that none of the previously proposed scaling relations of @xmath3 as a function of @xmath1 describes the entirety of the data well .
the reason for this is that the sfr ( @xmath3 ) depends on more than just gas density ( @xmath1 ) .
instead , star formation also strongly depends on the turbulence of the gas ( mach number and driving mode ) , the magnetic field , and on the virial parameter ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* sfk15 ) .
a complete understanding and prediction of star formation requires taking into account the dependences on these variables , in addition to gas ( surface ) density .
here we compare the prediction of @xmath1 based on inverting the k98 relation with the @xmath1 prediction based on the sfk15 framework developed here .
first , we note that a popular way of obtaining @xmath1 estimates in the absence of a direct measurement of it , is to invert the k98 relation , i.e. , to invert eq .
( [ eq : k98 ] ) , which yields @xmath157 = \left(\frac{{\sigma_\mathrm{sfr}}\,\left[{\mathrm{{\mathrm{m_{\odot}}}\,{\mathrm{yr}}^{-1}\,{\mathrm{kpc}}^{-2}}}\right]}{a}\right)^{1/n},\ ] ] with @xmath158 and @xmath11 ( k98 ) .
here we derived an alternative way to estimate @xmath1 from @xmath3 , which is given by eq .
( [ eq : gas ] ) , and contains additional dependences on the geometry ( @xmath96 ) , turbulence ( @xmath4 and @xmath14 ) , and on the magnetic field ( @xmath15 ) . for the @xmath1 estimates of the sami galaxies here , we fixed @xmath96 , @xmath14 and @xmath15 for simplicity , and only included the mach number based on the measured velocity dispersion as an additional parameter to @xmath3 ( compared to the k98 relation , which depends on @xmath3 only ) .
we now want to see how the @xmath1 estimates based on our new relation ( eq .
[ eq : gas ] ) compare to inverting the k98 relation ( eq .
[ eq : k98inverse ] ) .
[ fig : sf_law_comp ] shows the direct comparison of the two as a function of @xmath3 .
we plot the logarithmic difference of @xmath1(predicted ) and @xmath1(measured ) on the ordinate of fig .
[ fig : sf_law_comp ] for @xmath1(predicted ) based on k98 ( eq . [ eq : k98inverse ] ) in grey and @xmath1(predicted ) based on eq .
( [ eq : gas ] ) in colour .
in addition to the observational data already shown in figs .
[ fig : derivation ] and [ fig : gas_comp ] , we add direct estimates of @xmath1 for a subset of 56 star - forming sami galaxies for which _ herschel _ @xmath159 dust measurements were available , using the methods in @xcite . a detailed description of how the dust emission was converted to @xmath1 is provided in appendix [ app : herschel_dust_to_gas ] . in fig .
[ fig : sf_law_comp ] we see that our new method of estimating @xmath1 from @xmath3 given by eq .
( [ eq : gas ] ) is significantly better than simply inverting the k98 relation , eq .
( [ eq : k98inverse ] ) .
we find that our new relation provides @xmath1 estimates with an average deviation of 0.12 dex ( 32% ) , while inverting the k98 relation yields an average deviation from the true ( measured ) @xmath1 by 0.42 dex ( 160% ) .
this shows that our method provides a significantly better @xmath1 prediction from @xmath3 than inverting the k98 relation .
our improved @xmath1 estimate comes at the cost of requiring an estimate of the mach number ( velocity dispersion ) as an additional parameter for the reconstruction ( prediction ) of @xmath1 .
however , if @xmath3 is obtained from h@xmath2 ( as for the sami galaxies analysed here ) , then we have shown that the velocity dispersion of h@xmath2 can be used to estimate the mach number ( sec .
[ sec : mach ] ) .
even better @xmath1 predictions based on eq .
( [ eq : gas ] ) are expected if the exact scale height @xmath94 , the turbulence driving parameter @xmath14 , and the magnetic field plasma @xmath15 are available from future observations and/or by combining different observational datasets .
we presented a new method to estimate the molecular gas column density ( @xmath1 ) of a galaxy using only optical ifs data , by inverting the star formation relation derived in sfk15 .
our method utilises observed values of @xmath3 and velocity dispersion ( here from h@xmath2 ) as inputs and returns an estimate of the molecular @xmath1 .
the derivation of our method is explained in detail in sec .
[ sec : theory ] , with the final result given by eq .
( [ eq : gas ] ) .
we apply our new method to estimate @xmath1 for star - forming and composite / agn / shock galaxies classified and observed in the sami galaxy survey .
our main findings from this study are the following : * from the range in and mach number measured for the sami galaxies , we predict in the star - forming regions of our sami galaxy sample , consisting of 260 galaxies in total .
the predicted values of @xmath1 are similar to those of unresolved low - redshift disc galaxies observed in k98 .
while the k98 galaxies required co detections , here we estimate @xmath1 solely based on h@xmath2 emission lines . *
we classify each galaxy in our sample as star - forming or composite / agn / shock .
based on the sample - averaged @xmath145 and @xmath146 , and @xmath147 and @xmath148 for star - forming and composite / agn / shock galaxies , respectively , we estimate @xmath160 and @xmath150 , respectively ( see table [ tab : sum ] ) .
we therefore find that on average , the composite / agn / shock galaxies have enhanced @xmath3 , @xmath4 , and @xmath1 by factors of @xmath5 , @xmath6 , and @xmath7 , respectively , compared to the star - forming sami galaxies ( see table [ tab : sum ] ; for each individual sami galaxy , see table [ tab : galaxies ] ) .
* we discussed methods to account for finite spectral resolution and beam - smearing in sec .
[ subsec : veldisp ] . while the uncertainties are large in the velocity dispersion used to estimate the turbulent mach number of the molecular gas ( sec .
[ sec : mach ] ) , we show that the final estimate of @xmath1 is accurate to within a factor of 2 ( see sec . [
sec : uncertainties ] ) .
* we compare our new method of estimating @xmath1 from @xmath3 with a simple inversion of the k98 relation ( fig .
[ fig : sf_law_comp ] ) .
we find that our new method yields a significantly better estimate of @xmath1 than inverting the k98 relation , with average deviations from the intrinsic @xmath1 by 32% for our new method , compared to average deviations of 160% from inverting the k98 relation .
we thank mark krumholz and the anonymous referee for their useful comments , which helped to improve this work .
cf acknowledges funding provided by the australian research council s ( arc ) discovery projects ( grants dp150104329 and dp170100603 ) .
dms is supported by an australian government s new colombo plan scholarship .
support for amm is provided by nasa through hubble fellowship grant # hst - hf2 - 51377 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas5 - 26555 .
bag gratefully acknowledges the support of the arc as the recipient of a future fellowship ( ft140101202 ) .
ljk gratefully acknowledges the support of an arc laureate fellowship .
sb acknowledges the funding support from the arc through a future fellowship ( ft140101166 ) .
smc acknowledges the support of an arc future fellowship ( ft100100457 ) .
ns acknowledges support of a university of sydney postdoctoral research fellowship .
the sami galaxy survey is based on observations made at the anglo - australian telescope .
the sydney - aao multi - object integral field spectrograph ( sami ) was developed jointly by the university of sydney and the australian astronomical observatory .
the sami input catalogue is based on data taken from the sloan digital sky survey , the gama survey and the vst atlas survey .
the sami galaxy survey is funded by the arc centre of excellence for all - sky astrophysics ( caastro ) , through project number ce110001020 , and other participating institutions .
the sami galaxy survey website is http://sami - survey.org/.
table [ tab : galaxies ] lists the derived physical parameters of all sami galaxies analysed here . listed
are the first ten galaxies in each of our two galaxy classes ( star - forming and composite / agn / shock ) .
the complete table is available in the online version of the journal or upon request .
lccccccccccc gama & redshift & @xmath97 & @xmath161 & @xmath162 & @xmath3 & @xmath33 & @xmath91 & @xmath163 & @xmath12 & @xmath17 & @xmath1 + i d & & & & [ @xmath164 & @xmath165}$ ] & & @xmath166 $ ] & [ @xmath167 & @xmath165}$ ] & @xmath165}$ ] & @xmath168}$ ] + + @xmath169 & @xmath170 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & @xmath191 & @xmath192 + @xmath193 & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath202 & @xmath203 & @xmath204 + @xmath205 & @xmath206 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 + @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath176 & @xmath177 & @xmath224 & @xmath225 & @xmath226 + @xmath227 & @xmath228 & @xmath171 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 + @xmath238 & @xmath239 & @xmath195 & @xmath240 & @xmath241 & @xmath198 & @xmath242 & @xmath200 & @xmath201 & @xmath202 & @xmath243 & @xmath244 + @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath231 & @xmath250 & @xmath251 & @xmath252 & @xmath235 & @xmath253 & @xmath254 + @xmath255 & @xmath246 & @xmath247 & @xmath256 & @xmath249 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 + @xmath264 & @xmath265 & @xmath266 & @xmath248 & @xmath209 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 & @xmath272 & @xmath273 + & & & & & & & & & & &
+ + @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 + @xmath286 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath282 & @xmath294 & @xmath295 & @xmath244 + @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath173 & @xmath300 & @xmath301 & @xmath302 & @xmath201 & @xmath303 & @xmath304 & @xmath305 + @xmath306 & @xmath170 & @xmath307 & @xmath308 & @xmath173 & @xmath309 & @xmath199 & @xmath176 & @xmath177 & @xmath310 & @xmath311 & @xmath312 + @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath222 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 + @xmath324 & @xmath314 & @xmath325 & @xmath326 & @xmath317 & @xmath327 & @xmath328 & @xmath329 & @xmath330 & @xmath310 & @xmath295 & @xmath331 + @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 & @xmath343 + @xmath344 & @xmath345 & @xmath288 & @xmath196 & @xmath346 & @xmath347 & @xmath301 & @xmath348 & @xmath310 & @xmath271 & @xmath349 & @xmath350 + @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath355 & @xmath356 & @xmath357 & @xmath358 & @xmath202 & @xmath359 & @xmath360 & @xmath361 + @xmath362 & @xmath352 & @xmath363 & @xmath354 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath179 & @xmath370 + & & & & & & & & & & & + * notes . *
all galaxy parameters are based on a straight average over all valid spaxels and the uncertainties were propagated based on the average values .
column 1 : gama i d .
column 2 : redshift .
column 3 : ellipticity .
column 4 : number of valid spaxels for gas column density estimate .
column 5 : linear size of each spaxel .
column 6 : spaxel - averaged @xmath371 .
column 7 : spaxel - averaged turbulent mach number .
column 8 : spaxel - averaged gas volume density , estimated based on eq .
( [ eq : rho ] ) .
column 9 : spaxel - averaged freefall time based on eq .
( [ eq : tff ] ) .
column 10 : spaxel - averaged multi - freefall gas consumption rate , @xmath12 ; eq .
( [ eq : gastmff ] ) .
column 11 : spaxel - averaged single - freefall gas consumption rate , @xmath17 ; eq .
( [ eq : gastff ] ) .
column 12 : spaxel - averaged molecular gas surface density @xmath1 , estimated with eq .
( [ eq : gas ] ) .
[ tab : galaxies ]
to provide an independent measure of @xmath1 for our sami galaxy sample , we used the empirical relation determined by @xcite , correlating the total gas mass of galaxies with their sub - mm dust luminosities . using a sample of nearby galaxies
, @xcite found that the total ( atomic@xmath372molecular ) gas mass of galaxies ( @xmath373 ) could be determined within 0.12 dex using the monochromatic @xmath159 luminosity ( @xmath374 ) , with @xmath375 to determine the sub - mm luminosity of the sami galaxies , we made use of the _ herschel_-atlas survey @xcite , a wide 550 square degrees infrared survey of the sky by the _ herschel _ space observatory , that covers the gama regions from which the sami galaxy survey sample arise .
in particular , we cross - matched the 219 star - forming sami galaxies classified here against the single - entry source catalog from _
herschel_-atlas data release 1 @xcite .
of the 219 sami star - forming galaxies , 128 have _ herschel _ detections . of these ,
56 have significant detections ( signal - to - noise @xmath376 ) in the spire @xmath159 band . to convert the total gas mass to a gas surface density we required a surface area over which the infrared flux is emitted .
given the large beam size of the spire @xmath159 observations , the sami galaxies are unresolved . however , as can be seen in the radial profiles of the nearby galaxy sample used in @xcite ( in particular their figure 7 and online figures ) , the highest surface brightness regions occur within half an optical radius ( @xmath377 or @xmath378 based on * ? ? ?
* ) , with most of the infrared luminosity ( and molecular gas mass ) occurring within this radius .
@xcite further find that at this radius , the atomic and molecular gas surface densities are about the same ( the ratio of the total atomic and molecular gas masses inside @xmath379 is also about unity ) . based on those findings
, we approximated the molecular gas mass within @xmath379 with @xmath380 .
therefore , we derive the molecular @xmath1 through @xmath381 where the effective radius @xmath382 , and ellipticity @xmath97 , of the sami galaxies are as derived in the gama survey @xcite . , l. j. , & dopita , m. a. 2003 , in revista mexicana de astronomia y astrofisica conference series , vol .
17 , revista mexicana de astronomia y astrofisica conference series , ed .
v. avila - reese , c. firmani , c. s. frenk , & c. allen , 8384 | stars form in cold molecular clouds . however , molecular gas is difficult to observe because the most abundant molecule ( @xmath0 ) lacks a permanent dipole moment . rotational transitions of co are often used as a tracer of @xmath0 , but co is much less abundant and the conversion from co intensity to @xmath0 mass is often highly uncertain . here we present a new method for estimating the column density of cold molecular gas ( @xmath1 ) using optical spectroscopy .
we utilise the spatially resolved h@xmath2 maps of flux and velocity dispersion from the sydney - aao multi - object integral - field spectrograph ( sami ) galaxy survey .
we derive maps of @xmath1 by inverting the multi - freefall star formation relation , which connects the star formation rate surface density ( @xmath3 ) with @xmath1 and the turbulent mach number ( @xmath4 ) . based on the measured range of and , we predict in the star - forming regions of our sample of 260 sami galaxies .
these values are close to previously measured @xmath1 obtained directly with unresolved co observations of similar galaxies at low redshift .
we classify each galaxy in our sample as ` star - forming ' ( 219 ) or ` composite / agn / shock ' ( 41 ) , and find that in ` composite / agn / shock ' galaxies the average @xmath3 , @xmath4 , and @xmath1 are enhanced by factors of @xmath5 , @xmath6 , and @xmath7 , respectively , compared to star - forming galaxies .
we compare our predictions of @xmath1 with those obtained by inverting the kennicutt - schmidt relation and find that our new method is a factor of two more accurate in predicting @xmath1 , with an average deviation of 32% from the actual @xmath1 .
[ firstpage ] galaxies : ism galaxies : star formation galaxies : structure stars : formation techniques : spectroscopic turbulence |
understanding the formation and evolution of the minor planets in our solar system requires good knowledge of the physical parameters that describe each population of asteroids and comets .
models of asteroid migration , for example , are critically dependent on measurements of number counts , sizes and albedos .
visible light surveys have supplied most of the discoveries of asteroids and comets to date ; however , these surveys tend to be biased against low albedo objects , and so they are underrepresented in number counts .
furthermore , diameters derived from visible light observations are highly uncertain due to their linear dependence on albedo , which can vary by an order of magnitude for many types of asteroids .
radar observations , stellar occultations and in situ spacecraft imaging provide powerful means of obtaining precise diameter measurements , but these observations are limited to a small subset of the @xmath1500,000 minor planets known to exist today . while diameter and albedo can be measured more precisely than with visible light by applying radiometric models to objects for which thermal infrared data have been obtained , the size and albedo distributions obtained in this way will still be biased if the underlying source population was discovered by visible light surveys . in order to mitigate against these biases , it is necessary to undertake a survey that is capable of independent detection and discovery of asteroids at thermal infrared wavelengths . yet
the physical parameters such as size and albedo that are derived from such a new survey must be checked against other well - characterized datasets to ensure their reliability .
the _ wide - field infrared survey explorer _ ( wise ) is a nasa medium - class explorer mission designed to survey the entire sky in four infrared wavelengths , 3.4 , 4.6 , 12 and 22 @xmath2 m ( denoted @xmath3 , @xmath4 , @xmath5 , and @xmath6 respectively ) @xcite .
the final mission data products are a multi - epoch image atlas and source catalogs that will serve as an important legacy for future research .
the portion of the pipeline dedicated to finding minor planets ( called neowise ) has yielded observations of over 157,000 minor planets , including near - earth objects ( neos ) , main belt asteroids ( mbas ) , comets , hildas , trojans , centaurs , and scattered disk objects @xcite .
this represents an improvement of nearly two orders of magnitude more objects observed than wise s predecessor mission , the _ infrared astronomical satellite _ ( iras ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the wise survey began on 14 january , 2010 , and the mission exhausted its primary tank cryogen on 5 august , 2010 .
exhaustion of the secondary tank and the start of the neowise post - cryogenic mission occurred on 1 october , 2010 , and the survey ended on 31 january , 2011 .
in @xcite , we demonstrated that thermal models created for minor planets using the wise / neowise dataset produce results that are in good agreement with diameters independently measured by radar , stellar occultations , and in situ spacecraft imaging . by taking the previously measured diameters from methods derived independently of infrared thermal model assumptions , we were able to verify the accuracy of the color corrections given in @xcite for low temperature objects such as asteroids .
we used the previously measured diameters combined with the @xcite color corrections in conjunction with a faceted spherical thermal model based on the near - earth asteroid thermal model @xcite to predict the wise magnitudes for @xmath150 asteroids ; these were compared with the observed wise magnitudes .
we found that there were no systematic offsets between predicted and observed magnitudes for these objects , indicating that the color corrections given in @xcite adequately describe the system response . for objects with wise measurements in two or more bands with good signal - to - noise ( for which the beaming parameter @xmath7 can be fit )
, we found that diameters can be determined to within @xmath810% , and visible albedo to within @xmath820% . in this paper ,
we compare diameters and obtained with wise / neowise observations for 1742 asteroids to those determined by iras and find that they are in generally good agreement with some slight systematic differences .
this work , combined with @xcite , demonstrates that the neowise observations of minor planets will produce good physical parameters that will enable a wide range of scientific investigations .
we have assembled a list of objects that were observed by iras that neowise detected during the fully cryogenic portion of its mission . of the @xmath12200 asteroids observed by iras @xcite
, we identified neowise detections for 1742 objects .
the observations of these objects were retrieved by querying the minor planet center s ( mpc ) observation files to look for all instances of individual wise detections of the desired objects that were reported using the wise moving object processing system ( wmops ; * ? ? ? * ) .
the resulting set of position / time pairs were used as the basis of a query of wise source detections in individual exposures ( also known as level 1b " images ) using the infrared science archive ( irsa ) . in order to ensure that only observations of the desired moving object were returned from the query ,
the search radius was restricted to 0.3 arcsec from the position listed in the mpc observation file .
additionally , since wise collected a single exposure every 11 seconds and observes each part of the sky an average of 10 times , the modified julian date was required to be within 2 seconds of the time specified by the mpc .
the following flag values were allowed : cc_flags = 0 , p or p and ph_qual = a , b , or c. objects brighter than @xmath9 and @xmath10 magnitudes were assumed to have flux errors equivalent to 0.2 magnitudes due to changes to the shape of the point spread function as the objects became saturated , and a linear correction was applied to the @xmath5 magnitudes in this brightness regime ( the wise explanatory supplement contains a more detailed explanation ) . per the explanatory supplement , objects brighter than @xmath11 and @xmath12 were not used .
each object had to be observed a minimum of three times in at least one wise band , and it had to be detected at least 40% of the time when compared to the band with the maximum number of detections ( usually , though not always , @xmath5 ) .
the wmops system is designed to reject inertially fixed objects such as stars and galaxies in bands @xmath5 and @xmath6 .
nonetheless , the individual images at all wavelengths were compared with wise atlas coadd and daily coadd source lists to ensure that inertially - fixed sources such as stars and galaxies were not coincident with the moving object detections .
this check is particularly important in bands @xmath3 and @xmath4 where the density of background objects ( and hence the probability of a blended source ) is higher than at longer wavelengths .
any remaining blended sources in bands @xmath3 and @xmath4 were removed .
some objects were observed at multiple epochs , and observations separated by more than three days were modeled separately .
in the standard thermal model ( stm ) of @xcite , the temperature of an asteroid is assumed to be maximum at the subsolar point and zero on the point opposite of this ; this is the case of an object with zero thermal inertia .
in contrast , in the fast rotating model ( frm ) @xcite , the asteroid is assumed to be rotating much faster than its cooling time , resulting in a constant surface temperature across all longitudes .
the so - called beaming parameter " ( @xmath7 ) was introduced by @xcite in the stm to account for the enhancement of thermal radiation observed at small phase angles . the near - earth asteroid thermal model ( neatm ) of @xcite
also uses the beaming parameter @xmath7 to account for cases intermediate between the stm and frm models . in the stm , @xmath7 is set to 0.756 to match the occultation diameters of ( 1 ) ceres and ( 2 ) pallas , while in the frm , @xmath7 is equal to @xmath13 . with neatm , @xmath7 is a free parameter that can be fit when two or more infrared bands are available .
we modeled each object as a set of triangular facets covering a spherical surface with diameter equal to the ground - truth measurement ( c.f .
the temperature for each facet was computed , and the @xcite color corrections were applied to each facet .
the emitted thermal flux for each facet was calculated using neatm along with the bandcenters and zero points given in @xcite ; nightside facets were assumed to contribute no flux .
the emissivity , @xmath14 , was assumed to be 0.9 for all wavelengths ( c.f . * ? ? ?
the objects absolute magnitudes ( @xmath15 ) were taken from @xcite when available ; otherwise , the values were taken from the mpc s orbital element files .
unless a direct measurement of @xmath16 was available from @xcite , we assumed a @xmath16 value of 0.15 .
in general , minor planets detected by neowise in bands @xmath3 and @xmath4 contain a mix of reflected sunlight and thermal emission .
thus , it was necessary to incorporate an estimate of reflected sunlight into the thermal model in order to use data from bands @xmath3 and @xmath4 . in order to compute the fraction of reflected sunlight in bands @xmath3 and
@xmath4 , it was also necessary to compute the ratio of the albedo at these wavelengths compared to the visible abledo ( ) ; the solar spectrum was approximated as a 5778 k blackbody . as described in @xcite , we assumed that @xmath17 .
the flux from reflected sunlight was computed for each wise band using the iau phase curve correction of @xcite .
thermal models were computed for each object by grouping together observations no more than a three day gap between them .
error bars on the model magnitudes and subsolar temperatures were determined for each object by running 25 monte carlo ( mc ) trials that varied the objects @xmath15 values by 0.3 magnitudes and the wise magnitudes by their error bars using gaussian probability distributions .
the minimum magnitude error for all wise measurements fainter than @xmath9 and @xmath18 magnitudes was 0.03 magnitudes , per the in - band repeatability measured in @xcite . for objects brighter than @xmath9 and @xmath18 ,
the error bars were increased to 0.2 magnitudes , as these magnitudes represent the onset of saturation ( see the wise explanatory supplement ) .
for those objects for which @xmath7 and could not be fitted , @xmath7 was set to 1.0 and allowed to vary in the mc trials by 0.25 , and was set to 1.4 and allowed to vary by 0.5 .
these default values were used because they are the mean and standard deviation of @xmath7 and for the objects for which these parameters could be fit . the error bar for each object
s model magnitude was equal to the standard deviation of all the mc trial values .
although many asteroids are known to be non - spherical , the wise observations generally consisted of @xmath110 - 12 observations per object uniformly distributed over @xmath136 hours @xcite , so on average , a wide range of rotational phases was sampled .
although the variation in effective spherical diameter resulting from rotational effects tends to be averaged out , caution must be exercised when interpreting effective diameter results using spherical models for objects that are known to have large - amplitude lightcurve variations . in our sample , 179 objects have neowise observations at multiple epochs ( meaning that the groups of observations were separated by more than 10 days ) . of these , all but 24 had diameters that agreed to within 10% , and most of the remaining objects had @xmath5 peak - to - peak amplitudes @xmath00.3 mag , indicating that they are likely to be non - spherical .
non - spherical objects observed at different viewing geometries can lead to different diameters when using a spherical model .
however , since 87% of the objects with multi - epoch observations have diameters that agree to within 10% , we conclude that multiple observational epochs do not contribute significantly to the differences observed between iras and neowise diameters .
@xcite reported observations of @xmath12200 asteroid within the supplemental iras minor planet survey ( simps ) .
the diameters and albedos in that work were computed using the standard thermal model .
similar to the work of @xcite , who computed neatm diameters and albedos for 654 asteroids using iras fluxes , @xcite applied a spherical neatm model to @xmath11500 asteroids found in the simps and mid - course space experiment ( msx ; * ? ? ?
figures [ fig : iras_diameter_diff ] and [ fig : ryan_diameter_diff ] show the comparison between diameters found using neowise observations and those given by @xcite and @xcite , respectively .
figures [ fig : iras_albedo_diff ] and [ fig : ryan_albedo_diff ] show the comparison between neowise and tedesco / ryan albedos .
the diameters given by @xcite are systematically lower than the neowise diameters , while the diameters computed by @xcite are systematically higher . also , given by @xcite is systematically higher than the neowise values , but the @xcite values tend to be slightly lower . in figure
[ fig : radar ] , we show the diameters of @xcite and @xcite compared to the diameters of 78 objects with diameters measured by radar , occultation , or spacecraft imaging ( see @xcite for a complete list of references ) .
these objects have been selected to have @xmath5 peak - to - peak amplitudes @xmath190.5 magnitudes to avoid the worst complications caused by applying spherical models to highly elongated objects . as can be seen in the figure ,
the @xcite diameters are skewed toward slightly smaller sizes than the radar , etc .
measurements , whereas the @xcite diameters are biased toward larger sizes .
we propose that these biases are caused by the band - to - band color corrections derived for iras .
as described in @xcite , band - to - band corrections were derived by requiring that the 10 and 20 @xmath2 m iras observations simultaneously matched a diameter derived from a stellar occultation of ( 1 ) ceres . because the stm parameters were set to match ceres
, @xcite found a 7% difference between diameters derived for 13 objects with diameters measured from stellar occultations .
the neatm model applied by @xcite does not appear to address this issue .
furthermore , @xcite applied cutoffs to @xmath7 of 0.75 and 2.75 .
as described above , in our neatm models applied to the neowise data , we cut off @xmath7 at the limit predicted by the fast rotating model , @xmath13 . in the stm , @xmath7 was set to 0.756 by matching the diameters of ceres and pallas ; however , there is no reason in principle why it can not range to lower values . in our application of neatm , we allow @xmath7 to range arbitrarily low and find no fitted @xmath7 values less than @xmath10.53 . as described above , in @xcite , we used asteroids with previously known diameters and @xmath15 magnitudes to compare our model and observed wise magnitudes , and we found no systematic offsets . we conclude that the diameters for minor planets derived from neowise are in generally good agreement with those found by iras and are likely more free of systematic biases than the diameters provided in either @xcite or @xcite . together with @xcite ,
this demonstrates that the neowise dataset will produce good quality physical parameters for the @xmath0157,000 minor planets it contains .
bowell , e. , hapke , b. , domingue , d. , lumme , k. , peltoniemi , j. , harris , a. w. in asteroids ii : proceedings of the conference , tucson , az , mar .
8 - 11 , 1988 ( a90 - 27001 10 - 91 ) .
tucson , az , university of arizona press , 1989 , p. 524 - 556 .
harris , a. , mommert , m. , hora , j. , mueller , m. , trilling , d. , bhattacharya , b. , bottke w. , chesley , s. , delbo , m. , emery , j. , fazio , g. , mainzer , a , penprase , b. , smith , h. , spahr , t. , stansberry , j. , thomas , c. 2011 aj 141 , 75 comparison of the difference in diameters given by @xcite and the diameters found by applying a spherical neatm model to neowise measurements for 1742 asteroids .
it can be seen that the values from @xcite tend to be slightly smaller than neowise - derived diameters .
the neowise diameters diverge widely from iras for the smallest objects ; however , these objects are observed by neowise with high signal - to - noise ratio .
, width=576 ] comparison of the difference in diameters given by @xcite and the diameters found by applying a spherical neatm model to neowise measurements for 1155 asteroids .
the diameters found by @xcite tend to be larger than those found by neowise .
, width=576 ] comparison of the difference in albedos given by @xcite and the diameters found by applying a spherical neatm model to neowise measurements for 1742 asteroids .
it can be seen that the values from @xcite tend to be slightly higher than neowise - derived albedos .
, width=576 ] comparison of the difference in albedos given by @xcite and the albedos found by applying a spherical neatm model to neowise measurements for 1155 asteroids .
the albedos found by @xcite tend to be lower than those found by neowise .
, width=576 ] comparison of diameters found by radar , stellar occultation , or in situ spacecraft imaging for 78 asteroids and the diameters found by @xcite and @xcite .
the objects were selected to have @xmath5 peak - to - peak amplitudes @xmath190.5 . it can be seen that both methods of computing iras diameters are somewhat biased when compared to the radar diameters . in @xcite , we show that there is no systematic bias found in model and observed wise magnitudes for a similar set of objects with independently measured diameters . , width=576 ] | with thermal infrared observations detected by the neowise project , we have measured diameters for 1742 minor planets that were also observed by the _ infrared astronomical satellite _ ( iras ) .
we have compared the diameters and albedo derived by applying a spherical thermal model to the objects detected by neowise and find that they are in generally good agreement with the iras values .
we have shown that diameters computed from neowise data are often less systematically biased than those found with iras .
this demonstrates that the neowise dataset can provide accurate physical parameters for the @xmath0157,000 minor planets that were detected by neowise . |
the radio source j133658.3295105 ( object 28 in @xcite list ) in m83 , is described as the core of an fr ii radio source @xcite whose radio lobes are objects 27 and 29 in the same list .
the source presents characteristics of a radio galaxy or a quasar at a distance @xmath6 as discussed by @xcite .
these authors have found that at a distance d@xmath7=4.7gpc ( @xmath8=1 , q@xmath9=0.5 , h=75kms@xmath2mpc@xmath2 ) , j133658.3295105 @xmath10 and @xmath11 fit reasonably well with the expected frii radio - sources luminosities .
m83 and ngc5128 ( cena ) are the largest galaxies in one of the most active regions in the nearby sky , the hydra - centaurus group .
merger processes as well as galactic inner build - up are evident in both galaxies .
the radio source ( rs ) 28 appears projected at about 60 from the nucleus of m83 and the line joining rss 27 , 28 , and 29 appears to emerge from the m83 bulge as seen in figure [ fig : global ] ( and figure1 of * ? ? ?
this alignment was originally proposed by @xcite to be a jet from the nucleus of m83 . moreover ,
rss 30 , 32 , and 36 , and the m83 optical nucleus ( on ) itself appears to be strongly aligned with this structure , too ( figure [ fig : global ] , right ) the whole structure covers @xmath12 on the sky .
none of these radio sources are supernova remnants nor do they appear associated with hii regions @xcite .
rs 32 emits x - rays as a soft x - ray binary .
rs 36 was also described as an x - ray binary with evidence for x - ray spectral evolution .
the optical nucleus of m83 is also a radio and x - ray source @xcite .
these intriguing features encouraged us to try to determine more precisely the distance to rs 28 .
the complex structure of the central 300pc of the bulge of m83 suggests the possibility of it hosting a kick - off object , which has been theoretically predicted @xcite , but not yet observed @xcite .
indeed , the condensation traditionally recognized as the m83 optical nucleus ( on ) ( figure [ fig : kick - off]c ) is off - centered by @xmath13(@xmath1480pc ) to the northeast with respect to the center of the external @xmath15-band isophotes of the bulge @xcite . at the bulge center , @xcite
have found the kinematic center ( kc ) .
more recently , @xcite demonstrated the existence of a third condensation ( hidden nucleus , hn ) , also hidden at optical wavelengths , 7@xmath16(@xmath14140pc ) to the west - southwest of kc which seems to be a small galaxy being cannibalized @xcite .
there is an arclet @xmath17 to the southwest of kc ( figure [ fig : kick - off]c ) with more than twenty 30-dor - like giant hii regions @xcite . pa@xmath18
observed around kc and hn indicates the presence of regions that are actively forming stars @xcite .
the derived mass of on is @xmath19 @xcite .
these authors proposed a similar mass for kc , while @xcite determined a larger value of @xmath160 @xmath20m@xmath21 .
hn has a mass of 16@xmath20m@xmath21 @xcite .
the upper mass limits for the putative black holes ( bhs ) associated with on , kc , and hn are 0.21.0@xmath22 m@xmath23 @xcite .
numerical simulations show that on , kc , hn and the arclet of hii regions will merge , forming a single massive core in a few hundred myr @xcite .
parallel to the merging process at the center of m 83 , molecular clouds in the bar seem to be funneled inward @xcite , guaranteeing central mass accretion in the future .
we discuss here the optical observations of the radio source j133658.3295105 , taken to more precisely derive its distance and its nature .
in 2007 april , we obtained a 5 hr spectrum with the reosc echelle spectrograph in a simple dispersion mode at the casleo 2.2 m telescope , in argentina .
reosc has a 1024@xmath241024 tek ccd , with a 24@xmath25 m pixel size .
a 1200 lines mm@xmath2 grating was used , covering the wavelength range from 6200 to 6900 around h@xmath26 .
the dispersion was 32 mm@xmath2 , the reciprocal dispersion 0.76 pix@xmath2 , the resolution 2.5 , and the angular scale was 102 pix@xmath2 .
the slit ( at p.a.=152@xmath27 ) was positioned on an hii region 25 to the to the east - southeast of rs28 .
these 2 seeing observations confirmed the nature of the hii region . on the basis of an accurate astrometric analysis of the source in different spectral domains
, we planned the subsequent gemini multi object spectrographs - south ( gmos - s ) observations . figure [ fig : rs28_position ] shows the astrometry and the hii region to the east - southeast of rs28 on a ( h@xmath28 ) cerro tololo inter - american observatory ( ctio ) image .
rs 28 was observed by gmos - s at gemini south in 2007 june ( program gs-2007a - dd-17 ) .
we performed broad and narrow band imaging in the range 500 to 950 nm , centered at the position of the radio source emission peak with average seeing of @xmath29 .
the observed bands and exposure times were @xmath30 : 90s ; [ oiii ] : 121s ; [ oiii]c : 121s ; h@xmath0 : 91.5s ; h@xmath0c : 91.5s ; [ sii ] : 91.5s ; @xmath31 : 60s ; cat : 60s and @xmath8 : 91.5s .
spectroscopy in the spectral range 572 - 988 nm was performed with an effective exposure time of 3600s and average seeing of @xmath29 .
a @xmath32 long slit was positioned on rs28 and oriented along the line of radio sources shown in figure [ fig : global ] ( p.a.=140@xmath27 ) .
figure [ fig : kick - off ] shows the co - added image from the observed gmos - s bands @xmath33 , [ oiii ] , [ oiii]c , h@xmath0 , h@xmath0c , [ sii ] , @xmath8 , @xmath31 , and cat .
we estimate that the detection limit at signal - to - noise ratio ( @xmath34 is m@xmath35 .
this is based on an average seeing of @xmath29 , and takes into account the background noise from the m83 disk .
it is expected that the disk of m83 will dim background objects . @xcite
derive an azimuthal mean absorption @xmath36mag at r around 1kpc , which translates into @xmath37 mg ( assuming @xmath38 ) .
@xcite obtained , for three face - on galaxies , an inter - arm absorption 0.20.5mag smaller than the azimuthal mean at 800 nm .
that would lead to an absorption of the order of @xmath39mag for background objects at the projected position of rs28 on the m83 disk .
these considerations lead us to estimate that our composed image could detect more than 80% of the broad- and narrow - line active galactic nuclei ( agns ) at @xmath40 in the sexsi sample ( see figure6 in * ? ? ?
* ) , irrespective of the redshift .
the blow up of the composed gmos image of the 20 around rs 28 in figure [ fig : kick - off]b does not show any sign of an associated optical source . we have determined the detectability of rs 28 in our images using cen a as an example .
the radio source cen a originates from the elliptical galaxy ngc5128 , which hosts a seyfert2 nucleus with a strong dust lane across its main body .
quasars with strong activity of star formation would present much dust , and probably would suffer internal absorption as ngc5128 does .
the absolute brightness of ngc5128 is m@xmath41= -23.0mag at a luminosity distance 11mpc , @xmath42 .
if cen a was located at @xmath43 , ngc5128 would shine at m@xmath44=21.5mag , or m@xmath44=23.1mag if located at @xmath45 .
therefore , if located somewhere in the redshift range @xmath46 , a cen a type galaxy would be detectable by our imagery .
the angular distance between the maxima of cen a radio lobes brightness is @xmath143@xmath27 ( 580kpc ) about twice the value quoted by @xcite for rs28 at @xmath47 .
that relation is maintained if rs28 is at larger @xmath8 , since the angular size distance does not change significantly with @xmath8 , for @xmath48 .
our spectroscopy suggests that we should detect any distant background agn - typical emission lines . in the co - added spectrum , the detection limit for the continuum of a giant elliptical galaxy at redshift 2 is m@xmath49 ( s / n@xmath50 ) , and it is m@xmath51 for h@xmath52 detection in a qso at @xmath53 .
strong redshifted lines should fall into the spectral range covered by gmos for @xmath54 , as can be seen in the revised catalog of qsos @xcite .
the most conspicuous emission lines and the corresponding @xmath8 intervals are : [ oii ] @xmath553727 ( 0.34,1.4 ) ; [ nv ] @xmath553426 ( 0.46,1.77 ) ; [ mgii ] @xmath552798 ( 0.8,2.4 ) ; [ ciii ] @xmath551909 ( 1.6,3.9 ) ; [ heii ] @xmath551640 ( 2.1,4.7 ) ; [ civ ] @xmath551549 ( 2.2,5.1 ) ; [ siv ] @xmath551402 ( 2.6,6.7 ) ; and [ nev ] @xmath551240 ( 3.4,7.6 ) .
our observational constrains set a limit of the order of @xmath56 for large radio galaxies .
we can get an idea of the expected s / n by looking at spectra obtained for broad- and narrow - lines agns at different distances by the keck observatory with exposure times of 1 to 2 hs @xcite . for a similar exposure time ( 3600s ) , we expect for gemini a s / [email protected] times that of keck .
so , practically all lines , except for [ heii ] would be easily seen in our spectra .
therefore , from our optical spectroscopy , we can confidently say that if rs28 is a high - redshift radio source , it should be at @xmath57 .
the rs28 spectrum is compared in figure [ fig : espectro1 ] with that of a simultaneously observed , well - defined hii region located on the arm to the north of rs28 .
we see that h@xmath0 emission appears marginally at the rs28 spectrum , redshifted with respect to that of the hii region .
radial velocities are respectively 618@xmath5830kms@xmath2 and 522@xmath5912kms@xmath2 .
the hii region located to the east - southeast of rs28 ( figure [ fig : rs28_position ] ) is at v@xmath60=490@xmath6112kms@xmath2 , based on our reosc - casleo data .
reosc accuracy can be gaged by the nuclear spectrum taken on the same night and instrument configuration , which gives for the nucleus a v@xmath60=534@xmath5815kms@xmath2 .
the heliocentric velocity of m83 , from the nasa / ipac extragalactic database ( ned ) , is v@xmath60=513@xmath595kms@xmath2 .
two - dimensional co observations @xcite provide for the nuclear region a value of v@xmath62kms@xmath2 ( error guess is from the position accuracy in figure3 of sakamoto et al . ) .
this value has to be added to the lsr velocity of approximately 24kms@xmath2 ( obtained with ed murphy s calculator in http:@xmath63fuse.pha.jhu.edu ) to compare with our measurement .
we conclude that reosc radial velocities are reliable within the errors .
the radial velocity of rs28 is approximately 130@xmath5940kms@xmath2 larger than that of the close east - southeast hii region .
the nucleus and the two hii regions show similar radial velocities because m83 presents a small inclination of 24@xmath27 with respect to the plane of the sky and the direction in discussion is at about 130@xmath27 from the galaxy major axis and practically perpendicular to the bar ( figure [ fig : kick - off ] ) where noncircular motions are detected @xcite .
this result leads us to conclude that the h@xmath0 emission marginally detected at rs28 is emitted by an object that does not take part of the disk kinematics , whose projected velocity in the radial direction is of the order of 100kms@xmath2 .
these arguments point to a possible physical connection between rs28 and m83 .
chance alignments do occur .
a striking case is that of the two radio sources juxtaposed 1 to the northeast and southwest of scorpius ( sco ) x-1 , which for a long time were believed to be radio lobes of sco x-1 , until @xcite demonstrated by measuring sco x-1 proper motion that they are unrelated background sources . on the one hand
, deep optical imaging and spectroscopy on the location of the radio source rs28 do not confirm the high - redshift nature of this object , and on the other hand , gmos / gemini - s spectroscopy marginally shows the presence of h@xmath0 in emission at a redshift similar to that of the galaxy but from a source kinematically well - differentiated from disk objects . if its local character is confirmed , the frii radio source rs28 would have a total projected size @xmath10.5 - 1.0kpc .
the possibility of rs28 being ejected by a gravitational recoil , remnant of a previous m83 nuclear merger , remains open . in that case
, the driving engine within rs28 that produced the ejecta rs27 and rs29 would be an accretion disk like those postulated for microquasars in the galaxy and extragalactic radio sources .
the reason why rs27 and rs29 sit along the apparent kick line remains unclear .
physical properties of accreting bhs ( temperature , ejecta size , etc ) scale with some power of the bh mass @xcite . if ejecta size scales with the bh mass , a direct comparison with the microquasar 1e1740.7 - 2942 ( m@xmath64 10m@xmath65 , size@xmath15pc ; @xcite ) leads to a mass for a rs28 bh @xmath141000 to 2000m@xmath65 .
the characteristic blackbody temperature for eddington - limit accretion @xcite should be @xmath66 or about 1/3 to 1/5 of that of a galactic microquasar .
x - ray and/or near - ir spectroscopic observations will be necessary to scan redshifts @xmath67 , and also around @xmath68 , to unequivocally determine the distance to and the nature of j133658.3295105 .
we acknowledge the support of the conicet grant pip 5697 , and h.d .
acknowledges support from cnpq ( brazil ) .
we thank jean - rene roy , percy gomez , maria paz agero and elena terlevich for discussions .
we also thank ms .
pippa clarke for a careful reading of the manuscript .
finally we acknowledge the gemini observatory for the allocation of director s discretionary telescope time , through the proposal gs-2007a - dd-17 .
the gemini observatory is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : nsf ( usa ) , stfc ( united kingdom ) , nrc ( canada ) , arc ( australia ) , mincyt ( argentina ) , cnpq ( brazil ) , and conicyt ( chile ) .
the nasa / esa hubble space telescope is operated by aura under nasa contract nas 5 - 26555 . | we present gemini optical imaging and spectroscopy of the radio source j133658.3295105 .
this source has been suggested to be the core of an frii radio source with two detected lobes . j133658.3295105 and
its lobes are aligned with the optical nucleus of m83 and with three other radio sources at the m83 bulge outer region .
these radio sources are neither supernova remnants nor hii regions .
this curious configuration prompted us to try to determine the distance to j133658.3295105 .
we detected h@xmath0 emission redshifted by @xmath1130kms@xmath2 with respect to an m83 hii region 25 e - se of the radio source .
we do not detect other redshifted emission lines of an optical counterpart down to @xmath3 .
two different scenarios are proposed : the radio source is at @xmath4 , a much larger distance than the previously proposed lower limit @xmath5 , or the object was ejected by a gravitational recoil event from the m83 nucleus .
this nucleus is undergoing a strong dynamical evolution judging from previous 3d spectroscopy . |
a central aim of nanoscale science is to controllably modify material properties on a nanometre scale .
one of its greatest achievements has been manipulation of surface atomic structure with the scanning tunnelling microscope ( stm ) @xcite , and some of the most striking demonstrations of this capability the construction of complex arrays of adatoms by the maneuvering and positioning of individual adatoms @xcite . synthesized nanoscale structures such as these , as well as other surface nanostructures including islands @xcite and step arrays @xcite act as resonators for mobile surface electrons , with characteristics that may be controlled by the adjustment of geometrical parameters . here
we present a detailed analysis of the electronic structure of a particular class of resonator , surface nanocavities constructed from mn atoms on ag(111 ) . of particular interest
is the extent to which a simple scattering model , which takes into account the atomic structure of the resonator , can successfully describe its spatial and spectral electron distributions .
recent work @xcite has shown that the specific characteristics of individual resonators impact upon the electronic structure of adsorbates positioned within them , and may open up the possibility of atomic scale control of magnetic or chemical properties .
in addition , we consider the use of these artificial nanoscale structures for probing quasiparticle lifetimes close to the fermi energy . the lifetimes of excitations enter the description of many important surface phenomena such as the dynamics of charge and energy transfer , and for excitations in surface bands reflect a complex combination of decay processes via both surface - localised and extended states , with the additional influence of screening that varies rapidly as the electron density decreases outside the surface .
successful prediction of lifetimes remains an on - going theoretical problem @xcite , whilst measurements must overcome considerable challenges ( reviewed in ref.@xcite ) so that only recently @xcite has a consistent account of the lifetime of a prototype quasiparticle , a hole in a surface band , been constructed , based upon tunnelling spectroscopy at the surface state band edge on noble metal ( 111 ) faces @xcite .
the spatial decay of electron interference patterns observed in stm images provides an alternative approach first indicated by avouris @xcite , and has been used to measure lifetimes well removed from the fermi energy @xmath2 @xcite .
however , lifetime measurements for states close in energy to @xmath2 are still lacking but are of particular interest because the relative contributions of phononic and electronic decay channels are expected to change significantly in this range .
we indicate how electron resonators may be used for this purpose .
we study nanocavities constructed using mn adatoms on the ag ( 111 ) surface in a custom - built ultra - high vacuum ( uhv ) stm , operating at a temperature of @xmath3 k @xcite .
the mn atoms were evaporated onto a cold ag substrate that had been prepared by standard procedures in uhv .
electrochemically etched w tips were prepared in uhv by sputtering and annealing .
differential conductivity ( @xmath0 ) spectra were recorded under open feedback loop conditions using a lock - in amplifier with a sinusoidal modulation added to the sample bias .
moreover , maps of @xmath0 were recorded simultaneously with constant current imaging without opening the feedback .
60 @xmath4 constant current topograph of mn adatoms on ag(111 ) . at @xmath5 512 mv and @xmath6 731 m@xmath7
the mn atoms occasionally exhibit tip - induced mobility ( arrow ) .
image scanned at 1.9 s / line .
( b ) @xmath8 spectrum acquired on top of a single mn atom .
abrupt current fluctuations occur at @xmath5 1.5 v and @xmath6 1 g@xmath7 .
spectrum recorded at 133 mv / sec , starting at @xmath9 v. ] manganese atoms on ag(111 ) image as @xmath10 high and @xmath11 wide protrusions ( full width at half maximum ) at a sample voltage @xmath12 v ( fig . [ atom_stability]a ) , similar to previous observations of adatom species on noble metal surfaces @xcite . most of the stm images presented here have been recorded at @xmath12 v and a tunnelling current @xmath13 na , i.e. a tunnelling resistance @xmath14 g@xmath7 ; these values are chosen because the tip - sample interaction is sufficiently strong for @xmath15 below @xmath16 m@xmath7 that atoms are occasionally moved during scans , as demonstrated in figure [ atom_stability]a . in @xmath17 spectroscopy of single mn atoms
we observe abrupt current changes at @xmath18 v and @xmath19 g@xmath7 , figure [ atom_stability]b , indicating the onset of induced atomic mobility .
to achieve acceptable signal - to - noise ratios in tunnelling spectroscopy elevated modulation amplitudes were required ( @xmath20 mv@xmath21 ) .
we construct artificial structures via controlled relocation of individual mn atoms , using a sliding process pioneered by eigler @xcite .
a mn adatom is moved by placing the tip over the atom at large tip - sample separation ( typically @xmath14 g@xmath7 ) , reducing this separation ( @xmath22 v , @xmath23 m@xmath7 ) and then moving the tip slowly across the surface ( a rate of @xmath24 one lattice constant per second ) to the desired location .
the adatom follows the tip motion , and is left in position by retracting the tip . to model the experiments it is important to take into account the constant - current mode of operation that is used in obtaining the @xmath0 maps .
the tunnelling current is calculated as @xcite @xmath25 where @xmath26 is the local density of states ( ldos ) of the sample at the position @xmath27 of the tip and where the transmission probability @xmath28 is given by @xmath29 \label{transmissionfactor}\ ] ] with tip - sample distance @xmath30 and workfunctions @xmath31 ev and @xmath32 ev of tip and sample respectively @xcite .
simulating the feedback loop , @xmath30 was adjusted for each lateral position ( @xmath33 ) to ensure constant current @xmath34 , which in turn was numerically differentiated to obtain @xmath0 . in using ( [ tunnelcurrent ] ) we assume an electronically featureless tip , which is supported by the absence of any tip dependence to be seen during the experiments .
for @xmath35 we consider an array of @xmath36 two - dimensional @xmath1-wave scatterers with coordinates @xmath37 , @xmath38 .
this is the model introduced by heller _
et al _ @xcite .
we determine @xmath35 from the green function @xmath39 which is given by multiple - scattering theory in terms of the free - electron green function @xcite @xmath40 where @xmath41 and the @xmath42-matrices that characterise the point scatterers : @xmath43 a number of parameters enter the calculations .
we use @xmath42-matrices corresponding to perfectly absorbing scatterers , as identified by heller _
et al_. @xcite and used by them to describe the @xmath0 spectrum of a 60 fe atom corral on cu(111 ) .
this boundary condition reflects the strong coupling to bulk states @xcite that occurs at the adatoms .
the effective mass @xmath44 @xmath45 and surface state band edge @xmath46 mev have been determined previously @xcite .
the positions of the atoms and the lateral position of the tip are crucial in determining the exact shape of @xmath0 spectra .
for example , only the highest symmetry eigenstates contribute at the centre of a circular structure , but slightly off centre or if the shape is not exactly regular a mixing of state occurs , giving additional structure .
uncertainty enters these coordinates through the piezo calibration and possible anisotropy of the piezo sensitivity in the fast ( @xmath47 ) and slow ( @xmath48 ) scanning directions .
therefore an optimisation scheme was used based upon a @xmath49 fit using the levenberg - marquardt method in which the relative coordinates were fixed but independently scaled in the @xmath47 and @xmath48 directions .
after the initial determination , the piezo calibration parameters were held fixed in all subsequent calculations . in a second stage ,
the calculations have been extended further to include an energy dependent lifetime of the electronic states ( sect .
[ qwertz ] ) .
this was accomplished via a parameterised imaginary self energy @xmath50 added to the energy @xmath51 in computing @xmath52 from equations ( [ freeg])-([t ] ) .
the importance of the self energy in determining the peak amplitudes and widths in the ldos has been noted previously @xcite .
here we exploit this dependence to extract @xmath53 from experimental data .
figure [ box_pics ] displays data from an array of @xmath54 mn adatoms .
images a h are constant current topographs covering a range of different voltages , recorded at values indicated in the figure .
the adatoms , which appear as bright protrusions in every case , have been arranged into a rectangular geometry with approximate dimensions of @xmath55 @xmath4 . inside this structure
a weak corrugation is found , originating from the confinement of the surface state electrons .
a nonlinear height scale has been used for display , in order to render this weak corrugation easily discernible ; this has the side effect that the adatoms appear rather blunt . at negative voltages ( i.e. when tunnelling from occupied levels in the sample ) the constant current topographs exhibit a rectangular pattern of four maxima within the confining array of mn atoms , which is still discernible at voltages well below the bottom of the surface state band at @xmath56 mev .
this pattern persists owing to the preferential tunnelling from occupied states at the fermi level . at positive voltages ,
imaging unoccupied states , the pattern becomes more complex and varies with @xmath57 , the most significant tunnelling now occuring at energies corresponding to the varying bias voltage .
images i p show @xmath0 maps that were recorded simultaneously with the topographic scans .
below @xmath58 , the adatoms cause clear features , whereas no corrugation is discernible inside the array .
in contrast , above @xmath58 a clear wave pattern arises within the structure , while the adatoms are only weak features which conveniently serve to indicate the rectangular boundary .
the evolution with energy of the internal pattern is as expected , representing a progression through the nodal structures of the lowest lying `` particle - in - a - box '' eigenstates of the rectangular cavity , @xmath59 where @xmath60 and @xmath61 are the linear dimensions of the structure . at low energy , fig.[box_pics]j , a single
maximum occurs ( @xmath62 ) .
next , two maxima appear along the longer axis of the rectangle ( fig .
[ box_pics]k ) , corresponding to the @xmath63 level .
the @xmath64 mv map ( fig.[box_pics]l ) showing four maxima , two along each axis of the rectangle , reflects contributions from both the second ( @xmath63 ) and third ( @xmath65 ) levels .
the states are in reality resonances that in general overlap in energy , so that @xmath0 maps can contain contributions from several . at higher energies ,
the pattern becomes increasingly complex .
calculated results are displayed in fig .
[ box_theo ] where panels a g are topographs , h
n are @xmath0 maps and o - u are ldos maps .
close inspection and comparison with figs.[box_pics]b - h , and [ box_pics]j - p reveals that the experimental features are reproduced well by the model calculations .
as pointed out previously @xcite , experimental @xmath0 data recorded by the methods used here are affected by the variations in tip - sample separation which give rise to the corrugation in the topographs . consequently @xmath0 maps are not directly comparable to ldos maps , a fact clearly visible by comparing figs .
[ box_theo]l and s , or [ box_theo]m and t , although the similarity between the two tends to increase with increasing bias voltages .
indeed , around zero bias it is topographs and not @xmath0 that are are expected to resemble the ldos , and this is clearly the case in fig .
[ box_theo ] . spectrum and calculated ldos of the 28 atom rectangle .
right : @xmath0 spectrum and theoretical ldos of a @xmath66 diameter ring of 35 mn adatoms . ]
a more detailed comparison between experiment and theory is possible in the case of @xmath0 spectra .
an example measured near the centre of the rectangular resonator is shown in the left panel of fig .
[ box_spex ] . tunneling from a @xmath1-state tip directly at the centre of an ideal rectangular structure
can only take place into surface state levels with odd quantum numbers @xmath67 .
this is the origin of the large peaks at @xmath68 mev and between @xmath69 and @xmath70 mev .
irregularities in the structure and deviations in position of the tip from the the exact centre introduce additional weak tunneling channels , e.g. near @xmath71 mev , whilst the intrinsic width of the confined levels further modifies the spectrum . comparing the measured spectrum with
the ldos calculated using the two - dimensional scattering model that includes the structural irregularities and correct tip position , displayed in the same panel in fig .
[ box_spex ] , we see the latter faithfully reproduces the differential conductivity spectrum .
note that the characteristic temperature associated with kondo behaviour in mn / ag alloys is @xmath72 mk @xcite and thus does not affect the line shape @xcite under the present experimental conditions .
we have also performed similar experiments and calculations on other geometries , such as adatom rings .
the right panel of figure [ box_spex ] contains a comparison between the differential conductivity spectrum , measured approximately @xmath20 off centre in a @xmath66 diameter ring formed from 35 mn adatoms , and the corresponding ldos obtained using the @xmath1-wave scattering model .
the spectrum contains a large number of peaks , the off - centre position enabling low symmetry eigenstates to provide channels for tunnelling , but these are reproduced in the theoretical ldos spectrum calculated using the precise atomic positions of the mn ring .
we conclude that the @xmath1-wave scattering model provides an adequate basis for modelling the spatial and spectral electron distributions in electron resonators .
maps of the same structure with ( d - f , middle row ) and without ( g - i , bottom row ) the central atom . ]
having characterised the properties of electron resonators we have next investigated the interaction with an additional atom introduced into the nanocavity region .
these structures were formed by maneuvering one of the boundary mn atoms into the central region , and then moving a mn atom elsewhere on the surface into the vacant position in the boundary , restoring the integrity of the resonator . in figure [ full_empty ]
we show @xmath0 maps of a rectangular structure made from 28 mn atoms and measured at three different voltages , contrasting the images obtained with and without a mn atom inside the structure . at first glance the presence of the central atom appears to barely affect the wave pattern inside the rectangle , apart from in the immediate vicinity of the centre atom .
there , the difference is striking . at voltages of @xmath73 mv and @xmath74 mv where @xmath0 is at a minimum in the empty structure ( figs .
[ full_empty]g and i ) , @xmath0 exhibits a peak in the centre upon insertion of the atom ( figs .
[ full_empty]d and f ) . at @xmath75 mv , however , the maximum observed in the empty rectangle ( fig.[full_empty]h ) changes to a minimum ( fig .
[ full_empty]e ) .
spectra with increasing lateral distance from the centre of a 16 mn atom ring radius @xmath76 ( a ) without and ( b ) with a central mn atom , and spectra calculated using the two - dimensional scattering model of the ldos for the same systems , ( c ) without and ( d ) with the central atom .
the vertical dashed lines mark the energies of the main spectral features present at the centre of the empty structure . ]
this `` contrast reversal '' has been previously reported and explained as a consequence of the strong hybridisation between states on the mn atom and substrate surface electron levels @xcite , and has been reproduced in calculations of the local density of states directly on mn atoms in resonators of various geometries using full three - dimensional multiple scattering calculations . here
we consider how this interaction extends away from the central mn atom and the degree to which the two - dimensional scattering model describes the behaviour .
[ abstands_kurve80 ] shows a series of @xmath0 spectra measured at various distances from the centre of an empty adatom ring , and when a mn atom has been introduced .
also shown are the ldos calculations for identical geometries . comparing @xmath0 spectra in figs .
[ abstands_kurve80]a and b at central positions closest to the location of the new adatom , it is clear that spectral peaks in the empty resonator coincide with troughs of the occupied structure ( vertical dashed lines ) .
this is evidence for strong hybridisation @xcite . moving laterally away from this central position the characteristic peak structure of the empty resonator returns , with the exception of the lowest peak near @xmath77 mv which remains absent even at large displacements .
this prominent feature is associated with a nodeless resonator eigenstate that has its maximum at the centre of the ring , and so is particularly sensitive to the introduction of a mn atom into the central region .
the two right - most panels in fig .
[ abstands_kurve80 ] show the ldos calculated for this system using the two - dimensional scattering model , at lateral positions at which @xmath0 spectra were recorded .
the prominent features in the experimental spectra are reproduced well in the theoretical curves , again showing that the simple scattering model , using the exact atomic locations of the mn atoms making up the structure , gives an accurate and valid account of the electronic structure of resonators .
the calculations also confirm that the dominant state in the empty resonator is absent in the spectrum of states of the occupied resonator , so that the insensitivity of @xmath0 maps to the presence of the central atom away from its immediate vicinity that was suggested by fig .
[ full_empty ] clearly does not extend to all sample voltages .
note that in fig .
[ abstands_kurve80]d we have omitted theoretical spectra within @xmath78 of the central mn atom . within this radius ( which is much greater than the topographic radius of the mn atoms )
the differential conductivity also has large contributions from channels involving bulk states scattered by the adatom @xcite , and these are not included in the model .
also omitted in the model is any spatial information about the adatom orbitals that would also be required to model @xmath0 spectra ( and maps ) in its immediate vicinity .
the basic idea of our approach to lifetime measurements is straightforward @xcite .
exploiting the properties of the electron resonators identified in the previous sections we construct nanocavities with electronic features in the vicinity of @xmath2 .
the lineshape of the spectral features are affected by lifetime - limiting processes so that through analysis the lifetime @xmath79 can be determined . by changing the geometry of nanoscale structures it is possible to tune the eigenstate energies of the structures , and thus generate states at various energies in order to obtain the energy dependence of the lifetime .
the two - dimensional scattering model validated above enables the design of suitable structures .
we note that a fully quantitative analysis will also require information on the energy dependence of the tip density of states , which is not expected to be constant over a wide energy range , and likewise the bulk density of states of the crystal which causes a background signal in @xmath0 spectra . here
, we seek to demonstrate the concept and so neglect these influences .
following li et al .
@xcite we describe the lifetime effects in terms of an effective imaginary part @xmath53 of the self energy , where @xmath80 . of the self energy on the ldos of a 36 atom circle with 210 diameter .
( a ) calculated ldos at the centre for various @xmath53 indicated in the figure .
variation of peak ( b ) heights and ( c ) widths with @xmath53 . ] the effect of @xmath53 on the ldos of an artificial nanoscale resonator is shown in fig .
[ ei_theory]a for a 210 diameter ring .
the calculation for vanishing @xmath53 , i.e.infinite lifetime , yields a series of sharp resonances ( fig.[ei_theory]a , top curve ) .
note that the spectrum corresponds to the centre of the structure .
the high symmetry gives the `` cleanest '' spectrum , avoiding difficulties that exist at off - centre positions where resonances can overlap .
with increasing @xmath53 there is a reduction of the peak heights and a concomitant broadening , especially apparent for the lowest of the states which have the sharpest intrinsic profile .
these trends are quantified by fitting a series of lorentzians to the data , and figure [ ei_theory]b and c display the widths and height extracted through this analysis . as @xmath53 increases , the peak heights decrease monotonically , the lowest ( in energy ) peaks most rapidly , whilst there is an essentially linear dependence of the peak width on @xmath53 .
extracted from analysis of @xmath0 spectra of a 115 radius 35 atom circular resonator ( diamonds , squares ) and a 40 radius 16 atom circular resonator .
the @xmath53 values are plotted at the position of the respective spectral features .
the inset shows the equivalent lifetimes @xmath81 of hot holes and electrons . ]
figure [ ei_final ] shows values of @xmath82 obtained from an analysis of measurements from two nanostructures . for each
, @xmath49 fits were made over limited energy ranges ( @xmath83 mev ) adjusting coefficients of a linear parameterisation of @xmath53 .
we estimate the error of @xmath53 thus determined to be of the order of @xmath84 % .
the diamonds and squares are derived from @xmath0 spectra for a 35 atom circular resonator , and the filled circles were obtained from a 16 atom circle . below the fermi energy ( @xmath85 )
@xmath53 approaches a value obtained previously from spectra taken on clean terraces which yielded @xmath86 mev @xcite .
there is a clear increase of @xmath53 at large energies , and we find @xmath87 ev@xmath88 mev , which is comparable with an extrapolation down to this energy of data for the ag(111 ) surface state extracted from the spatial coherence of interference patterns at energies @xmath89 ev by brgi _
finally , at energies close to @xmath2 we observe a marked decrease of @xmath53 which corresponds to an increased lifetime of the corresponding quasiparticles ( inset in fig .
[ ei_final ] ) .
increased lifetimes are indeed expected as both the phononic and electronic decay channels become less efficient .
while further data close to @xmath2 is necessary to quantitatively characterize this effect the first results are encouraging .
future work will also address the relation of the self - energies determined from atomic arrays to that of an unperturbed surface .
we note that data from resonators of different diametres ( 40 and 115 ) yield similar @xmath82 .
this similarity appears to indicate that the influence of the array may be rather small .
using scanning tunnelling spectroscopy and spectroscopic imaging at low temperature we have investigated artificial mn atom arrays on ag(111 ) .
these act as electron resonators and we have demonstrated the degree to which variations in geometry may be used to control spatial and spectral distributions of confined surface state electrons .
this control may be exploited in manipulating the local density of states of adsorbates introduced into the resonator , or as we have demonstrated , to generate spectral features at specific energies in order to perform lineshape analyses from which quasiparticle lifetimes may be deduced .
extensive comparison of @xmath0 maps and spectra indicated the utility of a model based upon two - dimensional @xmath1-wave scatterers for describing and predicting the characteristics of specific resonators .
this work has been supported the dfg , the daad and the uk epsrc . | the electronic structure of artificial mn atom arrays on ag(111 ) is characterized in detail with scanning tunnelling spectroscopy and spectroscopic imaging at low temperature .
we demonstrate the degree to which variations in geometry may be used to control spatial and spectral distributions of surface state electrons confined within the arrays , how these are influenced by atoms placed within the structure and how the ability to induce spectral features at specific energies may be exploited through lineshape analyses to deduce quasiparticle lifetimes near the fermi level . through extensive comparison of @xmath0 maps and spectra
we demonstrate the utility of a model based upon two - dimensional @xmath1-wave scatterers for describing and predicting the characteristics of specific resonators . |
this article deals with the construction of magnetic monopoles , algebraic curves subject to certain transcendental constraints and number theory .
magnetic monopoles , or the topological soliton solutions of yang - mills - higgs gauge theories in three space dimensions , have been objects of fascination for over a quarter of a century .
bps monopoles in particular have been the focus of much research ( see @xcite for a recent review ) .
these monopoles arise as a limit in which the higgs potential is removed and satisfy a first order bogomolny equation @xmath1 ( together with certain boundary conditions , the remnant of the higgs potential ) . here
@xmath2 is the field strength associated to a gauge field @xmath3 , and @xmath4 is the higgs field .
these equations may be viewed as a dimensional reduction of the four dimensional self - dual equations upon setting all functions independent of @xmath5 and identifying @xmath6 .
just as ward s twistor transform relates instanton solutions in @xmath7 to certain holomorphic vector bundles over the twistor space @xmath8 , hitchin showed @xcite that the dimensional reduction leading to bps monopoles could be made at the twistor level as well .
mini - twistor space is a two dimensional complex manifold isomorphic to t@xmath9 , and bps monopoles may be identified with certain bundles over this space . in particular
a curve @xmath10 t@xmath9 , the spectral curve , arises in this construction and , subject to certain nonsingularity conditions , hitchin was able to prove all monopoles could be obtained by this approach @xcite .
nahm also gave a transform of the adhm instanton construction to produce bps monopoles @xcite .
the resulting nahm s equations have lax form and the corresponding spectral curve is again @xmath11 .
these curves @xmath11 are the curves of the title .
hitchin s construction ( as we shall soon recall ) places various transcendental constraints upon them and the outstanding and difficult problem is to construct curves satisfying these constraints .
now , given an appropriate curve , ercolani and sinha showed in their seminal paper @xcite how one could solve ( a gauge transform of ) the nahm equations in terms of a baker - akhiezer function for the curve @xmath11 .
the authors have extended this work @xcite and , given a curve , one can solve both the nahm data and reconstruct the desired monopole in terms of standard integrable systems constructions .
our purpose here is not to focus on this integrable system side of the story but on the curves @xmath11 and we shall only cite various formulae as needed to illustrate the type of detail needed of a curve in order to implement the construction . the number theory ( and in particular ` ramanujan ' of the title ) appear when we try to implement the transcendental constraints on the curve and construct the required quantities associated with the curve .
this is the story we now recount .
an outline of our article is as follows . in section 2
we recall the relevant aspects of the hitchin and ercolani - sinha constructions .
this recounts the constraints of hitchin and we highlight the ingredients needed to make effective the integrable system construction and show how these reduce to evaluating quantities intrinsic to the curve .
the first hurdle in implementing the construction is to analytically determine the period matrix for @xmath11 and then understand the theta divisor . in section 3
we will introduce a class of ( genus 4 ) curves for which we can do this .
they are of the form @xmath12 where @xmath13 , @xmath14 are distinct complex numbers .
( for appropriate @xmath13 this yields a charge 3 monopole . )
this class of curves was studied by wellstein over one hundred years ago @xcite and more recently by matsumoto @xcite . here
we will introduce our homology basis and define branch points in terms of @xmath15-constants following @xcite .
corresponding to ( some of ) hitchin s nonsingularity conditions ercolani and sinha obtain restrictions on the allowed period matrices for the spectral curve .
equivalent formulations of these conditions were given in @xcite .
the ercolani - sinha conditions are transcendental constraints and to solve these is the next ( perhaps _ the _ ) major hurdle to overcome in the construction . in section 4
we do this for our curves . at this stage
we have replaced the constraints by relations between various hypergeometric integrals . to simplify matters for the present paper we next demand more symmetry and consider in section 5 the genus 4 curves @xmath16 where @xmath17 is a certain real parameter .
this restriction has the effect of reducing the number of hypergeometric integrals to be calculated to two .
interestingly the relations we demand of these integrals are assertions of ramanujan only recently proven .
we will denote curves of the form ( [ bren03 ] ) as _ symmetric monopole _ curves ( though in fact they may not satisfy all of hitchin s nonsingularity conditions ) .
the tetrahedrally symmetric charge 3 monopole is of this form .
the curve ( [ bren03 ] ) covers a hyperelliptic curve of genus two and two elliptic curves .
we discuss these coverings . using weierstrass - poincar reduction theory we are able to express the theta function behaviour of these symmetric monopoles in terms of elliptic functions and fairly comprehensive results
may be obtained .
finally , in section 6 , we shall consider the curve ( [ bren03 ] ) associated with tetrahedrally symmetric 3-monopole when the above parameter @xmath18 .
this genus 4 curve covers 4 elliptic curves and all entries to the period matrices are expressible in terms of elliptic moduli .
the analytical means which we are using for our analysis involve thomae - type formulae , weierstrass - poincar reduction theory , multivariable hypergeometric function and higher hypergeometric equalities of goursat .
our conclusions in section 7 will highlight various of our results .
_ this article is based upon the second part of the preprint @xcite .
new theoretical results and the unwieldy length of that paper have led us to separate general constructions from the exploration of the transcendental constraints upon which we now report . _
in this section we shall recount the transcendental constraints placed upon the spectral curve coming from hitchin s construction .
we shall then describe those quantities related to the curve needed in the ercolani - sinha construction and its extensions .
using twistor methods hitchin @xcite has shown that each static @xmath0 yang - mills - higgs monopole in the bps limit with magnetic charge @xmath19 is equivalent to a spectral curve of a restricted form .
if @xmath20 is the inhomogeneous coordinate on the riemann sphere , and @xmath21 are the standard local coordinates on @xmath22 ( defined by @xmath23 ) , the spectral curve is an algebraic curve @xmath24 which has the form @xmath25 here @xmath26 ( for @xmath27 ) is a polynomial in @xmath20 of maximum degree @xmath28 .
the hitchin data constrains the curve @xmath11 explicitly in terms of the polynomial @xmath29 and implicitly in terms of the behaviour of various line bundles on @xmath11 .
if the homogeneous coordinates of @xmath30 are @xmath31 $ ] we consider the standard covering of this by the open sets @xmath32\,|\,\zeta_0\ne0\}$ ] and @xmath33\,|\,\zeta_1\ne0\}$ ] , with @xmath34 the usual coordinate on @xmath35 .
we will denote by @xmath36 the pre - images of these sets under the projection map @xmath37 .
let @xmath38 denote the holomorphic line bundle on @xmath22 defined by the transition function @xmath39 on @xmath40 , and let @xmath41 be similarly defined in terms of the transition function @xmath42 .
a holomorphic section of such line bundles is given in terms of holomorphic functions @xmath43 on @xmath44 satisfying @xmath45 .
we denote line bundles on @xmath11 in the same way , where now we have holomorphic functions @xmath43 defined on @xmath46 .
the hitchin data constrains the curve to satisfy : + * h1*. reality conditions
@xmath47 * h2*. @xmath48 is trivial on @xmath11 and @xmath49 is real .
+ * h3*. @xmath50 for @xmath51 .
+ the reality conditions express the requirement that @xmath11 is real with respect to the standard real structure on @xmath22 @xmath52 this is the anti - holomorphic involution defined by reversing the orientation of the lines in @xmath53 .
a consequence of the reality condition is that we may parameterise @xmath26 as follows , @xmath54 \prod_{k=1}\sp{r}(\zeta-\alpha_k)(\zeta+\frac{1}{\overline{\alpha}_k}),\qquad \alpha_r\in \mathbb{c},\ \chi_r\in\mathbb{r}.\ ] ] thus each @xmath26 contributes @xmath55 ( real ) parameters . certainly this constraint may be readily implemented .
the remaining two constraints are however transcendental .
the triviality of @xmath48 on @xmath11 means that there exists a nowhere - vanishing holomorphic section . in terms of our open sets
@xmath36 we then have two , nowhere - vanishing holomorphic functions , @xmath56 on @xmath57 and @xmath58 on @xmath59 , such that on @xmath60 @xmath61 ercolani and sinha utilize this to give an alternate characterisation we will present in due course . for a generic @xmath19-monopole
the spectral curve is irreducible and has genus @xmath62 .
this may be calculated as follows . for fixed @xmath20 the @xmath19 roots of @xmath63 yield an @xmath19-fold covering of the riemann sphere .
the branch points of this covering are given by @xmath64 this expression is of degree @xmath65 in @xmath20 and so by the riemann - hurwitz theorem we have that @xmath66 whence the genus as stated
. the @xmath67 monopole spectral curve is given by @xmath68 where @xmath69 is any point in @xmath53 .
in general the three independent real coefficients of @xmath70 may be interpreted as the centre of the monopole in @xmath71 .
strongly centred monopoles have the origin as center and hence @xmath72 .
the group @xmath73 of rotations of @xmath71 induces an action on @xmath22 via the corresponding @xmath74 transformations . if @xmath75 the transformation on @xmath22 given by @xmath76 corresponds to a rotation by @xmath15 around @xmath77 , where @xmath78 , @xmath79 , @xmath80 , and @xmath81 .
( here the @xmath82 transformation is given by the derivative of the @xmath20 transformation . )
the @xmath73 action commutes with the real structure @xmath83 .
although a general mbius transformation does not change the period matrix of a curve @xmath11 only the subgroup @xmath84 preserves the desired reality properties .
we have that @xmath85\sp{1/2},\ ] ] and @xmath86 \prod_{k=1}\sp{r}(\zeta-{\tilde\alpha}_k)(\zeta+\frac{1}{\overline{\tilde\alpha}_k}).\ ] ] in particular the form of the curve does not change : that is , if @xmath87 then so also @xmath88 .
it is perhaps worth emphasising that the reality conditions are an extrinsic feature of the curve ( encoding the space - time aspect of the problem ) whereas the intrinsic properties of the curve are invariant under birational transformations or the full mbius group .
such extrinsic aspects are not a part of the usual integrable system story . in order to implement the reconstruction of the nahm data and monopole associated to a curve @xmath11 various other quantities
are also required .
we shall now describe these and in so doing express ercolani and sinhas s alternate characterisation of one of hitchin s transcendental constraints .
the spectral curve ( [ spectcurve ] ) is an @xmath19-sheeted cover of @xmath9 . by a rotation
if necessary it is always possible to choose @xmath19 distinct preimages @xmath89 of @xmath90 and consequently the roots of @xmath91 near @xmath90 behave as @xmath92 where the @xmath93 may be assumed distinct . as a consequence
we see that at @xmath94 we have @xmath95 where @xmath96 is a local coordinate . also from ( [ arpar ] )
we have that at @xmath97 @xmath98 in view of ( [ infj ] ) there exists a differential @xmath99 , a sum of differentials of the second kind , such that @xmath100 here @xmath101 are a canonical homology basis for @xmath11 . on the integrable systems side of the story
this differential yields the flow on the jacobian corresponding to the time evolution of the system .
this direction is given by @xmath102 in practice one does not need @xmath103 explicitly but ( [ wvec ] ) and @xmath104.\label{etacondition}\ ] ] at this stage we have not imposed the hitchin constraints h2 , h3 on our curve @xmath11 .
ercolani and sinha now make the following connection between ( [ wvec ] ) and h2 . consider the logarithmic derivative of ( [ triv3 ] ) representing the triviality of @xmath48 on @xmath11 , @xmath105 ( hurtubise considered a similar construction in the @xmath106 case @xcite . ) now in order to avoid essential singularities in @xmath107 we have from ( [ rhodef ] , [ rhobdef ] ) that @xmath108 because @xmath109 is a function on @xmath110 , then @xmath111 for all cycles @xmath112 from @xmath113 .
a similar result follows for @xmath58 and upon noting ( [ derivative ] ) we may define @xmath114 further , in view of ( [ gaminf ] ) and ( [ foex ] ) , we may write @xmath115 where @xmath116 are canonically @xmath117-normalized holomorphic differentials . integrating @xmath99 around @xmath118-cycles leads to the ercolani - sinha constraints @xmath119 which are necessary and sufficient conditions for @xmath120 to be trivial when restricted to @xmath121 . thus @xmath122 therefore the vector @xmath123 , the period lattice for the curve @xmath11 , and so the winding - vector " vector @xmath124 is a half - period .
note that @xmath125 or otherwise @xmath99 would be holomorphic contrary to our choice . using the bilinear relations we have that , for any holomorphic differential @xmath126 , @xmath127 houghton , manton and ramo utilise this expression to express a dual form of the ercolani - sinha constraints ( [ esc1 ] ) .
define the 1-cycle @xmath128 then ( upon recalling that @xmath129 , where @xmath83 is the period matrix ) we have the equivalent constraint : @xmath130 the right - hand side of this equation is readily evaluated .
we may express an arbitrary holomorphic differential @xmath126 as , @xmath131 where @xmath132 is a polynomial of degree at most @xmath133 in @xmath20 . thus using ( [ rhodef ] )
we obtain @xmath134 upon using lagrange interpolation . at this stage
we have from the condition h2 , the following are equivalent : 1 .
@xmath120 is trivial on @xmath11 .
there exists a 1-cycle @xmath135 such that for every holomorphic differential @xmath126 ( [ holdefes ] ) , @xmath136 3 .
@xmath137 @xmath138 here ( 2 ) is the dual form of the ercolani - sinha constraints given by houghton , manton and ramo .
their 1-cycle generalises a similar constraint arising in the work of corrigan and goddard @xcite .
the only difference between ( 3 ) and that of ercolani - sinha theorem ii.2 is in the form of @xmath139 in which we disagree .
we also know that @xmath140 .
the ercolani - sinha constraints impose @xmath141 conditions on the period matrix of our curve .
we have seen that the coefficients @xmath26 each give @xmath55 ( real ) parameters , thus the moduli space of charge @xmath19 centred @xmath0 monopoles is @xmath142 ( real ) dimensional .
the 1-cycle appearing in the work of houghton , manton and ramo further satisfies [ hmrinvc ] @xmath143 .
this result is the dual of hitchin s remark ( * ? ? ?
* p164 ) that the triviality of @xmath120 together with the antiholomorphic isomorphism @xmath144 yields an imaginary lattice point with respect to @xmath145 .
the final constraint of hitchin and various quantities needed for our purposes are encoded in the expressions @xmath146 which we shall now unpack .
here @xmath147 are theta functions with characteristics , @xmath148 is the abel map , @xmath149 , and @xmath150 .
( our conventions for theta functions are given in the appendix . )
the vector @xmath151 plays a special role in the monopole construction .
it is defined by @xmath152 where @xmath153 is the vector of riemann constants ( our conventions regarding this are also given in the appendix ) .
we have that 1 .
@xmath151 is independent of the choice of base point of the abel map ; 2 .
@xmath154 ; 3 .
@xmath155 ; 4 .
for @xmath156 we have @xmath157 .
the point @xmath151 is the distinguished point hitchin uses to identify degree @xmath158 line bundles with @xmath159 .
the proof of these properties together with the following lemma further constraining the ercolani - sinha vector may be found in @xcite : [ ueventheta]@xmath160 is a non - singular even theta characteristic .
the expressions ( [ newbafnch ] ) are the components of a baker - akhiezer function ; they are sections of a line bundle @xmath161 .
the @xmath162 which we have left unspecified form a basis of the holomorphic sections of @xmath49 .
( they can be explicitly described in terms of theta functions and the quantities introduced already , but we do not need this for our present purposes . ) now the full condition h3 is that @xmath163 for @xmath149 .
this constraint must be checked using knowledge of the @xmath164 divisor .
the exact sequence @xmath165 given by multiplication by a section of @xmath166 does however give us the necessary condition @xmath167 if @xmath168 were trivial we would have a section , contradicting this vanishing result .
the same treatment given to the triviality of @xmath120 shows that if @xmath168 were trivial then @xmath169 .
therefore ( [ h0lz ] ) shows that @xmath170 for @xmath171 . thus @xmath172 is a primitive vector in @xmath173 and we obtain the final part of the ercolani - sinha constraints , @xmath174 the ercolani - sinha constraints ( [ escond ] ) or ( [ hmrescond ] ) place @xmath141 transcendental constraints on the spectral curve @xmath175 and a major difficulty in implementing this construction has been in solving these , even in simple examples .
the case of @xmath106 has been treated by several authors and going beyond these results we consider the case of @xmath176 in the remaining portion of this work . before doing this
let us record the remaining expressions needed for the integrable system reconstruction which are to be found in an object @xmath177 used by ercolani - sinha to reconstruct the nahm data .
the work @xcite gives the matrix @xmath177 ( which has poles of first order at @xmath178 ) may be written @xmath179\boldsymbol{u } - \widetilde{\boldsymbol{k}})}{\theta ( [ z+1]\boldsymbol{u } - \widetilde{\boldsymbol{k}})}\,e^{z(\nu_{l } - \nu_{j})}.\ ] ] here @xmath180 is the schottky - klein prime form , @xmath181 ( @xmath182 , @xmath183 ) is a non - singular even theta characteristic , and @xmath184 is determined ( for @xmath185 ) by @xmath186 .
the @xmath187 signs @xmath188 are arbitrary . in passing we note that a formula with similar features was obtained by dubrovin @xcite when giving a theta - functional solution to the euler equation describing motion of the @xmath19-dimensional rigid body .
( dubrovin s curve was of course different and was free of hitchin s constraints . )
this theorem encodes the data needed to reconstruct the nahm data and monopoles given a suitable spectral curve .
the whole construction is predicated on the theta functions built from the spectral curve .
thus we need 1 . to construct the period matrix @xmath83 associated to @xmath11 .
2 . to determine the half - period @xmath151 .
3 . to determine the ercolani - sinha vector @xmath139 .
4 . for normalised holomorphic differentials
@xmath189 to calculate @xmath190 . 5 . to determine @xmath191 .
6 . to determine @xmath103 and @xmath192
we shall now introduce the class of curves that will be the focus of our attention .
these are @xmath193 for suitable @xmath13 they correspond and @xmath194 $ ] . ] to centred charge three monopoles restricted by @xmath195 .
thus the eight dimensional moduli space of centred monopoles has been reduced to three dimensions .
the asymptotic behaviour of the curve gives us @xmath196 for notational convenience we will study ( [ cubic ] ) in the form ( @xmath197 , @xmath198 ) @xmath199 the moduli space of such curves with an homology marking can be regarded as the configuration space of six distinct points on @xmath9 .
this class of curves has been studied by picard @xcite , wellstein @xcite , shiga @xcite and more recently by matsumoto @xcite ; we shall recall some of their results . to make concrete
the @xmath15-functions arising in the ercolani - sinha construction we need to have the period matrix for the curve , the vector of riemann constants , and to understand the special divisors .
we shall now make these things explicit , beginning first with our choice of homology basis .
let @xmath11 denote the curve ( [ curvegena ] ) of genus four where the six points @xmath200 are assumed distinct and ordered according to the rule @xmath201 .
let @xmath202 be the automorphism of @xmath11 defined by @xmath203 the bilinear transformation @xmath204 @xmath205 and its inverse @xmath206 leads to the following normalization of the curve ( [ curvegen ] ) @xmath207 where @xmath208 fix the following lexicographical ordering of independent canonical holomorphic differentials of @xmath11 , @xmath209 to construct the symplectic basis @xmath210 of @xmath211 we introduce oriented paths @xmath212 going from @xmath213 to @xmath214 in the @xmath215-th sheet .
define 1-cycles @xmath216 on @xmath11 as follows @xmath217 the @xmath117-cycles of the homology basis are given in figure 1 , with the @xmath118-cycles shifted by one sheet .
we have the pairings @xmath218 , @xmath219 and therefore @xmath220 is a symplectic basis of @xmath221 . in the homology basis introduced
we have @xmath222 as @xmath223 for any cycle @xmath224 we have , for example , that @xmath225 for @xmath226 and @xmath227 , so completing the @xmath202 action on the homology basis . and
@xmath228 , width=226 ] and @xmath228 , width=113 ] denote vectors @xmath229 crucial for us is the fact that the symmetry ( [ curvesym ] ) allows us to relate the matrices of @xmath117 and @xmath118-periods . for any contour @xmath230 and one form
@xmath231 we have that @xmath232 .
if @xmath233 then , for example , @xmath234 leading to @xmath235 we find that @xmath236 where @xmath237 and @xmath238 .
this relationship between the @xmath117 and @xmath118-periods leads to various simplifications of the riemann identities , @xmath239 for @xmath240 and @xmath241 we obtain ( respectively ) that @xmath242 relations we shall employ throughout the paper . given @xmath243 and @xmath244 we now construct the riemann period matrix which belongs to the siegel upper half - space @xmath245 of degree 4 .
if one works with canonically @xmath117-normalized differentials the period matrix ( in our conventions ) is @xmath246 while for canonically @xmath118-normalized differentials it is @xmath247 .
clearly @xmath248 and we shall simply denote the period matrix by @xmath83 if neither normalization is necessary .
[ matsumoto1 ] let @xmath11 be the triple covering of @xmath249 with six distinct point @xmath250 , @xmath251 then the riemann period matrix is of the form @xmath252 where @xmath237 .
then @xmath253 is positive definite if and only if @xmath254 both wellstein and matsumoto give broadly similar proofs of ( [ taumat ] ) and we shall present another variant as we need to use an identity established in the proof later in the text . from ( [ brr ] )
we see that we have @xmath255 we know that @xmath243 is nonsingular and consequently @xmath256 and @xmath257 .
now @xmath258 which gives @xmath259 now from ( [ calba ] ) we see that @xmath260 from ( [ invcala ] ) we obtain @xmath261 and therefore @xmath262 finally one sees that @xmath263\left [ \rho h-\frac{(\rho-\rho\sp2)}{\delta } \boldsymbol{x}\boldsymbol{x}\sp{t}\right]=1,\ ] ] whence the result ( [ taumat ] ) follows for @xmath247 .
the remaining constraint arises by requiring @xmath264 to be positive definite .
we note that ( [ condition1 ] ) ensures that both @xmath256 and @xmath257 .
the branch points can be expressed in terms of @xmath15-constants . following matsumoto
@xcite we introduce the set of characteristics @xmath265 and denote @xmath266 ( see appendix a for out theta function conventions ) .
the characteristics ( [ characteristics ] ) are classified in @xcite by the representations of the braid group .
further , the period matrix determines the branch points as follows .
[ diez 1991 , matsumoto 2000 ] [ matsumoto2 ] let @xmath253 be the period matrix of ( [ curvegenb ] ) given in proposition [ matsumoto1 ] .
then @xmath267 these results have the following significance for our construction of monopoles .
first we observe that the period matrix is invariant under @xmath268 .
thus to our surface we may associate a point @xmath269\in \mathbb{b}\sp3=\{\mathbf{x}\in\mathbb{p}\sp3\,|\ , \boldsymbol{x}^t h \boldsymbol{x}<0\}\subset \mathbb{p}\sp3 $ ] and from this point we may obtain the normalized curve ( [ curvegenb ] ) .
it is known that a dense open subset of @xmath270 arises in this way from curves with distinct roots with the complement corresponding to curves with multiple roots .
correspondingly , if we choose a point @xmath269\in \mathbb{b}\sp3 $ ] we may construct a period matrix and corresponding normalized curve .
we note that with @xmath271 then @xmath272 and so we obtain @xmath273 and as a consequence corollary ( [ hmrinvc ] ) of houghton , manton and romo , @xmath143 .
more generally , let us write for an arbitrary cycle @xmath274 then the equality @xmath275 leads to the equation @xmath276 we have then that the matrix @xmath277 representing the involution @xmath83 on homology and ercolani - sinha vector satisfy @xmath278 a calculation employing the algorithm of tretkoff and tretkoff @xcite to describe the homology basis generators and relations , together with some analytic continuation of the paths associated to our chosen homology cycles ( with the sheet conventions described later in the text ) , yields that for our curve @xmath279 the matrix @xmath277 is not symplectic but satisfies @xmath280 where @xmath281 is the standard symplectic form .
( the minus sign appears here because of the reversal of orientation under the antiholomorphic involution . )
we shall now describe the vector @xmath151 and various related results , including the quantity @xmath284 . first let us record some elementary facts about our curve .
for ease in defining various divisors of the curve ( [ curvegena ] ) let @xmath285 be the three points over infinity and @xmath286 ( @xmath287 ) be the branch points
. then @xmath288 consideration of the function @xmath289 shows that @xmath290 .
the order of vanishing of the differentials @xmath291 , @xmath292 , @xmath293 and @xmath293 at the point @xmath294 are found to be @xmath295 , @xmath296 , @xmath297 and @xmath298 respectively , which means that the gap sequence at @xmath294 is @xmath296 , @xmath299 , @xmath300 and @xmath301 . from this
we deduce that the index of speciality of the divisor @xmath302 is @xmath303 .
because the genus four curve @xmath11 has the function @xmath304 of degree @xmath297 then @xmath11 is not hyperelliptic .
the function @xmath305 has divisor @xmath306 , with @xmath307 and @xmath308 such that @xmath309 is canonical .
this means that any other function of degree @xmath297 on @xmath11 is a fractional linear transformation of @xmath304 and that @xmath310 consists of precisely one point which is of order @xmath299 in @xmath159 ( * ? ? ?
* iii.8.7 , vii.1.6 ) .
the vector of riemann constants @xmath311 is a point of order @xmath299 in @xmath159 because @xmath312 is canonical ( * ? ? ?
* vi.3.6 ) .
let us fix @xmath313 to be our base point .
then as @xmath314 we have that @xmath315 . because @xmath316 we may identify @xmath317 as the unique point in @xmath310 .
we may further identify @xmath317 as the unique even theta characteristic belonging to @xmath164 .
with @xmath313 as our base point @xmath318 corresponds to the image under the abel map of the divisor of the function @xmath319 , and so vanishes ( modulo the period lattice ) .
thus for our curve @xmath320 is the unique even theta characteristic .
the point @xmath317 may be constructed several ways : directly , using the formula ( [ vecr ] ) of the appendix ( the evaluation of the integrals of normalised holomorphic differentials between branch points is described in appendix b ) ; by enumeration we may find which of the @xmath321 even theta characteristics @xmath322 $ ] leads to the vanishing of @xmath323(z;\tau)$ ] ; using a monodromy argument of matsumoto @xcite .
one finds that the relevant half period is @xmath324 $ ] .
the analysis of the previous paragraph , together with lemma [ ueventheta ] , tells us that @xmath139 must also be an even theta characteristic . again using that @xmath325 we have that @xmath326 ( with @xmath327 , @xmath328 , @xmath215 distinct ) and so @xmath329 . one sees from the above divisors ( in particular @xmath330 ) that @xmath331 .
thus @xmath332 and @xmath333 have order of vanishing differing by one for ( generic ) @xmath334 . from the results of the previous section we see that @xmath336 this has precisely the same divisor of poles as @xmath99 and we will use this to represent @xmath99 .
it is convenient to introduce the ( meromorphic ) differential @xmath337 the factor of three here being introduced to give the pairing @xmath338 we may therefore write @xmath339 the constants @xmath340 are found from the condition of normalisation @xmath341 where we have defined the vector of @xmath117-periods @xmath342 .
the vector of @xmath118-periods of @xmath343 is found to be @xmath344 .
the pairing with @xmath345 then yields the legendre relation @xmath346 now the @xmath118-periods of the differential @xmath347 give the ercolani - sinha vector . using ( [ gpdr ] )
we then obtain the equality @xmath348 finally , using ( [ gpdr ] ) , we may write @xmath349.\ ] ]
although we have now described how to evaluate most of the quantities needed for the reconstruction of a monopole , at least up to the evaluation of the integrals @xmath350 , we have not described how to solve the ercolani - sinha constraints for the spectral curve ( [ cubic ] ) which encode one of hitchin s transcendental constraints . to this
we now turn .
as we shall see , this reduces to constraints just on the four periods @xmath350 .
later we shall restrict attention to the curves ( [ bren03 ] ) , which has the effect of reducing the number of integrals to be evaluated to two and consequently simplifies our present analysis .
we shall work with the ercolani - sinha constraints in the form ( [ hmrescond ] ) .
let the holomorphic differentials be ordered as in ( [ diffbasis ] ) . then there exist two integer 4-vectors @xmath351 and values of the parameters @xmath352 and @xmath353 such that @xmath354 here @xmath355 depends on normalizations .
for us this will be @xmath356 to see this observe that ( [ hmrescond ] ) requires that @xmath357 in the parameterisation ( [ curvegena ] ) we are using we have that @xmath358 we wish @xmath359 with the value of @xmath355 stated .
consideration of the other differentials then yields ( [ esconda ] ) , transcendental constraints on the curve @xmath11 .
these constraints may be solved using the following result .
[ ourh2 ] the ercolani - sinha constraints ( [ esconda ] ) are satisfied for the curve ( [ cubic ] ) if and only if @xmath360 } = \frac{6 \hat\chi^{\frac{1}{3}}}{[\boldsymbol{n}.h \boldsymbol{n}-\boldsymbol{m}.\boldsymbol{n}+\boldsymbol{m}.h \boldsymbol{m}]}.\label{expxxx}\end{aligned}\ ] ] rewriting ( [ esconda ] ) we have that @xmath361 therefore @xmath362 upon using that the period matrix is symmetric and our earlier expression for @xmath363 .
rearranging now gives us that @xmath364 and so we have established ( [ expxx ] ) where @xmath365 there are several constraints .
first , the ercolani - sinha condition ( [ esonx ] ) is that @xmath366 \xi[h\boldsymbol{n } + \rho^{2}\boldsymbol{m } ] = \nu\ ] ] and consequently @xmath367\xi = \nu = 6\hat\chi^{\frac{1}{3}},\ ] ] thus establishing ( [ expxxx ] ) .
we remark that if @xmath368 is real , then @xmath369 may be chosen real and hence @xmath370 is real .
we observe that ( [ expxxx ] ) and ( [ expxxxx ] ) are consistent with @xmath371.\end{aligned}\ ] ] a further consistency check is given by ( [ relation ] ) . using the form of the period matrix , the legendre relation ( [ legndre ] ) and the proposition ( with @xmath372 ) we obtain ( [ relation ] ) . at this stage
we have reduced the ercolani - sinha constraints to one of imposing the four constraints ( [ expxx ] ) on the periods @xmath373 . in particular
this means we must solve @xmath374 which means @xmath375 $ ] .
further we have from the conditions ( [ condition1 ] ) that @xmath376= \sum_{i=1}^3 ( n_i^2-n_im_i+m_i^2)-n_4 ^ 2-m_4 ^ 2-m_4n_4<0 . \label{nmconditions}\end{aligned}\ ] ] our result admits another interpretation .
thus far we have assumed we have been given an appropriate curve and sought to satisfy the ercolani - sinha constraints .
alternatively we may start with a curve satisfying ( most of ) the ercolani - sinha constraints and seek one satisfying the reality constraints ( and any remaining ercolani - sinha constraints ) .
how does this progress ?
first note that the period matrix ( [ taumat ] ) for a curve satisfying ( [ nmconditions1 ] ) is independent of @xmath370 : it is determined wholly in terms of the ercolani - sinha vector .
let us then start with a primitive vector @xmath377 satisfying the hyperboloid condition ( [ nmconditions ] ) and lemma [ ueventheta ] . from this
we construct a period matrix and then , via proposition [ matsumoto2 ] , a normalized curve ( [ curvegenb ] ) .
now we must address whether the curve has the correct reality properties .
for this we must show that there exists a mbius transformation of the set @xmath378 to one of the form @xmath379 .
we will show below that this question may be answered , with the roots @xmath380 being determined up to an overall rotation . at this stage
we have ( using the rotational freedom ) a curve of the form @xmath381 to reconstruct a monopole curve we need a normalization @xmath382\sp{1/2}$ ] .
this is encoded in @xmath370 , which has not appeared thus far . to calculate the normalization we must calculate a period . then using ( [ nmconditions1 ] ) and ( [ expxxx ] ) we determine @xmath368 .
this is a constraint . for a consistent monopole curve
we require @xmath383\sp{1/6}.\ ] ] of course , to complete the construction we need to check there are no roots of the theta function in @xmath384 $ ] .
although the procedure outlined involves several transcendental calculations it is numerically feasible and gives a means of constructing putative monopole curves . to conclude we state when there exists a mbius transformation of the set @xmath385 to one of the form @xmath379 . for simplicity
we give the case of distinct roots : the roots @xmath378 are mbius equivalent to @xmath379 if and only if 1 .
if just one of the roots , say @xmath386 , is real and * @xmath387 then @xmath388 , * @xmath389 then @xmath390 , * @xmath391 then @xmath392 .
+ if all three roots are real then , up to relabelling , one of the above must hold .
all three roots are complex and , up to relabelling , @xmath393
in this section we shall consider the curve @xmath11 specialised to the form @xmath394 where @xmath17 is a real parameter . in this case branch points are @xmath395 where @xmath396 and @xmath397 are real , @xmath398{\frac{-b+\sqrt{b^2 + 4}}{2}}>0,\quad \beta=\sqrt[3]{\frac{-b-\sqrt{b^2 + 4}}{2}}<0,\quad \alpha^3\beta^3=-1.\ ] ] here
@xmath399 is real and we choose our branches so that @xmath369 is also real .
the effect of choosing such a symmetric curve will be to reduce the four period integrals @xmath400 to two independent integrals .
the tetrahedrally symmetric monopole is in the class ( [ curve ] ) .
we note that a general rotation will alter the form of @xmath401 .
thus the dimension of the moduli space is reduced from three by the 3 degrees of freedom of the rotations yielding a discrete space of solutions .
we are seeking then a discrete family of spectral curves .
we shall begin by calculating the period integrals , and then imposing the ercolani - sinha constraints .
we shall also consider the geometry of the curves ( [ curve ] ) . in terms of our wellstein parameterisation
we are working with @xmath402 ( @xmath403 ) .
we choose the first sheet so that @xmath404{(z^3-\alpha\sp3)(z^3+{1}/{\alpha\sp3})}$ ] is negative and real on the real @xmath405-axis between the branch points @xmath406 .
introduce integrals computed on the first sheet @xmath407 here @xmath408 is the standard gauss hypergeometric function and we have , for example , evaluated the first integral using the substitution @xmath409 and our specification of the first sheet .
we also have that @xmath410 our aim is to express the periods for our homology basis ( [ homology ] ) in terms of the integrals @xmath411 and @xmath412 .
consider for example @xmath413 = ( 1-\rho\sp2)\left[-\mathcal{i}_1(\alpha)+ \rho^2\mathcal{j}_1(\alpha)\right]\\ & = -2\mathcal{i}_1(\alpha)-\mathcal{j}_1(\alpha)-\rho\left [ \mathcal{i}_1(\alpha)+2\mathcal{j}_1(\alpha)\right].\end{aligned}\ ] ] here we have used that on the second sheet @xmath414 to obtain the last expression of the first line , and also that @xmath415 to obtain the final expression .
similarly we find ( upon dropping the @xmath396 dependence from @xmath416 and @xmath417 when no confusion arises ) that @xmath418 note that @xmath419 we next reduce the ercolani - sinha constraints to a number theoretic one . using ( [ expxx ] ) and ( [ ij ] ) we may rewrite the constraints as @xmath420 we may solve for the various @xmath421 in terms of @xmath422 as follows .
set @xmath423 then ( [ esred ] ) may be rewritten as @xmath424 giving @xmath425 this yields that the vectors @xmath426 , @xmath427 are of the form @xmath428 one may verify that for vectors of this form then @xmath429 as required by ( [ cplxmt ] ) . recall further that @xmath430 is to be a primitive vector : that is one for which the greatest common divisor of the components is 1 , and hence a generator of @xmath431 .
we see that @xmath430 is primitive if and only if @xmath432 from @xmath433 we obtain @xmath434 now given ( [ solvn ] ) we find that @xmath435 and so the constraint ( [ condition1 ] ) is satisfied if @xmath436= 2\xi^{2}(m_{1 } + n_{1})(m_{1 } - 2n_{1 } ) < 0.\ ] ] this requires @xmath437 in particular we have from ( [ esnu ] ) that @xmath438 thus we have to solve @xmath439 using the identity @xmath440 we then seek solutions of @xmath441 from ( [ esccnm ] ) the ratio of @xmath442 is negative .
consideration of the function @xmath443 ( see figure 2 for its plot ) shows that there exists unique root @xmath444 for each value @xmath445 and correspondingly a unique real positive @xmath446{t/(1-t)}$ ] .
[ figratio ] bringing these results together we have established : [ propesnt ] to each pair of relatively prime integers @xmath447 for which @xmath448 we obtain a solution to the ercolani - sinha constraints for a curve of the form ( [ curve ] ) as follows .
first we solve for @xmath449 , where @xmath450 then @xmath451 and we obtain @xmath353 from @xmath452 with @xmath453 .
thus far we have reduced the problem of finding an appropriate monopole curve within the class ( [ curve ] ) to that of solving the transcendental equation ( [ esct ] ) for which a unique solution exists .
can this ever be solved apart from numerically ? here we shall recount how a ( recently proved ) result of ramanujan enables us to find solutions .
let @xmath19 be a natural number . a modular equation of degree @xmath19 and signature @xmath454 ( @xmath455 ) is a relation between @xmath396 , @xmath397 of the form @xmath456 when @xmath457 we have the complete elliptic integral @xmath458 and ( [ modnr ] ) yields the usual modular relations . by interchanging @xmath459 we may interchange @xmath460 .
this , together with iteration of these modular equations , means we may obtain relations with @xmath19 being an arbitrary rational number .
our equation ( [ esct ] ) is precisely of this form for signature @xmath461 and starting with say @xmath462 .
ramanujan in his second notebook presents results pertaining to these generalised modular equations and various theta function identities .
for example , if @xmath106 in signature @xmath461 then @xmath396 and @xmath397 are related by @xmath463 he also states that ( for @xmath464 ) @xmath465 ramanujan s results were derived in @xcite ( see also @xcite ) , though some related to expansions of @xmath466 had been obtained earlier by j.m . and p.b .
borwein @xcite .
an account of the history and the associated theory of these equations may be found in the last volume dedicated to ramanujan s notebooks @xcite .
the associated theory of these modular equations presented in the accounts just cited is largely based on direct verification that appropriate expressions of hypergeometric functions satisfy the same differential equations and initial conditions and so are equal : we shall present a more geometric picture in due course .
analogous expressions to ( [ mod23 ] ) are known for @xmath176 , @xmath467 , @xmath301 and @xmath468 ( * ? ? ?
* respectively ) .
thus by iteration we may solve ( [ esct ] ) for rational numbers whose numerator and denominator have these as their only factors .
we include some examples of these in the table below .
thus to get the value @xmath299 for the ratio @xmath469 we set @xmath470 in ( [ mod23 ] ) and solve for @xmath471 taking the larger value @xmath472 ( the smaller value yielding the ratio @xmath473 ) .
@xmath474{2}-18\sqrt[3]{4})/250 & ( 44 + 38\sqrt[3]{2}+26\sqrt[3]{4})/3\\ \hline 5&-2&4&\frac{1}{2}+\frac{153\sqrt{3}-99\sqrt{2}}{250 } & 9\sqrt{458 + 187\sqrt{6}}\\ \hline \end{array}\ ] ] a theory exists then for solving ( [ modnr ] ) and this has been worked out for various low primes .
these results enable us to reduce the ercolani - sinha conditions ( [ esct ] ) to solving an algebraic equation .
we shall now describe some geometry underlying our curves ( [ curve ] ) which will lead to an understanding of the results of the last section .
we shall first present a more computational approach , useful in actual calculations , and then follow this with a more invariant discussion .
we begin with the observation that our curves each cover four elliptic curves .
[ coverlem ] the curve @xmath475 with arbitrary value of the parameter @xmath17 is a simultaneous covering of the four elliptic curves @xmath476 , @xmath477 as indicated in the diagram , where @xmath478 is an intermediate genus two curve : 0.5 cm ( 60,30)(0,0 ) ( 9,23)(0,0)@xmath479 ( 6,20)(-4,-1)20 ( -8,20)(0,0)@xmath480 ( -18,15)(0,0)@xmath481 ( -26,0)(0,0)@xmath482 ( -20,12)(-1,-1)9 ( -16,12)(1,-1)9 ( 0,0)(0,0)@xmath483 ( -9,10)(0,0)@xmath484 ( -26,10)(0,0)@xmath485 ( 10,19)(1,-1)16 ( 16,10)(0,0)@xmath486 ( 28,0)(0,0)@xmath487 ( 12,20)(2,-1)34 ( 38,10)(0,0)@xmath488 ( 55,0)(0,0)@xmath489 the equations of the elliptic curves are @xmath490 where the jacobi moduli , @xmath491 are given by @xmath492 with @xmath493 the covers @xmath494 are given by @xmath495 and @xmath496 the elliptic curves @xmath477 are equianharmonic ( @xmath497 ) and consequently have vanishing @xmath328-invariant , @xmath498 .
the derivation of the covers @xmath499 and the underlying curves is straightforward .
the pullbacks @xmath500 of these covers are @xmath501{z_2}\\ y = w_2\end{cases}\end{aligned}\ ] ] showing that the degrees of the cover are 2 and 3 respectively .
a direct calculation putting these elliptic curves into weierstrass form shows @xmath497 and hence the elliptic curves @xmath477 are equianharmonic .
their @xmath328-invariants are therefore vanishing and @xmath477 are birationally equivalent .
to derive the covers @xmath502 we first note that the curve @xmath11 is a covering of the hyperelliptic curve @xmath503 of genus two , @xmath504 the cover of this curve is given by the formulae @xmath505 the curve @xmath503 covers two - sheetedly the two elliptic curves @xmath476 given in ( [ curvepm ] ) @xmath506 composition of ( [ piast ] ) and ( [ cover2a ] ) leads to ( [ pipm ] ) . using these formulae direct calculation then yields the holomorphic differentials of @xmath11 are mapped to holomorphic differentials of @xmath476 , @xmath477 as follows @xmath507 where @xmath508 are given in ( [ klm ] ) .
the absolute invariants @xmath509 of the curves @xmath476 are @xmath510 evidently @xmath511 in general , as well @xmath512 ; therefore these elliptic curves are not birationally equivalent to that one appearing in hitchin s theory of the tetrahedral monopole which is equianharmonic @xcite .
we observe that the substitution @xmath513 leads to the parameterisation of jacobi moduli being @xmath514 which ramanujan used in his hypergeometric relations of signature 3 , see e.g. @xcite .
the @xmath15-functional representation of the moduli @xmath491 and parameter @xmath515 can be found in ( * ? ? ? * section 9.7 ) , @xmath516 we shall now describe the geometry of the covers we have just presented explicitly .
our curve has several explicit symmetries which lie behind the covers described .
we will first describe these symmetries acting on the field of functions @xmath517 of our curve as this field does not depend on whether we have a singular or nonsingular model of the curve ; we will subsequently give a projective model for these , typically working in weighted projective spaces where the curves will be nonsingular .
viewing @xmath518 and @xmath519 as functions on @xmath11 we see that @xmath520 has symmetries ( @xmath521 ) @xmath522 together these yield the group @xmath523 , with @xmath524 and @xmath525 . when @xmath526 , the dihedral symmetry @xmath527 is enlarged to tetrahedral symmetry by @xmath528 with @xmath529 being generated by @xmath530 and @xmath531 .
now to each subgroup @xmath532 we have the fixed field @xmath533 associated to the quotient curve @xmath534 .
the canonical curve of a non - hyperelliptic curve of genus 4 is given by the intersection of an irreducible quadric and cubic surface in @xmath535 .
in our case the quadric is in fact a cone and we may represent our curve @xmath11 as the nonsingular curve as the plane curve given by the vanishing of @xmath536 the curve is singular .
when @xmath17 is real the point @xmath537=[0,0,1]$ ] is the only singular point of @xmath11 with delta invariant @xmath298 and multiplicity @xmath297 yielding @xmath538 .
] in the weighted projective space @xmath539\,|\,[z , t , w]\sim [ \lambda z,\lambda t , \lambda\sp2 w]\}$ ] given by the vanishing of @xmath540 the group @xmath541 acts on this as ( @xmath542 , @xmath543 ) @xmath544\rightarrow [ z , t,\rho w]\sim [ \rho z,\rho t , w],\\ \mathrm{b}:\ & [ z , t , w]\rightarrow [ \rho z , t,\rho w]\sim [ \rho\sp2 z,\rho t , w ] , \\
\mathrm{c}:\ & [ z , t , w]\rightarrow [ t ,- z ,- w]\sim [ \imath t , -\imath z , w].\end{aligned}\ ] ] the fixed points of these actions on @xmath11 and quotient curves are as follows : @xmath545 : : there are 6 fixed points , @xmath546 $ ] , where @xmath547 are the two roots of @xmath548 . for other points
we have a @xmath549 map @xmath550 .
an application of the riemann - hurwitz theorem shows the genus of @xmath551 to be @xmath552 . @xmath530
: : this has no fixed points and an application of the riemann - hurwitz theorem shows the genus of @xmath553 to be @xmath554 .
@xmath555 : : there are 6 fixed points , @xmath556 $ ] , where @xmath557 is a root of @xmath558 . here
the riemann - hurwitz theorem shows the genus of @xmath559 to be @xmath560 . by using the invariants of @xmath561
we may obtain nonsingular projective models of @xmath533 .
take for example @xmath562 with invariants @xmath563 , @xmath564 and @xmath304 ( in degree @xmath299 ) .
then we obtain the quotient curve @xmath565 in @xmath566\}$ ] .
the genus of the quotient is seen to be @xmath296 .
we recognise this as the curve @xmath567 .
one verifies that @xmath568 giving us the invariant differential ( [ invdiff1 ] ) .
similarly , by taking @xmath569 and @xmath570 , we also obtain equianharmonic elliptic curves .
the invariants of the involution @xmath571 are again all in degree @xmath299 and now are @xmath563 , @xmath572 and @xmath304 . by taking @xmath573 we may identify @xmath574 .
the invariant of @xmath575\rightarrow[\rho z , t , w]$ ] is @xmath576 and the curve @xmath577 in @xmath578\}$].using the formula for the genus of a smooth curve of degree @xmath579 in @xmath580 , @xmath581 the genus is seen to be @xmath296 . now
( [ invdiff2 ] ) is the invariant differential for this action . if we had taken @xmath582 with invariants @xmath576 , @xmath583 and @xmath304 we obtain the curve @xmath584 in @xmath585 ( which is equivalent to @xmath586 in @xmath587 ) .
the genus of this quotient is seen to be @xmath295 .
we obtain the genus @xmath299 curve @xmath503 as follows .
the invariants of @xmath588 are @xmath589 , @xmath590 , @xmath591 and @xmath304 , subject to the relation @xmath592 .
the curve @xmath11 may be written @xmath593 , and hence @xmath594 .
this curve has genus 2 in @xmath595\}$ ] and may be identified with @xmath503 .
be setting @xmath596 this curve takes the form @xmath597 in @xmath598\}$ ] and the identification with @xmath503 in the affine chart of earlier is given by @xmath599 , @xmath600 . in this latter form
we find that the action of @xmath555 is given by @xmath601\rightarrow [ -u,\nu ,- w]\sim [ u,-\nu , w]$ ] which is the hyperelliptic involution ; further quotienting yields a genus 0 curve .
the remaining genus 1 curves @xmath602 are identified with the quotients of @xmath503 by @xmath603{4+b^2}$ ] , @xmath604{4+b^2}u$ ] , @xmath605 .
this action has invariants @xmath606 ( in degree 2 ) , @xmath607{4+b^2}u$ ] ( in degree 1 ) , and @xmath355 ( in degree 3 ) . the resulting degree 6
curve is @xmath608 where , as previously , @xmath609{4+b^2}$ ] .
these curves have genus @xmath296 in @xmath610\}$ ] . to complete the identification with @xmath602
we compute the @xmath328-invariants of these curves . in the affine patch with @xmath611 which looks like @xmath612 ( the other affine patches have orbifold singularities and hence this choice ) the curve takes the form @xmath613 the @xmath328-invariants of these curves agree with ( [ invariants ] ) and hence the identifications as stated .
both the differentials @xmath614 and @xmath615 are invariant under @xmath530 .
these may be obtained by linear combinations of @xmath616 ( [ invdiffpma ] ) .
the latter differentials are those invariant under the symmetry of ( [ hcurve ] ) @xmath617 which yield the quotients @xmath602 .
a birational transformation makes this symmetry more manifest .
let @xmath618 then ( [ hcurve ] ) transforms to @xmath619 which is manifestly invariant under @xmath620 , @xmath621 .
the substitution @xmath622 reduces the canonical differentials @xmath623 and @xmath624 to the canonical differentials the elliptic curves @xmath625 which correspond to our earlier parameterisations .
we have seen that complete abelian integrals of the curve @xmath11 ( [ bren03 ] ) are given by hypergeometric functions .
the same is true for the various curves given in lemma [ coverlem ] covered by @xmath11 . relating the periods of @xmath11 and the curves it covers leads to various relations between hypergeometric functions , and this underlies the higher hypergeometric identities of goursat @xcite .
goursat gave detailed tables of transformations of hypergeometric functions up to order four that will be enough for our purposes .
the simplest example of this is the cover @xmath626 for which @xmath627 and @xmath628 .
one then finds for example that @xmath629 where both @xmath630 and @xmath355 are evaluated on the first sheet . a change of variable shows that @xmath631 now the left - hand side of equation ( [ gour1 ] ) is @xmath632 (
the minus sign arising when we go to wellstein variables @xmath633 ) and this has been evaluated in ( [ integralsij ] ) .
comparison of these two representations yields the hypergeometric equality @xmath634 which is one of goursat s quadratic equalities @xcite ; see also ( * ? ? ?
2.11 , eq .
further identities ensue from the coverings @xmath635 and we shall describe these as needed below .
we remark that the curve ( [ hcurve ] ) already appeared in hutchinson s study @xcite of automorphic functions associated with singular , genus two , trigonal curves in which he developed earlier investigations of burkhardt @xcite .
these results were employed by grava and one of the authors @xcite to solve the riemann - hilbert problem and associated schlesinger system for certain class of curves with @xmath636-symmetry .
it is possible for the theta functions associated to a period matrix @xmath83 to simplify ( or admit reduction ) and be expressible in terms of lower dimensional theta functions .
such happens when the curve covers a curve of lower genus , but it may also occur without there being a covering . reduction may be described purely in terms of the riemann matrix of periods ( see @xcite ; for more recent expositions and applications see @xcite,@xcite ) .
a @xmath637 riemann matrix @xmath638 is said to admit _ reduction _ if there exists a @xmath639 matrix of complex numbers @xmath112 of maximal rank , a @xmath640 matrix of complex numbers @xmath641 and a @xmath642 matrix of integers @xmath643 also of maximal rank such that @xmath644 where @xmath645 . when a riemann matrix admits reduction the corresponding period matrix may be put in the form @xmath646 where @xmath647 is a @xmath648 matrix with rational entries and the matrices @xmath649 and @xmath650 have the properties of period matrices . because @xmath647 here has rational entries there exists a diagonal @xmath651 matrix @xmath652 with positive integer entries for which @xmath653 . with @xmath654 the theta function associated with @xmath83
may then be expressed in terms of lower dimensional theta functions as @xmath655(dw;d\tau\sp{\#}d).\ ] ] our curve admits many reductions . of itself
this just means that the theta functions may be reduced to theta functions of fewer variables .
it is only when the ercolani - sinha vector correspondingly reduces that we obtain real simplification . in the remainder of this section
we shall describe these reductions and later see how dramatic simplifications occur .
first let us describe the riemann matrix of periods .
we may evaluate the remaining period integrals as follows .
let @xmath656 then for @xmath657 we have that @xmath658 where it is again supposed that the integrals @xmath659 and @xmath660 are computed on the first sheet .
we have already computed @xmath411 and @xmath412 .
the integrals @xmath659 and @xmath660 are found to be @xmath661 with @xmath662 .
we observe that the relations @xmath663 follow from the above formulae .
the vectors @xmath664 are @xmath665 one may can easily check that @xmath666 we then have that @xmath667 the ercolani - sinha conditions , @xmath668 written for the vectors @xmath669 lead to the equations @xmath670 which were obtained earlier .
a calculation also shows that the relation ( [ cplxmt ] ) @xmath671 yielding a nontrivial check of our procedure . the integrals between infinities may be reduced to our standard integrals by writing @xmath672 where we write @xmath673 and @xmath674
is any of the branch points .
these are then calculated to be @xmath675 our riemann matrix admits a reduction with respect to any of it columns .
we will exemplify this with the first column , a result we will use next ; similar considerations apply to the other columns .
now from the above and ( [ expxx ] ) it follows that @xmath676 where @xmath643 is the @xmath677 integral matrix @xmath678 then to every two ercolani - sinha vectors @xmath426 , @xmath427 we have that @xmath679 the number @xmath579 here is often called the hopf number . in particular for @xmath680 then @xmath643 is of maximal rank and consequently our riemann matrix admits reduction .
let us now focus on the consequences of reduction for symmetric monopoles .
[ symwp ] for the symmetric monopole we may reduce by the first column using the vector ( [ escond2 ] ) whose elements are related by ( [ solvn ] ) , with @xmath447 . then @xmath681 and for @xmath680 there exists an element @xmath682 of the symplectic group @xmath683 such that @xmath684 letting @xmath685 then @xmath686,n_1 - 2m_1-p\,[m_1 - 2n_1]).\ ] ] when @xmath687 a further symplectic transformation allows the simplification @xmath688 . under @xmath682
the ercolani - sinha vector transforms as @xmath689 the proof of the theorem is constructive using work of krazer , weierstrass and kowalewski .
martens @xcite has given an algorithm for constructing @xmath682 which we have implemented using @xmath690 . because @xmath682 depends on number theoretic properties of @xmath691 and @xmath692 the form is rather unilluminating and we simply record the result ( though an explicit example will be given in the following section ) .
what is remarkable however is the simple universal form the ercolani - sinha vector takes under this transformation .
this has great significance for us as we next describe . using ( [ redthetagen]),([taureduced ] ) and
say @xmath693 we have that a smaller multiple than @xmath694 would suffice here with correspondingly fewer terms in the sums @xmath695 . ]
@xmath696(dw;d{\tau}^{\#}d)\\ & = \sum_{m=0}\sp{d-1 } \theta \left[\begin{array}{c}\frac{m}{d}\\0\end{array}\right](dz;d(\rho+2))\,\theta((w_1+\frac{m\alpha}{d},w_2,w_3);{\tau}^{\#}),\end{aligned}\ ] ] where we have genus one and three theta functions on the right hand - side here .
comparison of ( [ newbafnch ] ) and ( [ essymptrans ] ) then reveals that the theta function dependence of the baker - akhiezer is given wholly by the genus one theta functions .
further simplifications ensue from the identity @xmath697(dz;d\tau)=\mu(\tau)\,\prod_{l=0}\sp{d-1 } \theta \left[\begin{array}{c}\frac{\epsilon}{d}\\ \frac { \epsilon'}{d}+\frac{d-(1 + 2l)}{2d}\end{array}\right](z;\tau),\ ] ] where @xmath698 is a constant .
we then have [ symel ] for symmetric monopoles the theta function @xmath405-dependence of ( [ newbafnch ] ) is expressible in terms of elliptic functions .
thus far we have not discussed the final hitchin constraint for symmetric monopoles .
this theorem reduces the problem to one of the zeros of elliptic functions . , @xmath699 .
, width=264 ] , @xmath699 . ,
width=264 ] the graph in figure [ fig:21 ] shows the real and imaginary parts of the theta function denominator of @xmath177 for the @xmath700 , @xmath699 symmetric monopole , the @xmath701 ramanjuan case .
these vanish at @xmath702 and @xmath703 as desired , but additionally one finds vanishing at @xmath704 and @xmath705 . calculating
the theta function with shifted argument in the numerator shows that there is no corresponding vanishing and consequently @xmath177 yields unwanted poles in @xmath706 .
thus the @xmath700 , @xmath699 curve does not yield a monopole . a similar evaluation of the relevant @xmath707 , @xmath708 and @xmath709
, @xmath710 theta functions also reveals unwanted zeros and of the those cases from our table of symmetric 3-monopoles only the tetrahedrally symmetric case has the required vanishing .
indeed extensive numerical calculations suggests : for a symmetric monopole the denominator of @xmath177 has @xmath711 zeros and consequently the tetrahedrally symmetric monopole is the only monopole in this class .
although other techniques exist for the study of monopoles which might allow one to prove that the tetrahedrally symmetric monopole is the only monopole in the class of symmetric monopoles we do nt know of a suitable theory that counts the number of times a _ real _ line ( interval ) intersects the theta divisor .
we would welcome such a theory .
our results do however allow us to say more about the curve of the tetrahedrally symmetric monopole . before turning to a more detailed examination of this case in our next section
we first describe how to calculate the remaining quantities appearing in the formula for @xmath177 . here
we follow section 5.4 .
we calculate the @xmath117-periods of the differential @xmath343 in a manner similar to the period integrals already calculated .
introduce integrals on the first sheet @xmath712 evidently @xmath713 and @xmath714 and one finds that @xmath715 we find , as before in the case of holomorphic differentials , that @xmath716 the legendre relation ( [ legndre ] ) gives a non trivial consistency check of our calculations .
this may be written in the form of the following hypergeometric equality @xmath717 and this may be established by standard means . to calculating @xmath335 using ( [ nuijgpgr ] )
introduce the differential of the second kind , @xmath718 with second order pole at @xmath295 on all sheets , @xmath719 ( here we took into account @xmath720 for @xmath721 . ) then @xmath722 the last integral in ( [ nuij1s ] ) may also be expressed in terms of hypergeometric functions as follows .
first we remark that @xmath723 where @xmath724 .
next , for the integrals on the first sheet we have @xmath725
the curve of the tetrahedrally symmetric monopole is of the form @xmath726 in this case we may take @xmath727 for these values we may explicitly evaluate the various hypergeometric functions . using ramanujan s identity ( [ ramf23 ] ) together with the standard quadratic transformation of the hypergeometric function @xmath728 ( valid for @xmath729 , @xmath730 ) we find that @xmath731 ( in verifying this we note that @xmath732 is the relevant value leading to our @xmath449 in ( [ ramf23 ] ) . )
now this last hypergeometric function is related to an elliptic integral we may evaluate ( * ? ? ?
* p 86 ) , @xmath733 bringing these results together we finally obtain @xmath734 then from ( [ esevch ] ) we obtain that @xmath735 this agrees with the result of @xcite .
we also note that upon using goursat s identity ( * ? ? ?
* ( 39 ) ) @xmath736 we may establish the result of @xcite based on numerical evaluation , that @xmath737 using these results and those of the previous section we have , the tetrahedral 3-monopole for which @xmath526 admits the @xmath83-matrix of the form @xmath738 we have already seen that the symmetric monopole curve @xmath11 covers two equianharmonic torii @xmath477 . for the value of the parameter @xmath526
the curve covers three further equianharmonic elliptic curves .
these may be described as follows . for @xmath739 , @xmath300 ,
@xmath467 let @xmath740 be defined by the formulae @xmath741 then @xmath742 and we have the following relations between holomorphic differentials @xmath743 the final of these rational maps was introduced by @xcite and has the following significance .
let @xmath350 and @xmath744 be the @xmath117 and @xmath118-periods of the differential @xmath745 and denote by @xmath746 , @xmath747 the @xmath117 and @xmath118-periods of the elliptic differential @xmath748 .
then @xmath749 where @xmath750 is the matrix @xmath751 satisfying the condition @xmath752 introduce the homology basis for the elliptic curve as shown in figure [ fig : elldiffbasis ] and set @xmath753 then @xmath754 from the reduction formula ( [ differentials5 ] ) we next conclude that @xmath755 and therefore have that @xmath756 equations ( [ xy ] ) and ( [ xy1 ] ) permit us to express @xmath757 and comparison with ( [ absym ] ) yields the given @xmath750 . the condition ( [ hopf1 ] ) is checked directly .
the number 4 appearing in ( [ hopf1 ] ) means that the cover @xmath758 given in ( [ cover345 ] ) be of degree 4 .
we remark that the matrix @xmath750 of the proposition is obtained from the @xmath643 of ( [ escond3 ] ) by @xmath759 , which simply reflects our choice of homology basis .
thus we are discussing the reduction of the previous section . indeed with @xmath760 \equiv \left(\begin{array}{cc } a & b \\
\end{array}\right)\ ] ] we find that the @xmath83 matrix ( [ tetratau ] ) transforms to @xmath761.\ ] ] combined with theorem [ symwp ] we may reduce our expression for the tetrahedral monopoles @xmath177 to one built out of jacobi elliptic theta functions .
although monopoles have been studied now for many years and from various perspectives , relatively few analytic solutions are known .
the outstanding problem is constructing curves satisfying the transcendental constraints of hitchin placed upon it . using the connection with integrable systems
this article has sought to make effective such solutions .
it is nevertheless only early steps upon this road .
here we have seen how one class of transcendental constraints may be replaced by a number theoretic problem . in applying our techniques beyond the known case of charge two we considered the restricted class of charge three monopoles ( [ welstein99 ] ) which includes the tetrahedrally symmetric monopole .
this family of curves has many arithmetic properties that facilitates analytic integration .
in particular the period matrix may be explicitly expressed in terms of just four integrals . using this
we were able to explicitly solve the ercolani - sinha constraints that are equivalent to hitchin s transcendental condition ( h2 ) of the triviality of a certain line bundle over the spectral curve ( proposition [ ourh2 ] ) .
our approach reduces the problem to that of determining certain rationality properties of the ( four ) relevant periods .
( our result also admits another approach to seeking monopole curves : we may solve the ercolani - sinha constraints and then seek to impose hitchin s reality conditions on the resulting curves . ) to proceed further in this rather uncharted territory we further restricted our attention to what we have referred to as symmetric 3-monopoles " whose spectral curve has the form ( [ bren03 ] ) .
this reduced the required independent integrals from four to two , each of which were hypergeometric in form , and the rationality requirement is now for the ratio of these ( proposition [ propesnt ] ) . extensions of work by ramanujan mean this latter question may be replaced by number theory and of seeking solutions of various algebraic equations ( depending on the primes involved in the rational ratio ) .
examples of such solutions were given ( again including the tetrahedral case ) .
we further examined the symmetries and coverings of these symmetric curves and their relation to higher goursat hypergeometric identities .
having at hand now many putative spectral curves we proceeded to evaluate the remaining integrals needed in our construction .
remarkably we discovered that application of weierstrass reduction theory showed that the ercolani - sinha vector transformed to a universal form and that all of the theta function @xmath405-dependence for symmetric 3-monopoles was expressible in terms of elliptic functions ( theorems [ symwp ] , [ symel ] ) .
the final selection of permissible spectral curves at last reduced to the question of zeros of these elliptic functions .
we are hampered by not knowing a suitable theory for the calculation of the number of times a real line interval intersects the theta divisor .
numerical calculation of these zeros led to a universal conjecture ( 5.6 ) suggesting that of the symmetric 3-monopoles only the tetrahedral monopole has the required zeros .
our final section then was devoted to the charge three tetrahedrally symmetric monopole .
here we were able to substantially simplify known expressions for the period matrix of the spectral curve as well as prove a conjectured identity of earlier workers .
again an explicit map was given and we have been able to reduce entirely to elliptic functions .
the final comparison with the nahm data of @xcite will be left for a subsequent work .
this paper was conceived at the epsrc funded newton institute programme `` integrable systems '' in 2001 and had early encouragement from hermann flaschka and nick ercolani .
the many technical hurdles encountered in this work meant growth has come in spurts rather than continuous progress and it has occupied our thoughts for much of this time .
we are grateful to an epsrc small grant and most recently a misgam small grant enabling the authors to come together to draw the work into its present form . over the intervening years
we have benefited from discussions and correspondence with many colleagues and we wish to thank : nigel hitchin , conor houghton and paul sutcliffe for their remarks on monopoles ; david calderbank , miles reid , richard thomas and armando treibich for their geometric and algebro - geometric advice ; raimundas vidunas and adri olde daalhuis for references on matters hypergeometric ; mike eggar and keiji matsumoto for help with topological aspects of our curve ; and chris eilbeck , tamara grava , yuri fedorov , john mckay and stan richardson for discussions pertaining to riemann surfaces . we have enjoyed a problem touching on so many aspects of mathematics .
for @xmath762 the canonical riemann @xmath15-function is given by @xmath763 the @xmath15-function is holomorphic on @xmath764 and satisfies @xmath765 where @xmath766 .
the riemann @xmath15-function @xmath767 with characteristics @xmath768 is defined by @xmath769 where @xmath770 .
this is also written as @xmath771(\boldsymbol{z};\tau).\ ] ] for arbitrary @xmath770 and @xmath772 the following formula is valid @xmath773 the function @xmath774 is called the @xmath15-constant with characteristic @xmath775 .
we have @xmath776 the following transformation formula is given in ( * ? ? ?
* p85 , p176 ) . for any @xmath777 and
@xmath778 we put @xmath779 where @xmath780 is the row vector consisting of the diagonal components of @xmath3 .
then for every @xmath781 we have @xmath782 in which @xmath783 is a @xmath300-th root of unity depending only on @xmath784 while @xmath785 and @xmath786 is a complex number independent of @xmath83 and @xmath405 such that @xmath787 . *
the vector of riemann constants * the convention we adopt for our vector of riemann constants is @xmath788 in the jacobi inversion .
this is the convention used by farkas and kra and the negative of that of mumford ; the choice of signs appears in the actual construction of @xmath789 , such as ( 2.4.1 ) of farkas and kra .
then @xmath790 the vector of riemann constants depends on the homology basis and base point @xmath647 . if we change base points of the abel map @xmath791 then @xmath792 . with this convention
@xmath793 * theta characteristics * the set @xmath794 of divisor classes @xmath795 such that @xmath796 , the canonical class , is called the set of _ theta characteristics _ of @xmath11 .
the set @xmath794 is a principal homogeneous space for the group @xmath797 , the group of @xmath299-torsion points of the group @xmath798 of degree zero line bundles on @xmath11 .
equivalently this may be viewed as the @xmath299-torsion points of the jacobian , @xmath799 .
geometrically if @xmath370 is a holomorphic line bundle on @xmath11 such that @xmath800 is holomorphically equivalent to @xmath801 then the divisor of @xmath370 is a theta characteristic .
if @xmath802 is a holomorphic line bundle of order @xmath299 , that is @xmath120 is holomorphically trivial , then the divisor of @xmath803 is also a theta characteristic .
thus there are @xmath804 theta characteristics .
we may view @xmath805 as a vector space of dimension @xmath806 over @xmath807 .
this vector space has a nondegenerate symplectic ( and hence symmetric as the field is @xmath807 ) form defined by the weil pairing .
if @xmath795 and @xmath808 are divisors with disjoint support in the classes of @xmath809 and @xmath810 respectively , and @xmath811 , @xmath812 then the weil pairing is @xmath813 where if @xmath814 then @xmath815 .
mumford identifies @xmath807 with @xmath816 by sending @xmath295 to @xmath296 and @xmath296 to @xmath817 .
( in general we may consider @xmath818 , the @xmath454-torsion points of @xmath798 , and the weil pairing gives us a nondegenerate antisymmetric map @xmath819 where @xmath820 are @xmath454-th roots of unity . )
the @xmath807 vector space @xmath797 may be identified with @xmath821 and with this identification @xmath822 is simply the cup product .
define @xmath823 by @xmath824 where @xmath825 is the line bundle with divisor @xmath795 .
then @xmath826 any function @xmath827 satisfying this identity is known as an arf function , and any arf function is given by @xmath827 for some theta characteristic with corresponding line bundle @xmath370 .
thus the space of theta characteristics may be identified with the space of quadratic forms ( [ quadform ] ) .
we shall now describe how to integrate holomorphic differentials between branch points .
we use the fact that for non - invariant holomorphic differentials ( as we have ) @xmath828 indeed , if @xmath231 is any holomorphic differential on a compact riemann surface which is an @xmath829-fold branched cover of @xmath830 then @xmath831 , where @xmath832 are the preimages of @xmath833 .
then @xmath834 and consequently @xmath835 with similar expressions obtained for @xmath836 and @xmath837 . further utilising @xmath838 and @xmath839
we may write @xmath840 appropriate linear combinations of these yield @xmath841 for @xmath226 . for example @xmath842 in order to be able to integrate a holomorphic differential between any branch point we must show how we may integrate such between @xmath843 and @xmath844 on any branch .
now we use that there exist meromorphic functions @xmath845 and @xmath846 ( for each @xmath327 , @xmath328 ) with ( respective ) divisors @xmath847 thus for any normalized holomorphic differential @xmath189 @xmath848 and @xmath849 , where @xmath173 is the period lattice .
these equalities hold ( modulo a lattice vector ) for a path of integration on any branch and so , for example , @xmath850
we wish to determine when there is a mbius transformation between the sets @xmath851 and @xmath852 .
the former corresponds to reality constraints on our data arising from @xmath853 while the latter may be constructed from the period matrix of the curve in terms of various theta constants .
if we have a period matrix satisfying @xmath854 then we must satisfy @xmath853 .
+ at the outset we note that the mbius transformation @xmath643 sending @xmath855 and its inverse @xmath856 @xmath857 are given by @xmath858 the transformation @xmath859 may be represented by the @xmath860 matrix @xmath861 and upon setting @xmath862 we may determine a @xmath860 representation of ( [ mob1 ] ) .
the rotation @xmath866 transforms the set @xmath561 to one of the form @xmath867 where @xmath868 ( @xmath869 ) . upon setting @xmath870
, @xmath871 will transform the latter set to one of the form @xmath872 .
finally the scaling @xmath873 given by @xmath874 transforms @xmath561 to @xmath875 .
such a set is of the desired form @xmath876 and is characterised by 3 ( real ) parameters . with @xmath877 , @xmath878 , @xmath879
we see we have @xmath880 , @xmath881 , @xmath882 . from a set @xmath883 and a choice of @xmath15 and @xmath884 ( equivalently , a rotation )
we may reconstruct @xmath561 . more generally , let us consider images @xmath885 under mbius transformations .
up to a relabelling of roots we have four possibilities of those roots we map to @xmath886 : @xmath887 we have already considered @xmath888 in the previous paragraph .
for completeness let us give @xmath889 for the various cases and the various restrictions arising + a. @xmath890 b. @xmath891 c. @xmath892 d. @xmath893 the constraints @xmath894 for case @xmath895 may be obtained as follows .
further composing the mbius transformation leading to @xmath895 with that giving @xmath896 , @xmath897 , @xmath898 gives us case @xmath888 for which we know the constraint .
this second mbius transformation is given by @xmath899 and we may transfer the constraint of @xmath888 to @xmath895 . similarly composing @xmath900 with @xmath901 yields case @xmath888 up to a relabelling of roots . geometrically cases @xmath888 , @xmath895 , @xmath900 consist of the following .
a circle passes through @xmath902 . under a mbius transformation to the set @xmath903 the circle becomes the real axis and so @xmath904
this is the real parameter appearing in each of these cases
. a similar argument composing @xmath905 with @xmath906 will give the constraints ( [ mobcd ] ) .
v. z. enolski and t. grava , _
singular @xmath636-curves and the riemann - hilbert problem _ , int .
notices * 32 * ( 2004 ) , 16191683 .
n. ercolani and a. sinha , _ monopoles and baker functions _ , commun . math . phys .
* 125 * ( 1989 ) , 385416 .
k. matsumoto , _ theta constants associated with the cyclic triple coverings of the complex projective line branching at six points _ , publ .
inst . math .
sci * 37 * ( 2001 ) , 419440 , arxiv : math.ag/0008025 . | we develop the ercolani - sinha construction of @xmath0 monopoles and make this effective for ( a five parameter family of centred ) charge 3 monopoles . in particular we show how to solve the transcendental constraints arising on the spectral curve . for a class of symmetric curves
the transcendental constraints become a number theoretic problem and a recently proven identity of ramanujan provides a solution . |
in large surveys , often a few objects will be found that are unusual in one respect or another .
our spectroscopic survey of candidate star clusters in the field of m31 @xcite has provided such a case . from the 1400 objects
observed , we expected ( and found ) a large number of misclassified stars , due to the heterogeneous quality of the imaging material used to identify cluster candidates .
all but a few of these stars are in the foreground , as revealed by their dwarf spectral types .
our catalog magnitude limit of v=20 required that any m31 stars observed would have to be bright giants or supergiants .
of the few stars we observed in that area that were clearly not from the foreground milky way disk , one star proved to be unusual in two ways : its spectrum is that of an f supergiant with weak balmer emission , and its heliocentric radial velocity is @xmath0 , which is the most negative velocity yet found for a single star .
the star is called b030d in the bologna catalog of m31 objects @xcite , and `` bol d30 '' in simbad .
this star is unlikely to be a member of m31 s disk : it lies close to the center of m31 on the sw side , at a place where the heliocentric velocity of the disk is @xmath1 ( the minimum disk velocity for m31 reaches @xmath2 ) .
a difference of even 230 with respect to the disk circular velocity is much too high for any disk or thick disk star in m31 .
while m31 s halo velocity distribution is not well known , a velocity of @xmath3 with respect to m31 s center ( which has v= @xmath4 ) is very high , even for a halo star .
could it be an unusual star in the halo of the milky way ?
m31 lies at @xmath5 = 121.2 , @xmath6 = @xmath7 .
accounting for the sun s motion with respect to the lsr and the lsr s motion around the galactic center , 12 , 7 ) and the lsr velocity is 220 @xcite ] reduces the velocity to @xmath8614 with respect to the milky way center . a recent estimate of the galaxy s escape speed ( 544 , * ?
* ) suggests that it is unlikely that this star is bound to the milky way ; and its highly _ negative _ velocity and galactic longitude imply that it can not be a hypervelocity star ejected from the galactic center such as those identified by @xcite .
this paper describes the analysis of the unusual spectrum of the star b030d and the reasoning that led us to conclude that the star is most likely a member of the m31 halo s `` giant stream '' @xcite close to the pericenter of its orbit .
b030d resides about 0.3 sw of the center of m31 ( 4 kpc in projection , see fig [ m31 ] ) , and was originally cataloged by @xcite as a `` bluish and diffuse '' object , based on photographic plates .
it was first observed spectroscopically by us as part of our mmt hectospec spectroscopic study of star clusters in m31 @xcite . from this good ( but not ideal ) spectrum
we immediately realized that the object was not a star cluster , which is confirmed by the images made available as part of the local group galaxy survey ( lggs , * ? ? ? * see our fig .
[ m31 ] ) .
we also learned of its unusual velocity and estimated its spectral type to be f1 , though its luminosity class was difficult to discern .
the higher order balmer lines were narrow , as would be the case for a low gravity star , but the h@xmath9 absorption line was weaker than expected for such a star , and h@xmath10 was very weak .
we used the ubv photometry of @xcite to obtain more information about the nature of the star .
table [ ubvtable ] summarizes ubv photometry and other basic information on this star .
here we see that b030d has v=19.02 .
figure [ ubv ] plots u@xmath8b vs b@xmath8v for all stars from the m31 catalog of @xcite with v brighter than 19.5 .
b030d is shown by a large black circle . where do foreground stars from the milky way lie in this diagram ? in this magnitude range , and at the relatively low galactic latitude of m31 , we would expect to see stars from the thick disk and the halo of the milky way . the dashed line in figure [ ubv ] shows the sequence of solar abundance main sequence stars from @xcite , reddened by the likely foreground reddening to m31 , e(b
v)=0.13 @xcite .
metal poor dwarfs will be found above the dashed line , and indeed the sequence of milky way halo dwarfs can be seen above the line , terminating near b@xmath8v=0.5 , u@xmath8b=0.15 .
b030d s b@xmath8v is too blue to belong to the old halo of the milky way unless it is a field star blue straggler ( e.g. * ? ? ? * ) .
blue stragglers are metal - poor halo stars which are either significantly younger than most of the halo or the product of a merged binary . judging from this diagram , b030d s ubv colors are at best merely consistent with it being a main sequence star . however , there are other options . where do young supergiant stars from m31 appear in this diagram ?
their position depends strongly on their reddening .
massey s spectroscopically confirmed m31 o@xmath8f supergiants are shown in figure [ ubv ] as red solid points , and the supergiant locus from @xcite is shown as a solid black line .
this locus has also been reddened by e(b@xmath8v)=0.13 .
the reddening line , plotted above the supergiants , shows where more highly reddened supergiants would appear in this diagram .
it can be seen that most of the supergiants from massey et al are highly reddened early type stars .
this is not unexpected : we would expect such young stars to be in dusty star forming regions and so be reddened by dust in both the milky way and m31 .
however more recently , @xcite have studied candidate yellow supergiants in m31 , and have identified a number of stars with strong absorption from the triplet of o i@xmath11 , a sensitive indicator of low surface gravity @xcite .
these stars are certainly members of m31 , and are shown in figure [ ubv ] as blue circles .
several of these stars have similar colors to b030d .
unfortunately , b030d lies in the part of the u@xmath8b vs b@xmath8v diagram where the main sequence and the supergiant lines cross .
thus we see that the ubv colors of b030d are consistent with either a very low gravity star such as an m31 supergiant , albeit one with very little reddening from m31 s disk , or with a halo blue straggler from the milky way .
( we have already noted , however , that its velocity is so high that if it were in the milky way it would be unlikely to be bound to our galaxy . ) we subsequently observed the star with hectospec in november 2007 , along with a number of m31 stellar targets as comparisons , selected on the basis of color and magnitude to bracket the properties of b030d .
for the second set of observations we again used the hectospec multi - fiber spectrograph on the mmt @xcite , with exposure time 4800s on target and 2400s off target to supply a local background to subtract from the spectra .
the spectra were reduced in the standard way for fiber spectra ( e.g. , * ? ? ?
* ) , except that those off - target exposures were used as local backgrounds .
the spectra have a resolution of 5.0 and cover the region 3700 - 9200 .
we note that in regions where the bulge of m31 is prominent , such as is the case for b030d , not subtracting a local background can affect the derived velocity by as much as 100 .
velocities were found using template spectra made from the m31 observations described in @xcite .
the spectra were flux - calibrated using spectra of flux standard stars , but for our comparison purposes here , we rectified the spectra using arbitrary but consistent continuum settings .
these comparison spectra help sort out the uncertainty in ubv diagram discussed above .
stars in the `` main sequence '' locus that starts at bottom right and reaches to the center of the ubv plot show the broad balmer lines and weak gravity - sensitive lines typical of dwarf stars .
their velocities are consistent with membership of the thick disk and halo of the milky way , with the redder ( closer ) stars having thick disk kinematics and the bluer ( more distant ) stars having halo kinematics .
stars in the `` supergiant '' locus show narrow strong lines , and the presence of gravity sensitive lines such as ti ii , fe ii@xmath12 , sr ii @xmath13 and the o i@xmath11 triplet .
the velocities of those stars are consistent with membership of the m31 disk .
we selected a few spectra for detailed comparison here , based on the similarity of their ubv colors to b030d s .
these stars are shown in figure [ ubv ] with large symbols .
figures [ members_blue ] and [ nonmembers_blue ] show the blue spectra of b030d compared to these m31 supergiants and milky way dwarfs , respectively .
it can be seen that the balmer lines of b030d are as narrow as those found in the m31 supergiants , distinguishably narrower than the foreground dwarfs . also ,
the ti ii , fe ii@xmath12 and sr ii@xmath13 lines are readily apparent in b030d and difficult to see in the foreground dwarfs .
finally , the ch features contained in the g band are partially resolved in b030d ; where the g band is visible in the foreground dwarfs , it is not resolved .
all of these spectroscopic indicators suggest a low gravity for b030d .
note , however , the relatively weak h@xmath9 line in b030d , compared to the other m31 stars . turning to the red parts of the spectrum ( figures [ members_red ] and [ nonmembers_red ] ) ,
again we see that the o i@xmath11 line is easily visible in b030d and the m31 supergiants , and absent or weak in the foreground dwarfs .
figure [ oi ] shows the results of measuring the equivalent width of that line for all of our high s / n ratio stars , plotted against the ratio of the ca ii h&k line strengths ( as a proxy for spectral type - the index saturates for spectral types later than g , * ? ? ? * ) .
g and k foreground dwarfs form the group around ca ii = 1.25 and o i = 0 .
the m31 f supergiants , along with b030d , clearly separate from the galactic mainsequence stars . in view of our qualitative approach here ,
we do not try to derive an accurate luminosity for b030d using the relationship between m@xmath14 and o i@xmath11 equivalent width derived in @xcite , but we do note that their relation and our o i measurement gives an m@xmath15 consistent with @xmath16 ( see below ) . in summary , the spectroscopy confirms that b030d is a high luminosity star , at the distance of m31 , not a main sequence star from the milky way halo .
however , its lower order balmer lines are unusually weak .
though we have identified the star as an f supergiant , there remains the detail of the weak balmer absorption .
figure [ halpha ] is an enlargement of the h@xmath10 part of the spectrum of b030d and a few comparison m31 f supergiants .
clearly there is blue - shifted balmer emission at h@xmath10 in b030d , and that also explains why the h@xmath9 line is weaker than expected .
since there are no accompanying forbidden lines , and moreover the offset exposures show no emission at all , we must assume the emission is not from the surrounding area but from the star itself , and likely due to a shell .
shell emission is common among the rare , luminous high - latitude galactic f variable stars , such as 89 her @xcite and hr4912 @xcite .
the v magnitude of the star is 19.0 . using the distance modulus of 24.43 @xcite and reddening of 0.13 @xcite , we derive m@xmath17 .
b030d thus lies in an interesting evolutionary state . with an effective temperature of @xmath18 6500 k ( estimated from its spectral type ) and a luminosity of @xmath18 15,000 l@xmath19 ( log l = 4.2 ) .
its position in the hertzsprung gap guarantees that it is evolving rapidly through the hr diagram . here
, we consider four likely options for its evolutionary status and propose new observations that can distinguish among them . _
luminous blue variable ( lbv ) : _ m31 contains @xmath18 20 lbv s , evolved massive stars which vary erratically in brightness and undergo periods of enhanced mass loss ( e.g. , * ? ? ?
* and references therein ) .
although b030d is roughly 10 times fainter than the typical lbv , its optical spectrum resembles spectra of s dor in its cool state @xcite . compared to a ` cool ' s dor , the balmer emission in b030d is weaker and the optical colors much redder .
_ post - main - sequence close binary : _
@xcite proposed that the physical appearance of an evolved close binary can be indistinguishable from an lbv . in this mechanism ,
a main - sequence secondary star accretes material at a high rate from a roche lobe - filling post - main - sequence primary star . a luminous disk surrounding the secondary produces the f - type spectrum and erratic brightness variations ; mass lost from the disk forms an extended shell that produces balmer and other low excitation emission lines .
@xcite noted that this mechanism is a convenient explanation for af and and other lbv s outside the spiral arms of a star - forming galaxy .
the high luminosity in b030d requires an accretion rate of @xmath20 m@xmath21 yr@xmath22 onto a @xmath23 m@xmath24 main - sequence - star .
this high accretion rate can produce an f - type optical spectrum with normal optical colors ( e.g. , the symbiotic binary ci cyg ; * ? ? ? * ) .
_ post - asymptotic giant branch star : _ post - agb stars are high luminosity stars that have recently ejected a planetary nebula .
as they eject their hydrogen envelopes , they evolve at roughly constant luminosity to higher effective temperatures .
although many post - agb stars have much hotter temperatures than b030d , @xcite show f - type optical spectra for several post - agb stars with luminosities similar to the luminosity of b030d .
nearly all post - agb stars are surrounded by extended shells of dust and gas ; thus , these stars often have large infrared excesses from the surrounding dust and radio emission from the gaseous shell . with m@xmath15
= @xmath85.5 , b030d is brighter than the typical post - agb star @xcite .
for the solar metallicity typical of m31 @xcite , stars with initial masses of 4 @xmath25 produce post - agb stars with white dwarf core masses of 0.78 @xmath25 and log l = 4.2 @xcite .
however , the m31 giant stream ( see below ) probably has a lower metallicity of roughly 1/3 solar @xcite . for this metallicity ,
the mass of a post - agb progenitor with log l = 4.2 is roughly 33.5 @xmath25 .
_ symbiotic nova : _ the symbiotic ( or very slow ) novae are eruptive variables in wide binaries with a red giant primary . in these systems ,
the white dwarf secondary accretes matter from the red giant and undergoes occasional thermonuclear runaways . during an optical maximum which can last many decades , a symbiotic nova has an f - type supergiant spectrum with balmer emission lines ( e.g. , pu vul ; * ? ? ?
although symbiotic novae occupy the same region of the h - r diagram as post - agb stars @xcite , they often have much smaller shells of gas and dust . without additional data , choosing
the most likely of these possibilities is difficult .
however , the relatively normal u@xmath8b colors suggest the post - agb and symbiotic nova interpretations are more likely than the lbv - like possibilities .
higher resolution optical spectra would provide a useful diagnostic .
lbv s typically show higher velocity and more variable winds than post - agb stars or symbiotic novae .
many post - agb stars and symbiotic stars are spectroscopic binaries with periods of 1 - 3 yr .
measuring an orbital period from radial velocity data could rule out several alternatives .
infrared spectroscopy should provide a clear discriminant between a post - agb star and a symbiotic nova . in a symbiotic nova ,
infrared spectra show strong co absorption lines characteristic of red giant photospheres ( e.g. , * ? ? ?
in a post - agb star , co features at 2.3 and 4.6 @xmath26 m suggest the much cooler excitation temperatures ( t @xmath18 300 k ) expected from a circumstellar shell ( e.g. , * ? ? ?
* ; * ? ? ?
there is one similar star we have observed in m31 , with h@xmath10 in emission but no other nebular lines : f13633 ( j004025.26 + 403926.9 ) , which has the same luminosity but perhaps lower surface gravity than b030d .
it is located in a one of the spiral arms , and appears to be double , in the lggs image .
we concluded in section 2 that b030d is a high luminosity star at roughly the distance of m31 . in section 4
we went on to discuss four possible options : a very young , massive star type ( a luminous blue variable ) or three stellar types which are associated with older populations ( a post - mainsequence close binary , a post - agb star or a symbiotic nova ) .
projected close to m31 s center but with a 355 velocity difference with disk stars there ( see [ ubvtable ] ) , it would seem more reasonable to associate the star with m31 s halo .
however , studies of the age distribution of m31 s halo show no evidence for very young stars @xcite , so if the star is a luminous blue variable , we need to seek possible mechanisms to accelerate a young disk star to such a high velocity .
there are a number of such possibilities , such as the interaction of a binary with the super - massive black hole in the center of either m31 or m32 , which would disrupt the binary and produce a `` hyper - velocity star '' @xcite , interactions in the core of a dense star cluster , or a `` runaway '' caused by a supernova explosion in a binary system @xcite . while such interactions are likely to be very rare , they are not impossible .
however , there is a simpler , more likely explanation for the star s large velocity which involves a well - studied component of m31 s halo , the giant stream , which was first discovered using star counts by @xcite .
subsequent kinematical studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have shown that the stream has a highly elliptical orbit with perigalacticon very close to m31 s center ( within 2 kpc or less ) and apogalacticon at 125 kpc @xcite .
models fit to the stream have suggested velocities at its perigalacticon as low as v@xmath27 @xcite .
the planetary nebulae studied by @xcite include ones that are possible stream members with v@xmath28 as low as @xmath29 , although this pn has only one line at 4992 ( from a hectospec spectrum of ours ) , thus the possibility that the object is actually at high redshift can not be dismissed . in figure
[ stream ] , we show velocities of pne , drawn from @xcite and our hectospec observations ( paper in preparation ) , along with the orbit of the stream proposed in @xcite . clearly some pne fall along the orbit of the stream , as does b030d . stars with masses of 3 - 4 @xmath25
have main sequence lifetimes much shorter than the youngest stars in the giant stream .
thus , a post - agb member of the giant stream with m@xmath15 = @xmath85.5 is unlikely .
however , b030d could be a post - agb runaway star ejected from the disk of m31 at high velocity . in the milky way
, 5% to 10% of 34 @xmath25 b - type stars are runaways ( stone 1991 ) .
although runaways with velocities of 200@xmath8300 km s@xmath22 relative to the disk are rare @xcite , theoretical simulations produce high velocity runaways at the projected distance of b030d from the center of m31 ( e.g. , the milky way study of * ? ? ?
* ) . in this case
, b030d would have evolved into a post - agb star after ejection as a runaway .
the post - agb possibility seems remote and we conclude that the most likely origin for the star is as part of the giant stream .
@xcite show that the stream contains stars as young as 4 gyr , so the two older options discussed in section 3 ( a post - mainsequence close binary or a symbiotic nova ) are the best possibilities .
b030d is a curiosity . not only does it have an extremely negative velocity , but it is an unusual and rare kind of star as well . in our research into spectroscopic databases
, we could not find any exact analogues that matched the details of its spectrum with regard to the balmer emission .
that problem was certainly mostly due to the lack of modern , digital spectra of the classes of unusual stars described above , but the short life of this kind of f shell star must also be part of the story .
in addition to ir observations , a further piece of data that may help to determine the exact classification would be the star s variability .
we examined the star on both epochs of the palomar sky survey , but could not find a significant change in brightness over that period ( 36 years ) , though the different passbands used for the two surveys makes a comparison difficult .
modern digital images of m31 could of course provide more stringent limits on variability .
we would like to thank dan fabricant for leading the effort to design & build the hectospec fiber positioner and spectrograph , perry berlind & mike calkins for help with the observations , john roll , maureen conroy & bill joye for their many contributions to the hectospec software development project .
we acknowledge important discussions with warren brown , margaret geller & phil massey .
bath , g. t. 1979 , , 282 , 274 battistini , p. , bonoli , f. , braccesi , a. , fusi - pecci , f. , malagnini , m. l. , & marano , b. 1980 , , 42 , 357 bond , h. e. , & fullton , l. k. 1999 , 11th european workshop on white dwarfs , 169 , 439 brown , t. m. , smith , e. , ferguson , h. c. , rich , r. m. , guhathakurta , p. , renzini , a. , sweigart , a. v. , & kimble , r. a. 2006 , , 652 , 323 caldwell , n. , harding , p. , morrison , h. , rose , j. a. , schiavon , r. , & kriessler , j. 2009 , , 137 , 94 climenhaga , j. l. , harris , b. l. , & smolinski , j. 1981 , revista mexicana de astronomia y astrofisica , vol . 6 , 6 , 233 fabricant , d. , et al .
2005 , , 117 , 1411 fardal , m. a. , guhathakurta , p. , babul , a. , & mcconnachie , a. w. 2007 , , 380 , 15 fitzgerald , m. p. 1970
, , 4 , 234 freedman , w. l. , & madore , b. f. 1990 , , 365 , 186 kenyon , s. j. , & truran , j. w. 1983 , , 273 , 280 kenyon , s. j. 1986 , , 91 , 563 kenyon , s. j. , oliversen , n. a. , mikolajewska , j. , mikolajewski , m. , stencel , r. e. , garcia , m. r. , & anderson , c. m. 1991 , , 101 , 637 lambert , d. l. , hinkle , k. h. , & luck , r. e. 1988 , , 333 , 917 lawlor , t. m. , & macdonald , j. 2006 , , 371 , 263 leonard , p. j. t. , hills , j. g. , & dewey , r. j. 1993 , luminous high - latitude stars , 45 , 386 martin , j. c. 2006 , , 131 , 3047 massey , p. 2000 , , 112
, massey , p. et al . 2006 , aj , 131 , 2478 massey , p. , mcneill , r. t. , olsen , k. a. g. , hodge , p. w. , blaha , c. , jacoby , g. h. , smith , r. c. , & strong , s. b. 2007 , , 134 , 2474 mcconnachie , a. w. , irwin , m. j. , ibata , r. a. , ferguson , a. m. n. , lewis , g. f. , & tanvir , n. 2003 , , 343 , 1335 merrett , h. r. , et al .
2003 , , 346 , l62 merrett , h. r. , et al .
2006 , , 369 , 120 mihalas , d. , & binney , j. 1981 , san francisco , ca , w. h. freeman and co. , 1981 .
608 p. , m31 members & & & & & & & & & + b030d & j004141.49 + 410308.2 & 0:41:41.5 & 41:03:08.0 & @xmath30 & @xmath31 & 19.02 & 0.33 & 0.08 & 0.22 & 0.27 + f15459 & j004030.62 + 404523.8 & 0:40:30.6 & 40:45:23.8 & @xmath32 & @xmath33 & 18.33 & 0.29 & 0.14 & 0.18 & 0.22 + y16803 & j004034.00 + 405358.3 & 0:40:34.0 & 40:53:58.3 & @xmath34 & @xmath35 & 18.02 & 0.46 & 0.21 & 0.27 & 0.29 + f13633 & j004025.26 + 403926.9 & 0:40:25.3 & 40:39:26.9 & @xmath36 & @xmath37 & 19.16 & 0.33 & 0.14 & 0.26 & 0.30 + f24367 & j004117.90 + 404845.0 & 0:41:17.9 & 40:48:45.0 & @xmath38 & @xmath39 & 19.10 & 0.37 & 0.03 & 0.25 & 0.29 + y30936 & j004201.09 + 403951.9 & 0:42:01.1 & 40:39:51.9 & @xmath40 & @xmath41 & 18.46 & 0.36 & 0.24 & 0.26 & 0.36 + foreground & & & & & & & & & + a25217 & j004123.97 + 405249.0 & 0:41:24.0 & 40:52:48.7 & @xmath42 & & 16.73 & 0.42 & -0.05 & 0.27 & 0.22 + f25865 & j004128.60 + 410229.6 & 0:41:28.6 & 41:02:29.6 & @xmath43 & & 18.29 & 0.43 & -0.10 & 0.28 & 0.37 + f26766 & j004134.97 + 410537.0 & 0:41:35.0 & 41:05:37.0 & @xmath44 & & 17.94 & 0.50 & -0.07 & 0.32 & 0.34 + f26781 & j004135.14 + 410116.3 & 0:41:35.1 & 41:01:16.3 & @xmath45 & & 16.60 & 0.43 & -0.02 & 0.29 & 0.30 + f44756 & j004252.79 + 404605.6 & 0:42:52.8 & 40:46:05.6 & @xmath46 & & 17.75 & 0.39 & 0.06 & 0.23 & 0.29 + | we report on a single star , b030d , observed as part of a large survey of objects in m31 , which has the unusual radial velocity of @xmath0 . based on details of its spectrum , we find that the star is an f supergiant , with a circumstellar shell
. the evolutionary status of the star could be one of a post - mainsequence close binary , a symbiotic nova , or less likely , a post - agb star , which additional observations could help sort out .
membership of the star in the andromeda giant stream can explain its highly negative velocity . |
experimentally , in _ pp _ and _ pp _ collisions , not only the average particle multiplicity is increasing with energy but , as well , the normalised width of the multiplicity distribution is an increasing function of energy @xcite . in other words , both @xmath2 and @xmath3 increase with energy .
the increase of @xmath2 is expected : more particles are produced when more energy is available .
the growth of @xmath4 is , however , not so obvious .
the general argument seems to be that the increase of energy implies phase - space opening .
there is more room for partons , strings , clusters or clans , more room for multiple collisions and different impact parameters contributions , more room for harder collisions and additional qcd branching @xcite . in all these approaches ,
if there are no limitations , fluctuations increasingly dominate over average multiplicity : @xmath4 increases and , asymptotically , it may reach a constant value ( kno scaling @xcite ) .
this kind of behaviour , as pointed out in @xcite , has been observed long ago in cosmic ray physics @xcite , with the negative binomial parameter @xmath5 stabilising at a value of the order of 2.53 , for @xmath6 tev .
however , if saturation phenomena occur @xcite , i.e. , too many objects occupy a finite region of interaction , then fluctuations are severely limited and @xmath4 will start to decrease . on general grounds ,
this is not unreasonable . as the number of intermediate objects ( or final particles )
is bounded by @xmath7 ( conservation of energy ) , while the region of interaction is constrained by the stronger froissart bound , @xmath8 , there exists a real possibility for saturation .
it is also well known that , at fixed energy , when triggering on rare events @xcite or selecting central impact parameter , @xmath9 , collisions @xcite , the kno distribution , due to the higher density , is narrower .
we shall work in the framework of the dual string model @xcite .
saturation , in this case , results from clustering and percolation of strings . in a collision , @xmath10 strings are produced and they may form clusters with different number of strings ( or different sizes ) .
these clusters act as random , particle emitting , sources .
following @xcite , we write for the multiplicity distribution , @xmath11 , @xmath12 where @xmath13 is a normalised weight function , @xmath14 is a clustering variable , @xmath15 a convolution function for a cluster of strings , @xmath16 being the single string average multiplicity . we assume for @xmath15 a poisson distribution ,
this being consistent with the single string distribution of low energy @xmath17annihilations .
note that , as emphasised in @xcite , an equation similar to ( [ eq:1 ] ) , with the same function @xmath18 , can be written for the inclusive @xmath19 distribution , @xmath20 . from ( [ eq:1 ] )
we obtain @xmath21 and @xmath22 defining now @xmath23 as the number of clusters , @xmath24 the number of strings in a cluster and @xmath25 the average number of strings , which must satisfy the sum rule @xmath26 , we shall write @xmath27 such that , finally , we obtain @xmath28 and @xmath29 the relevant parameter in string clustering ( and percolation ) is the transverse density @xmath30 , @xmath31 where @xmath32 is the transverse radius of the string ( we shall use @xmath33 fm ) and @xmath34 is the effective radius of the interaction region , with @xmath35 fm in _
_ collisions .
it is the parameter @xmath30 that controls clustering : when @xmath24 strings cluster , the effective colour charge is not proportional to @xmath24 but , due to the vectorial nature of colour summation @xcite , is proportional to @xmath36 , where @xcite @xmath37 such that , instead of ( [ eq:8 ] ) , one should write @xmath38 equation ( [ eq:9 ] ) for @xmath0 remains , of course , unchanged .
note that @xmath39 is a decreasing function of @xmath30 .
in order to show , clearly , the problem of the competition between the opening of new possibilities and limitations due to saturation , let us introduce a very simple , and known to everybody , toy model .
imagine that one has @xmath10 identical coins to distribute over @xmath40 identical boxes .
the reader may imagine the distribution of @xmath10 partons or strings over the interaction region of size @xmath40 , or @xmath41 ( in units of @xmath42 ) .
this is a purely combinatorial problem , without percolation . what is the distribution of coins in the boxes ? in ref .
@xcite it was shown that in the thermodynamical limit only the transverse density @xmath43 is relevant , and that the first moments of the distribution @xmath44 of the number of strings per cluster ( the probability of having a @xmath24-cluster ) @xmath45 and @xmath46 if we use this toy model as input for @xmath18 in ( [ eq:1 ] ) , and take , as mentioned before , the poisson distribution for @xmath47 , then , see ( [ eq:9 ] ) , one obtains @xmath48 in figure [ fig:1 ] we present plots of ( [ eq : toy8 ] ) and ( [ eq : toy9 ] ) .
the average multiplicity is , naturally , a growing function of the density .
the parameter @xmath5 goes to infinity at small @xmath30 ( one just has 1-clusters ) , decreases and reaches a minimum , @xmath49 , at @xmath50 , then moves to infinity again as @xmath51 ( the distribution is dominated by @xmath52-clusters ) .
we have checked by monte carlo simulation that the distribution @xmath44 of the model is similar to a gamma distribution .
in order to make contact with real physics , we need to estimate the dependence of the parameter @xmath5 , experimentally measured in _ pp_collisions as a function of the energy , on the density parameter @xmath30 . in other words
, we have to find the energy dependence of @xmath30 . as @xmath30 , see ( [ eq:10 ] ) , depends on @xmath10 and @xmath41 , we need to find the energy dependence of @xmath10 and @xmath41 . for the _ pp _ average charged particle multiplicity in the dual string model , we shall write @xcite : @xmath53 , \label{eq : phys12}\ ] ] where @xmath54 is the single valence string average multiplicity ( assumed constant ) , @xmath55 the sea string multiplicity reduction factor ( @xmath56 ) .
notice that in ( [ eq : phys12 ] ) one always has 2 valence strings and @xmath57 sea strings . at low energy ,
@xmath58 gev , experimentally @xmath59 , and theoretically @xmath60 ( the two valence strings ) and of course @xmath61 and @xmath62 .
notice that ( [ eq : phys12 ] ) is a natural generalisation of ( [ eq:13 ] ) when there are two kinds of strings . as the slope parameter @xmath63 is a direct measure of the square of the interaction radius @xmath64
, we write , from ( [ eq:10 ] ) , @xmath65 with @xmath66 , such that at @xmath67 gev , @xmath35 fm and @xmath68 . by using equations (
[ eq : phys12 ] ) and ( [ eq : phys13 ] ) , and experimental information on @xmath69 @xcite and on @xmath63 @xcite , we can in principle construct @xmath70 and @xmath71 : from @xmath71 we obtain @xmath72 , eq . ( [ eq : toy9 ] ) . in fig .
[ fig:2]a we show our fit to @xmath63 and in fig .
[ fig:2]b our fit to @xmath69 .
these parameterisations are then used in eq .
( [ eq : phys13 ] ) and eq .
( [ eq : phys12 ] ) . intuitively , one expects the number of strings , @xmath10 , to increase with @xmath7 .
the results from our fits are shown in fig .
[ fig:3 ] : @xmath10 increases with energy and , within a factor 2 , is not substantially different from monte carlo direct calculations @xcite .
naively , as @xmath73 , one also expects @xmath30 to grow with energy .
however , as the interaction area also grows with energy , the situation is not so clear . from ( [ eq : phys12 ] ) and ( [ eq : phys13 ] ) it is easily seen that the @xmath30 dependence on @xmath7 , @xmath71 , is determined by the ratio @xmath74 . in figure [ fig:3]b we present @xmath71 ; in the figure
is shown as well the percolation threshold : @xmath75 . finally , in fig . [ fig:3]c
, we show the dependence of @xmath5 on energy , see ( [ eq : toy9 ] ) and fig .
[ fig:3]b , in comparison with experimental points .
if the `` coins in boxes '' model reflects the physical situation , we expect @xmath5 to reach a minimum at @xmath76 tev , in the lhc energy region .
the behaviour we predict is qualitatively different from what is expected on the basis of extrapolations of other phenomenological models : for instance , consider the wrblewski regularity @xcite : @xmath77 ; this leads to an asymptotic ( @xmath78 ) kno scaling , thus to a levelling off of @xmath5 ( see dashed line in fig . [ fig:3]c . ) other proposals ( e.g. @xmath79 in @xcite , and the pythia monte carlo @xcite , respectively dotted line and cross in fig .
[ fig:3]c ) give a monotonically decreasing @xmath5 ; for a discussion of the situation in which @xmath5 continues to decrease at high energy and reaches values smaller than 1 , see @xcite .
we have argued that the energy behaviour of the parameter @xmath5 ( or @xmath80 ) of the multiplicity distribution in _ pp _ collisions is sensitive to the energy dependence of the parton / string transverse density @xmath30 .
if impact parameter overcrowding never occurs ( @xmath30 is always small ) @xmath5 will decrease with energy , reaching , eventually , some asymptotic constant value .
if overcrowding takes place ( @xmath30 increases rapidly with energy ) then @xmath5 necessarily will change its behaviour and start increasing with energy .
the turning point may occur at an energy of the order of 510 tev .
it was shown in @xcite that the parameter @xmath5 of the multiplicity distribution is the parameter @xmath5 of the @xmath81-distribution : @xmath82 .
this parameter @xmath5 is also a decreasing function of energy in the @xmath83 tev region . in figure [ fig:4 ]
we have compared experimental data on @xmath81 distributions to the mentioned parameterisation , with @xmath5 given by the multiplicity distribution values at the same energy .
the agreement is reasonable , and improving with energy .
r. ugoccioni would like to thank for discussions a. prokudin on elastic collisions and m. monteno on monte carlo tuning .
e.g. ferreiro and c. pajares thank the financial support of the cicyt of spain through the contract fpa2002 - 01161 and of xunta de galicia through the contract pgidit03pxic20612pn , physics reports * 236 * , 225 ( 1994 ) ; b. andersson , _ the lund model _
( cambridge university press , cambridge , u.k . ,
1996 ) ; a. giovannini and r. ugoccioni , phys .
d * 59 * , 094020 ( 1999 ) ; s.g .
matinyan and w.d .
walker , phys .
rev . * d59 * , 034022 ( 1998 ) ; a. giovannini and l. van hove , z. phys .
c * 30 * , 391 ( 1986 ) ; yu.l .
dokshitzer , v.a .
khoze , a.h .
mueller and s.i .
troyan , _ basics of perturbative qcd _ ( editions frontires , gif - sur - yvette , 1991 ) . ,
physics reports * 100 * , 1 ( 1983 ) ; a.h . mueller and j. qiu , nucl .
* b268 * , 427 ( 1986 ) ; l. mclerran and r. venugopalan , phys .
d * 49 * , 2233 ( 1994 ) ; phys .
d * 49 * , 3352 ( 1994 ) ; yu.v .
kovchegov , phys . rev .
d * 54 * , 5463 ( 1996 ) .
* b306 * , 312 ( 1993 ) ; z. phys .
* c63 * , 507 ( 1994 ) ; m.a .
braun and c. pajares , eur .
j. * c16 * , 349 ( 2000 ) ; phys .
lett . * 85 * , 4864 ( 2000 ) ; n. armesto and c. pajares , int .
j. mod .
* a15 * , 2019 ( 2000 ) . ,
b * 109 * , 124 ( 1982 ) ; m. adamus et al .
( na22 collaboration ) , phys .
b * 186 * , 223 ( 1987 ) ; n. amos et al . ( isr ) , nucl .
b * 262 * , 689 ( 1985 ) ; r.e .
breedon et al .
( ua6 collaboration ) , phys .
b * 216 * , 459 ( 1989 ) ; f. abe et al .
( cdf collaboration ) , phys .
d * 50 * , 5518 ( 1994 ) .
t. sjstrand _
et al . _ ,
. commun . * 135 * , 238 ( 2001 ) ; parameters tuned as detailed in c.m .
et al . _ , acta phys .
pol . * b35 * , 433 ( 2004 ) ; a. moraes in _ workshop mc4lhc _
( cern , 31 july 2003 ) , and _ physics at lhc _ ( prague 2003 ) .
the actual pythia 6.214 point in the figure is from m. monteno ( private communication ) . | it is experimentally observed that the width of the kno multiplicity distribution or the negative binomial parameter , @xmath0 for _ pp _ collisions , in the energy region @xmath1 gev , is an increasing function of the energy .
we argue that in models with parton or string saturation such trend will necessary change : at some energy the distribution will start to become narrower . in the framework of percolating strings , we have estimated the change to occur at an energy of the order of 510 tev . , , , |
extending some previous work @xcite , @xcite , @xcite , @xcite , @xcite on the rationality of the exponential transform we here go on to consider what we believe is the most general type of domains for which the exponential transform , in an extended sense , can be expected to have a `` core '' consisting of a rational function . around this core
there will then be `` satellites '' of some rather trivial factors , depending on the regimes at the locations of the independent variables .
the `` domains '' we consider will actually be covering surfaces over the riemann sphere .
the terminology `` quadrature riemann surface '' has already been introduced by m. sakai @xcite for the type of domains in question .
an alternative name could be `` multi - sheeted algebraic domain '' , to extend a terminology used by a. varchenko and p. etingof @xcite . in the present work we shall mostly use the latter terminology because we will not emphasize so much the quadrature properties , but rather take as a starting point the way the domains are generated , namely by a pair of meromorphic function on a compact riemann surface . given a bounded domain @xmath0 , the traditional exponential transform @xcite , @xcite , @xcite , @xcite , @xcite of @xmath1 is the function of two complex variables defined by @xmath2 \quad ( z , w\in \c).\ ] ] for the unit disk @xmath3 it is ( see @xcite ) : @xmath4 using the schwarz function @xcite , @xcite for @xmath5 , @xmath6 ( [ et3 ] ) can be written @xmath7 where @xmath8 is the rational function @xmath9 and @xmath10 is the characteristic function of @xmath11 . the expression ( [ expelimd ] )
reveals the general structure of the exponential transform of any algebraic domain .
the function @xmath8 is an instance of the elimination function , which can be defined by means of the meromorphic resultant @xmath12 of two meromorphic functions @xmath13 and @xmath14 . the resultant is defined as the multiplicative action of @xmath14 on the divisor @xmath15 of @xmath13 , namely @xmath16 , and the elimination function is @xmath17 . in the case of the unit disk , or any quadrature domain , the relevant elimination function which enters into the exponential transform is the one with @xmath18 , @xmath19 .
this means that we in the present example get @xmath20 @xmath21 as desired .
the aim of the present paper is to generalize the formula ( [ expelimd ] ) as far as possible .
this will involve an extended exponential transform in four complex variables and an analogous extended elimination function in four variables , defined in terms of a conjugate pair of meromorphic functions on a fairly general compact symmetric riemann surface .
the paper is organized as follows .
sections [ sec : cauchy ] and [ sec : resultant ] contain general preliminary material . in section [ sec : multi ] we introduce the concepts of multi - sheeted algebraic domains and quadrature riemann surfaces and prove that they are equivalent . the main result is stated in section [ sec : main ] and proved in section [ sec : proof ] .
section [ sec : example ] , finally , is devoted to examples , namely the ellipse and neumann s oval .
the cauchy transform of a bounded density function @xmath22 in @xmath23 is @xmath24 typically the functions @xmath22 which will appear in this paper will be like the characteristic function of a domain , or the corresponding integer valued counting function for a multi - sheeted domain . from @xmath25 , the density @xmath22 can be recovered by @xmath26 to be interpreted in the sense of distributions .
if one writes the definition of the cauchy transform as @xmath27 one realizes that it suffers from a certain lack of symmetry .
a more balanced object would be the `` double cauchy transform '' , @xmath28 in fact , this double transform is much richer than the original transform , and after exponentiation it gives the by now quite well studied @xcite , @xcite , @xcite , @xcite , @xcite ( etc ) exponential transform : @xmath29 the original cauchy transform can be recovered as @xmath30 at least if @xmath22 vanishes in a neighborhood of infinity .
one disadvantage with the double cauchy transform is that the formula ( [ dbarcauchy ] ) turns into the more complicated @xmath31 on the other hand we have the somewhat nicer looking @xmath32 where @xmath33 denotes the dirac distribution .
now , even the double cauchy transform is not entirely complete .
it contains the cauchy kernel @xmath34 , which is a meromorphic differential on the riemann sphere with a pole at @xmath35 , but it has also a pole at @xmath36 .
it is natural to make the latter pole visible and movable .
that would have the additional advantage that one can avoid the two cauchy kernels which appear in the definitions of the double cauchy transform and the exponential transform to have coinciding poles ( namely at infinity ) .
thus we arrive naturally at the extended cauchy and exponential transforms : @xmath37 @xmath38 if the points @xmath39 , @xmath40 , @xmath41 , @xmath42 are taken to be all distinct , then both transforms are well defined and finite for any bounded density function @xmath22 on the riemann sphere .
for example , with @xmath43 , @xmath44 turns out to be the modulus squared of the cross - ratio .
see @xcite for further details .
here we shall briefly review the definitions of the meromorphic resultant and the elimination function , as introduced in @xcite , referring to that paper for any details . if @xmath13 is a meromorphic function on any compact riemann surface @xmath45 we denote by @xmath15 its divisor of zeros and poles , symbolically @xmath46 .
if @xmath47 is any divisor and @xmath14 is a meromorphic function we denote by @xmath48 the multiplicative action of @xmath14 on @xmath47 .
for example , if @xmath49 , @xmath50 , then @xmath51 .
now the meromorphic resultant between @xmath13 and @xmath14 is , by definition , @xmath52 whenever this makes sense .
the elimination function is @xmath53 where @xmath54 are parameters .
it is always a rational function in @xmath39 and @xmath40 , more precisely of the form @xmath55 where @xmath56 , @xmath57 and @xmath58 are polynomials , and it embodies the necessary ( since @xmath45 is compact ) polynomial relationship between @xmath13 and @xmath14 : @xmath59 we also have the extended elimination function , defined by @xmath60 to relate the elimination function to the exponential transform one needs integral formulas for the elimination function .
if @xmath13 is meromorphic on @xmath45 with divisor @xmath15 , let @xmath61 be a @xmath62-chain such that @xmath63 and such that @xmath64 has a single - valued branch , which we denote @xmath65 , in @xmath66 .
then @xmath65 can be viewed as a distribution on @xmath45 , and its exterior differential in the sense of distributions ( or currents ) is @xmath67 here @xmath68 is the @xmath62-form current supported by @xmath61 and defined locally , away from @xmath69 , as the differential ( in the sense of currents ) of that function @xmath70 which is @xmath71 on the right - hand side of @xmath61 , zero on the left - hand side .
globally @xmath68 is not exact ( despite the notation ) , not even closed . to be precise , @xmath72 where @xmath73 denotes the finite distribution of point masses ( or charges ) corresponding to @xmath15 .
we shall also need the fact that @xmath68 has the period reproducing property @xmath74 holding for any smooth @xmath62-form @xmath75 .
now we have ( essentially theorem 2 in ) @xmath76.\ ] ] the integrand is a @xmath77-form current with support on the @xmath62-chains @xmath78 and @xmath79 ( because away from these curves the integrand contains @xmath80 ) , so the integral is rather a line integral than an area integral .
in fact , the above can also be written @xmath81,\ ] ] which perhaps clarifies the connection to the definition of the elimination function .
the boundary of a quadrature domain ( algebraic domain ) is an algebraic curve , but by no means every algebraic curve arises in this way .
however , the gap between the two classes of objects can be reduced considerably by extending the notion of a quadrature domain , allowing it to have several sheets and to be branched over the riemann sphere .
this will take essentially one half of all algebraic curves into the framework of quadrature domains and exponential transforms .
one step in this direction was taken in sakai , where a notion of quadrature riemann surface was introduced in a special case .
below we shall take some further steps .
let @xmath45 be any symmetric compact ( closed ) riemann surface .
slightly more generally , we shall allow @xmath45 to be disconnected , namely to be a finite disjoint union of riemann surfaces .
the symmetry means that @xmath45 is provided with an anticonformal involution @xmath82 , @xmath83 .
if @xmath45 is disconnected then @xmath84 is allowed to map one component of @xmath45 onto another .
let @xmath85 denote the set of fixed points of @xmath84 .
simple examples of symmetric riemann surfaces are @xmath86 ( the riemann sphere ) with the involution being either @xmath87 or @xmath88 . in the first case
@xmath89 and @xmath90 has two components , in the second case @xmath85 is empty and hence @xmath90 has only one component .
one can also think of identifying the points @xmath91 and @xmath92 .
the identification spaces become , in the first case ( @xmath93 ) the unit disk together with its boundary , and in the second case ( @xmath94 ) the projective plane , thus a nonorientable surface . in general , the orbit space @xmath95 , obtained by identifying @xmath91 and @xmath92 for any @xmath96 , is a klein surface , possibly with boundary . a klein surface @xcite
is defined in the same way as a riemann surface except that it is allowed to be nonorientable and that both holomorphic and antiholomorphic transition functions between coordinates are allowed .
the possible boundary points of @xmath97 are those coming from @xmath85 under the identification . from the klein surface @xmath97 , @xmath45 can be recovered by a natural doubling procedure , described in section 2.2 of @xcite and in @xcite , for example .
the latter reference actually describes several types of doubles ( the complex double , the orienting double and the schottky double ) , but the description in @xcite will be enough for our purposes . in case
@xmath97 is orientable and has a boundary the doubling procedure gives what is usually called the schottky double , named after the inventor of the idea , f. schottky @xcite .
the idea was later extended to more general surfaces by f. klein @xcite . continuing the discussion of @xmath45 , @xmath84 and @xmath97 ,
if @xmath45 is connected but @xmath98 disconnected , then @xmath98 has exactly two components , say @xmath99 and @xmath100 , and @xmath84 maps each of them onto the other .
it then follows that @xmath101 can be identified with @xmath99 ( or @xmath100 ) , and in particular that @xmath97 is orientable , hence is ( after choice of orientation ) an ordinary riemann surface with boundary .
if @xmath98 is connected then @xmath97 necessarily is nonorientable .
it is relevant to allow @xmath45 to have several components .
for example , in section [ sec : example ] we will encounter the double of the riemann sphere , which simply is two riemann spheres with the opposite conformal structure and with @xmath84 mapping one onto the other .
now to the definition of `` multi - sheeted algebraic domain '' .
there are two ingredients .
the first is a compact symmetric riemann surface @xmath102 such that @xmath45 is connected and such that @xmath90 has two components , one of which , call it @xmath99 , is to be selected .
we could equally well have started with @xmath99 , to be any riemann surface with boundary ( bordered riemann surface ) , and then let @xmath45 be the double of @xmath99 .
the second ingredient is a nonconstant meromorphic function @xmath13 on @xmath45 .
[ def : multi ] a _ multi - sheeted algebraic domain _ is ( represented by ) a pair @xmath103 , where @xmath99 is a bordered riemann surface and @xmath13 is a nonconstant meromorphic function on the double @xmath45 of @xmath99 .
two pairs , @xmath103 and @xmath104 are considered the same if there is a biholomorphic mapping @xmath105 such that @xmath106 and @xmath107 .
the equivalence simply means that it is the image @xmath108 , with appropriate multplicities , which counts .
note that also @xmath109 is a multi - sheeted algebraic domain , if @xmath103 is .
trivial examples of a multi - sheeted algebraic domain are obtained by taking @xmath110 and @xmath111 .
then with @xmath13 any nonconstant rational function @xmath112 will be a multi - sheeted algebraic domain , as well as @xmath113 .
some further examples will be discussed in section [ sec : example ] .
along with @xmath13 , meromorphic on @xmath45 , the symmetry @xmath84 provides one more meromorphic function on @xmath45 , namely @xmath114 with @xmath115 the polynomial in ( [ qpr ] ) for @xmath116 , the map @xmath117 parametrizes the curve @xmath118 ( or , better , its projective counterpart ) .
this parametrization is one - to - one ( except for finitely many points ) if and only if @xmath13 and @xmath119 generate the field of meromorphic functions on @xmath45 ( form a primitive pair in the terminology of @xcite )
. this will generically be the case if @xmath13 is chosen `` at random '' , but there are certainly many counterexamples .
for example , @xmath13 may be already symmetric in itself , i.e. , @xmath120 , and then @xmath13 and @xmath119 is a primitive pair only if @xmath45 is the riemann sphere and @xmath13 is a mbius transformation
. if @xmath13 and @xmath119 are not a primitive pair , then the polynomial @xmath115 in ( [ qpr ] ) is reducible . in case
@xmath99 is planar ( i.e. , is topologically equivalent to a planar domain ) and @xmath13 is univalent on @xmath99 without poles on @xmath121 , then @xmath122 is an ordinary algebraic domain , in other words a classical quadrature domain . in this case @xmath13 and @xmath119 do form a primitive pair ( see @xcite ) .
the reference to `` quadrature '' can be explained in terms of a residue calculation . in the present generality
it is natural to use the spherical metric on the riemann sphere in place of the customary euclidean metric .
this will remove some integrability problems , and the point of infinity can be treated on the same footing as other points .
quadrature domains for the spherical measure have been previously discussed , at least in fluid dynamic contexts , for example in @xcite ( hele - shaw flow ) , @xcite ( vortex patches ) .
let @xmath123 be a function holomorphic in a neighborhood of @xmath121 and assume , for simplicity , that @xmath13 has no poles on @xmath124 ( such poles will actually cause no problems anyway ) . at least away from poles of @xmath13 we have , by exterior differentiation , @xmath125 therefore , if @xmath13 has poles of orders @xmath126 at @xmath127 then @xmath128 @xmath129 @xmath130 here it turns out that the terms @xmath131 cancel with corresponding terms ( with negative sign ) in the residue contribution unless @xmath119 happens to have zeros at the points @xmath132 . in any case , with or without cancellations , the right member above is of the form @xmath133 i.e. , equals the action on @xmath123 by a distribution with support in finitely many points .
the quadrature nodes @xmath134 are the poles of @xmath135 .
if @xmath13 is univalent the above identity becomes an ordinary quadrature identity ( although for the spherical measure ) of the form @xmath136 holding for all integrable analytic functions @xmath14 in @xmath122 . here the right member is given by @xmath137 , which still is the action of a distribution with finite support .
when @xmath13 is not univalent one should still think of the quadrature identity in the same way as in ( [ quadratureidentity ] ) , just with the difference that @xmath1 is a region with several sheets over the riemann sphere .
the test functions should be allowed to take different values at points lying above one and the same point , but on different sheets .
thus those of the form @xmath138 ( i.e. , those which would become @xmath139 in a formulation like ( [ quadratureidentity ] ) ) are too special .
this is most easily expressed by pulling everything back to @xmath99 , in which case we simply have the originally obtained identity @xmath140 for @xmath123 holomorphic in a neighborhood of @xmath121 , and by approximation for functions @xmath123 holomorphic and integrable ( with respect to @xmath141 ) in @xmath99 .
an equivalent formulation is that there exists a positive divisor @xmath47 in @xmath99 such that @xmath142 holds for every holomorphic and integrable function @xmath123 in @xmath99 with @xmath143 . a _ quadrature riemann surface ( for the spherical metric )
_ is a pair @xmath103 , where @xmath99 is a bordered riemann surface and @xmath13 is a nonconstant meromorphic function on @xmath99 such that @xmath144 and such that ( [ qi ] ) ( or ( [ qid ] ) ) holds for some @xmath145 ( respectively @xmath47 ) and the indicated classes of functions @xmath123 .
the notion of equivalence between pairs is the same as in definition [ def : multi ] .
if @xmath13 is meromorphic on @xmath45 then we have @xmath146 where @xmath147 is the order of @xmath13 .
thus we have proved one direction of the following .
a pair @xmath103 is a multi - sheeted algebraic domain if and only if it is a quadrature riemann surface .
it remains to prove that a quadrature riemann surface is a multi - sheeted algebraic domain .
in many special cases this has already been done ( for the euclidean metric ) , see for example @xcite , @xcite , @xcite .
we shall discuss here the general case under the simplifying assumption that @xmath13 is meromorphic in a neighborhood of @xmath148
. it will be convenient to introduce some further notation . with @xmath149 any subset and @xmath47 any divisor on @xmath45
we denote by @xmath150 the set of functions @xmath123 meromorphic in a neighborhood of @xmath151 and with @xmath143 .
similarly , @xmath152 denotes the set of @xmath62-forms @xmath153 , meromorphic in a neighborhood of @xmath151 and satisfying @xmath154 .
we shall also use standard notations for cohomology groups .
now , returning to the previous residue calculation one realizes that what needs to be proven is that if @xmath155 holds for all @xmath156 , for some sufficiently large divisor @xmath47 , then @xmath13 extends to a meromorphic function on @xmath45 .
on @xmath85 we have @xmath157 where the right member is holomorphic in a neighborhood of @xmath85 and can be viewed as representing an element in the cohomology group @xmath158 .
when @xmath47 is strictly positive this group is trivial since , by serre duality @xcite , @xcite , @xmath159 ( here star @xmath160 denotes dual space ) , hence there exist @xmath161 such that @xmath162 in a neighborhood of @xmath85 .
clearly @xmath163 for @xmath156 since the integrand is holomorphic in @xmath99 , hence ( [ qibdryd ] ) reduces to the statement that @xmath164 for @xmath156 . at this point
we may use a general duality theorem ( also related to serre duality ) going back to the work of j. silva @xcite and g. kthe @xcite , extended by a. grothendieck @xcite and , in the form we need it , c. auderset @xcite .
it states that the bilinear form in @xmath123 and @xmath165 defined by the left member of ( [ duality ] ) induces ( with the path of integration moved slightly into @xmath100 ) a non - degenerate pairing @xmath166 which exhibits each of the quotient spaces as the dual space of the other . in view of this ( [ duality ] )
implies that @xmath167 , in particular that @xmath168 since @xmath13 is meromorphic in a neighborhood of @xmath121 , ( [ omega ] ) implies that also @xmath119 is meromorphic there , hence that @xmath13 actually is meromorphic on all @xmath45 , as was to be proved . the usage of the general duality theorems above can be replaced by more direct arguments , like applying ( [ qi ] ) to suitable cauchy kernels .
specifically we may choose , with @xmath169 , @xmath170 where the right member is a kernel which in the case of the riemann sphere is the usual cauchy kernel @xmath171 and which has counterparts with good enough properties on all compact riemann surfaces ( see @xcite ) .
the point @xmath172 is needed in higher genus .
actually the duality theorem discussed above can be proved using this kernel .
let @xmath45 , @xmath99 , @xmath173 , @xmath13 be as in section [ sec : multi ] , more precisely such that @xmath103 is a multi - sheeted algebraic domain . to account for the multiplicities of @xmath108 as a covering of the riemann sphere we introduce the integer - valued counting function , or mapping degree , @xmath174 pointwise well - defined for @xmath175 .
it is understood that points @xmath91 are counted with the appropriate multiplicities .
set also @xmath176 this is a multi - valued algebraic function in the complex plane which contains all local schwarz functions of @xmath177 , because for @xmath178 one of the values of @xmath179 is @xmath180 .
we will have to make expressions like @xmath181 well - defined , i.e. , single - valued , despite @xmath179 itself being multi - valued .
the natural definition is the following : @xmath182 here @xmath183 denotes the restriction of the divisor @xmath184 to @xmath99 and the right member then is the multiplicative action of @xmath185 on @xmath183 . to spell it out ,
let @xmath186 with repetitions according to multiplicities .
then @xmath187 which is a natural definition in view of ( [ schwarz ] ) .
clearly @xmath188 is an analytic function of @xmath39 in regions where @xmath189 is constant .
now , for the main result , we have two functions which we want to relate to each other : one is the weighted exponential transform @xmath190,\ ] ] @xmath191,\ ] ] which can be viewed as a kind of potential of @xmath22 , and the other is the elimination function , which is defined by algebraic means and always is a rational function , namely of the form @xmath192 the latter expression comes form ( [ qpr ] ) together with the observation that the one variable polynomials cancel in the four variable case .
the nature of @xmath193 depends on the locations of the points @xmath39 , @xmath40 , @xmath41 , @xmath42 , more precisely on the the values of @xmath22 at these points .
the main result is the following .
[ thm : emathcale ] let @xmath103 be a multi - sheeted algebraic domain .
then , in the above notations , @xmath194 @xmath195 @xmath196
extending ( [ dbarcauchy1 ] ) to four variables gives @xmath197 which tells that the function @xmath198 is analytic in @xmath39 in regions where @xmath189 is constant , namely in each component of @xmath199 .
hence so is the exponential of it , namely @xmath200 augmenting this we have that also the function @xmath201 is analytic in @xmath199 , away from poles caused by the presence of @xmath179 .
now we claim that it is even better than that : @xmath202 is meromorphic everywhere , hence is a rational function . to prove this it is enough to prove that @xmath202 is continuous across @xmath177 .
it is well - known that @xmath203 is continuous in @xmath39 ( see for example @xcite ) , so we only have to bother about the other factor .
but it is easy to realize that also this is continuous : spelling out as in ( [ schwarzrho ] ) and assuming for example that @xmath189 increases by one unit as @xmath177 is crossed at a certain place we find that the factor @xmath204 changes from @xmath205 to @xmath206 which clearly is a continuous change since the new point @xmath207 starts up on @xmath85 .
similarly for the factor @xmath208 repeating the above argument for @xmath40 , @xmath41 , @xmath42 it follows that the function @xmath209 @xmath210 is rational in the variables @xmath39 @xmath211 , @xmath41 , @xmath212 .
thus , since also @xmath213 is rational in these variables , it is enough to prove that the formula in the statement of the theorem holds just locally , somewhere .
we may then choose @xmath39 , @xmath40 , @xmath41 , @xmath42 close to each other , so that in particular @xmath214 .
in addition we may assume that this value is the smallest value of @xmath22 occurring on @xmath215 .
the case that it is zero can be treated exactly as in the proof of theorem 6 in @xcite , which concerns the special case @xmath216 .
so let us for example assume that @xmath217 ( the general case will be similar ) .
thus @xmath13 attains the values @xmath39 , @xmath40 , @xmath41 , @xmath42 exactly once in @xmath99 , and these four points on the riemann sphere are close to each other .
let @xmath218 be an arc in @xmath215 from @xmath42 to @xmath40 ( e.g. , the geodesic arc ) .
then the function @xmath219 has a single - valued branch , call it @xmath220 , in @xmath221 , hence @xmath222 is single - valued in @xmath223 .
we may consider @xmath224 as a @xmath62-chain , and as such it has the same role for @xmath225 as @xmath61 has for @xmath65 in ( [ dlogf ] ) , so that @xmath226 if @xmath13 has degree @xmath147 , then @xmath224 consists of @xmath147 small arcs , one of which is located on @xmath99 .
let @xmath227 be that arc and let @xmath228 be the reflected arc in @xmath100 . in the sequel we shall use @xmath229 in the restricted sense @xmath230 .
thus @xmath231 @xmath232 using ( [ elimint ] ) , ( [ dlogf ] ) and ( [ inttau ] ) we now get @xmath233\ ] ] @xmath234\cdot\exp[\int_{\tilde{\sigma}}(\frac{df}{f - z}-\frac{df}{f - a})].\ ] ] here we start by rewriting the last factor according to @xmath235_{\widetilde{f^{-1}(b)}}^{\widetilde{f^{-1}(w)}}\ ] ] @xmath236 note that functions like @xmath237 are single - valued in the present case ( cf .
( [ schwarzrho ] ) ) .
next , the first factor can be integrated partially , to become @xmath238 @xmath239\cdot\ ] ] @xmath240\ ] ] @xmath239\cdot\ ] ] @xmath241\ ] ] @xmath239 \cdot\frac{s(z)-\bar w}{s(z)-\bar b } \cdot\frac{s(a)-\bar b}{s(a)-\bar w}.\ ] ] we have to rework all expressions which are not yet in the form appearing in the statement of the theorem .
so we next turn our attention to the first factor in the last obtained expression .
this can be rewritten as @xmath242\ ] ] @xmath243\ ] ] @xmath244\cdot\ ] ] @xmath245\ ] ] @xmath246\cdot\ ] ] @xmath247\cdot\ ] ] @xmath248 @xmath249\cdot\ ] ] @xmath250\ ] ] @xmath251 now putting all the pieces together we obtain the formula in the statement of the theorem .
let @xmath47 denote the domain inside the ellipse @xmath252 with semiaxes @xmath253 and foci @xmath254 .
the exterior domain @xmath255 is known to be a null quadrature domain @xcite for the euclidean metric . for the spherical metric
it is a two point quadrature domain .
indeed , the schwarz function for the ellipse is @xmath256 and for @xmath123 holomorphic in @xmath257 and smooth up to the boundary we have @xmath258 the residues come from the zeros of @xmath259 in the exterior of the ellipse , and straight - forward computations show that there are exactly two such zeros , located on the imaginary axis and more precisely given by @xmath260 thus we have a quadrature identity of the form @xmath261 @xmath262 being the residue of @xmath263 at @xmath264 . the exterior of the ellipse is the conformal image of the unit disk under the joukowski map @xmath265 thus @xmath266 , or simply @xmath267 with @xmath39 denoting the identity function , is a ( single - sheeted ) algebraic domain .
the same function @xmath13 maps the exterior of the unit disk onto a multi - sheeted algebraic domain , i.e. , @xmath113 is ( or represents ) such a domain .
it covers @xmath47 twice and @xmath255 once , in other words the counting function is @xmath268 since there are several sheets the associated quadrature identity is best expressed in a form pulled - back to @xmath269 , i.e. , on the form ( [ qi ] ) . a slightly weaker form is @xmath270 where @xmath271 and @xmath272 is the residue of @xmath263 at @xmath273 .
this form is weaker because it only uses test functions @xmath274 that take the same values on the two sheets over @xmath47 . by inversion in the unit circle and a rotation by @xmath275 degrees ( for convenience )
the ellipse transforms into a curve known as neumann s oval @xcite , @xcite , @xcite , @xcite with equation @xmath276 the exterior of the ellipse transforms into a bounded domain @xmath1 , and since inversions and rotations are rigid transformations with respect to the spherical measure , also @xmath1 will be a two point quadrature domain for the spherical measure .
the formula is immediately obtained by inversion and rotation of ( [ qiellipse ] ) .
the domain @xmath1 also satisfies a quadrature identity for the euclidean measure ( indeed , both types of quadrature identities are equivalent to @xmath179 being meromorphic in @xmath1 ) .
the latter quadrature identity is somewhat simpler , namely @xmath277 it turns out to be a quite rewarding task to investigate the algebraic curves corresponding to the level curves of the left member in ( [ neumann ] ) , and in particular to determine which of them correspond to multi - sheeted algebraic domains .
most types of phenomena which could possibly show up really do show up among these curves .
this task is what we are going to undertake for the remainder of this section . in order to simplify a little we first scale so that the quadrature nodes above become @xmath278 .
this means that @xmath279
. then we need only one parameter ( in place of the two , @xmath41 and @xmath42 ) , which we take to be @xmath280 the quadrature identity now becomes @xmath281 holding for all integrable analytic functions @xmath123 in @xmath1 .
set @xmath282 which is the polarized version ( @xmath283 polarizes into @xmath284 ) of the left member in ( [ neumann ] ) .
there are exactly two open sets for which the quadrature identity ( [ twopointqi ] ) holds , namely @xmath285 and @xmath286 = \{z\in \c : q(z,\bar z)<0\}\cup \{0\}.\ ] ] the latter is just the completion of the former with respect to one missing point .
the domain @xmath1 ( or @xmath287 $ ] ) can be viewed as two disks glued together , or `` smashed '' , or `` added '' , and has been studied by many authors , for example @xcite , @xcite , @xcite . to analyse the level curves of @xmath288 we set , for any @xmath289 , @xmath290 let @xmath291 be the corresponding homogenous polynomial .
we shall keep @xmath292 fixed and just vary @xmath293 .
the real locus of @xmath294 in @xmath295 is @xmath296 it represents the intersection with @xmath297 ( `` the real '' ) of the complex locus in @xmath298 , @xmath299 which has a natural completion in the projective space @xmath300 as @xmath301 here @xmath298 is embedded in @xmath300 so that @xmath284 corresponds to @xmath302 . by real points in @xmath298
we mean points @xmath284 satisfying @xmath303 , and these are the fixed points of the involution @xmath304 . in projective coordinates the involution is @xmath305 .
in addition to the anticonformal involution @xmath84 , the curve @xmath306 has the conformal symmetries @xmath307 and @xmath308 .
let @xmath309 be the compact symmetric riemann surface corresponding to @xmath310 .
as point sets they are identical except for a few singular points on @xmath311 , which are resolved on @xmath312 .
the real locus of @xmath294 corresponds to the fixed point set @xmath313 of @xmath314 .
a first observation is that the function @xmath315 has five stationary points in the complex plane .
there are two global minima , on the level @xmath316 , there are two saddle points on the level @xmath317 and there is one local maximum ( at the origin ) , on the level @xmath318 .
these three values of @xmath293 will correspond to changes of regime for the algebraic curve @xmath319 . solving the equation @xmath319 for @xmath40 as a function of @xmath39 gives the schwarz functions for the curves in the real locus .
the result is @xmath320 we see that @xmath321 in general has four branch points .
the levels @xmath293 at which @xmath288 has stationary points are exactly those values of @xmath293 for which some or all of these branch points resolve : for @xmath322 the square root resolve completely into second order polynomials , and for @xmath318 one pair of branch points resolves .
let @xmath323 denote the genus of @xmath312 , or equivalently of @xmath324 .
the degree of @xmath294 is four , hence the genus formula in algebraic geometry @xcite , @xcite tells that @xmath325 where @xmath326 is a certain number related to the singular points .
an analysis , carried out in detail in @xcite , shows that @xmath311 passes through the points @xmath327 and @xmath328 and that it at each of these points has two simple cusps of multiplicity one with distinct tangent directions .
in particular , the points @xmath327 and @xmath328 are singular , and it turns out that each of them gives the contribution @xmath71 to @xmath329 .
except for the above two points of infinity , @xmath311 stays in @xmath298 .
thus , what remains of the genus formula is @xmath330 where @xmath331 denotes the contribution to @xmath329 which comes from finite singular points . by ( [ genusformula ] ) , @xmath332 , so there is at most one finite singular point , and this must be visible in the real because nonreal singular points necessarily come in pairs .
@xmath333 for @xmath317 , @xmath294 factors as @xmath334 each factor defines its own algebraic curve and riemann surface .
these have genus zero and are moreover symmetric : the zero locus of each factor is preserved under the involution @xmath335 .
the real locus @xmath336 is the union of two intersecting circles , those of radius @xmath337 and centers @xmath278 .
the intersection points , @xmath338 are saddle points for the function @xmath315 .
@xmath339 for @xmath316 , @xmath294 factors as @xmath340 where again each factor defines its own algebraic curve and riemann surface of genus zero .
however , in the present case they are not symmetric , instead the involution maps each of these riemann surfaces onto the other . the real locus consists only of the two points @xmath341 .
this can easily be understood by observing that the value @xmath316 is the global infimum of @xmath288 . in both of the reducible cases bezout
s theorem says that there should be four points of intersection between the two curves ( since these have degree two ) .
these intersection points are the two points of infinity @xmath327 and @xmath328 plus , in the first case the intersection points of the two circles ( in the real ) , and in the second case the two points @xmath341 .
besides the above two special values of @xmath293 , also the quadrature value @xmath318 is exceptional .
this is a local maximum value for @xmath288 .
the local maximum is attained at @xmath342 , and the corresponding point @xmath343 on the algebraic curve is a singular point ( since both partial derivatives of @xmath56 vanish there ) .
thus @xmath344 in the genus formula ( [ genusformula ] ) , hence @xmath345 . * for @xmath349
, @xmath313 has one component and there are no singular points visible in the real .
therefore , by ( [ genusformula ] ) and the remark following it @xmath312 has genus one .
this means that the symmetry line @xmath313 is not able to separate @xmath312 into two halves ( see more precisely discussions in section 2.2 in @xcite ) .
thus @xmath350 has only one component , and it will not generate any algebraic domain ( even multi - sheeted ) , despite the nice picture in the real , with a smooth algebraic curve bounding a simply connected region ( figure [ picture1 ] , left ) .
@xmath312 can be viewed as the double of a mbius band .
* at @xmath318 the genus of @xmath312 collapses to zero , @xmath313 has still only one component even though @xmath336 has gotten an additional point , @xmath343 , which is a singular point of @xmath311 ( resolved on @xmath312 ) .
it follows that @xmath350 has two components , one of which , say @xmath99 , is mapped conformally onto the quadrature domain @xmath287 $ ] by the analytic function @xmath13 which corresponds to the projection @xmath351 on @xmath311 . * for @xmath352
the genus of @xmath312 is again one , and the singular point @xmath343 in the previous case has now grown up to a curve , hence @xmath313 has two components .
also @xmath350 has two components , say @xmath353 , and @xmath312 can be viewed as the double of @xmath99 ( or @xmath100 ) , which topologically is an annulus .
+ the meromorphic function @xmath354 which corresponds to @xmath351 on @xmath311 is however no longer univalent ( not even locally univalent ) on what corresponds to @xmath99 in the previous case .
therefore this gives a now only multi - sheeted algebraic domain , with @xmath108 consisting of a main piece which contains the points @xmath278 and the origin ( the dashed area in figure [ picture1 ] , right ) plus a smaller piece around the origin ( the bounded undashed region ) .
the latter piece thus is covered twice , and the two sheets are connected via two branch points of the schwarz function . * for @xmath317 the curve is reducible , hence @xmath312 is the union of two independent riemann surfaces , both of genus zero and symmetric under @xmath314 . in the real locus @xmath355 we simply have two intersecting circles , those centered at @xmath278 and having radius @xmath337 ( figure [ picture2 ] , left ) . explicitly : @xmath356 @xmath312 is the union of two riemann spheres and can be viewed as the double of two disks . *
for @xmath357 we are back to the case of genus one with @xmath313 and @xmath350 both having two components .
however , the situation has changed in the sense that the involution goes the other way ( like @xmath358 in place of @xmath359 in a right - angled period parallelogram ) , it may be more natural in this case to think of @xmath312 as the double of a cylinder than as the double of an annulus ( even though these two types of domains are topologically equivalent ) .
the real locus consists of two closed curves , one enclosing two branch points close to @xmath360 , the other enclosing two branch points close to @xmath361 , and with @xmath354 as before , @xmath13 maps @xmath99 ( say ) onto the dashed region to the right in figure [ picture2 ] ( right ) covered twice and the ( unbounded ) undashed region covered once .
* when @xmath316 the two closed curves in the real locus of the previous case have shrunk to two points , the minimum points @xmath362 of @xmath288 .
the curve is reducible and @xmath312 hence is the disjoint union of two riemann surfaces ( of genus zero ) , but these are not symmetric under @xmath84 . instead the involution @xmath84 maps each of them onto the other .
the algebraic curve @xmath336 consists of two pieces , which meet each other in two points of tangency .
this is what is seen in the real locus .
thus @xmath312 can be viewed as the double of the riemann sphere , and @xmath314 has no fixed points . *
when @xmath363 finally , there is no real locus at all .
the genus is one , but now @xmath84 has no fixed points at all ( @xmath313 is empty ) . therefore no algebraic domain ( even multisheeted )
can be associated to this case .
@xmath312 can be viewed as the double of klein s bottle ( see again @xcite , and also @xcite for doubles of nonorientable surfaces ) .
* for @xmath349 : a mbius band * for @xmath318 : a disk . * for @xmath366 : an annulus . * for @xmath317 : two disjoint disks . * for @xmath357 : a cylinder . * for @xmath316 : a sphere . * for @xmath363 : a klein s bottle .
the pair @xmath367 defines a single - sheeted algebraic domain for @xmath318 and a multi - sheeted algebraic domain for @xmath352 and @xmath357 .
for @xmath317 it defines two single - sheeted algebraic domains ( which intersect in the complex plane ) .
the exponential transform of the above multi - sheeted algebraic domains is given by theorem [ thm : emathcale ] , where the elimination function is @xmath368 with @xmath294 as in ( [ qalpha ] ) and the schwarz function is given by ( [ w(z ) ] ) .
varchenko and p.i . , etingof ,
_ why the boundary of a round drop becomes a curve of order four _ , american mathematical society ams university lecture series , vol .
* 3 * , providence , rhode island 1992 | we introduce multi - sheeted versions of algebraic domains and quadrature domains , allowing them to be branched covering surfaces over the riemann sphere .
the two classes of domains turn out to be the same , and the main result states that the extended exponential transform of such a domain agrees , apart from some simple factors , with the extended elimination function for a generating pair of functions . in an example we discuss the algebraic curves associated to level curves of the neumann oval , and determine which of these give rise to multi - sheeted algebraic domains . |
ngc 6822 is the most nearby dwarf irregular galaxy outside the milky way / lmc / smc system and is therefore a prime target to study in detail the interplay of star formation and the atomic interstellar medium ( ism ) . at a distance of only 490 kpc ,
neutral hydrogen observations reach an unprecedented physical resolution and stellar population studies can be performed even with ground - based telescopes .
extensive studies of the distribution and evolution of the stellar populations in ngc 6822 were done by e.g. , @xcite , @xcite and @xcite .
these studies find a star formation history which is stochastic , low - level , but on average constant , with an increase in star formation rate ( sfr ) of a factor of a few around @xmath4 myr ago , which persists until the present .
various estimates for the recent sfr @xcite all indicate a low value between @xmath5 and @xmath6 , with an average of 0.06 @xmath7@xcite .
ngc 6822 is a metal poor galaxy , with an ism metal abundance of @xmath8 0.2 - 0.3 @xmath9 @xcite .
similar values are found for the stellar abundances @xcite .
the intermediate - age stellar population , as traced by c and m stars , was mapped by @xcite , who found that these are distributed in an elliptical halo that does neither trace the distribution of the hi , nor that of the main optical component .
a number of c stars are found outside the hi disc , though this may be a common phenomenon for dwarf galaxies ( e.g. @xcite ) .
the most recent observations in the 21-cm line of neutral hydrogen are presented in @xcite ( hereafter dbw00 ) and @xcite .
these show that the galaxy is dark matter - dominatated , gas - rich , and reveal many interesting features , such as a highly structured hi disc , a super - giant hi shell , and the presence of a companion cloud . in this letter , we discuss the properties and distribution of the stellar population , with emphasis on the young blue stars .
ngc 6822 was observed with the 40@xmath10 telescope at siding spring observatory , australia from 5 - 12 july 1999 .
we used the imager with a 2048@xmath112048 tek ccd with a pixel scale of 0.6@xmath10 pixel@xmath12 , resulting in an effective field of view of @xmath13 .
the entire extent of the hi disc of ngc 6822 was covered in 8 pointings in @xmath14 and @xmath15 . each field
was observed for 2500s in @xmath15 and 3000s in @xmath14 . the total area covered was @xmath16 square arcmin .
standard stars were observed throughout the run .
the seeing was fairly constant at 1.5@xmath102@xmath10 .
conditions were photometric for 3 nights and variable during the rest of the run .
we used the overlap regions between fields to put non - photometric images on the same flux scale .
we found that due to this procedure absolute photometry is accurate to @xmath17 mag .
we used the dophot @xcite crowded field photometry package to create a catalog of stellar objects in @xmath14 and @xmath15 in the observed area .
best results were obtained by using an analytical point spread function , adjusted to the typical seeing size for each field , and using a @xmath18 pixels box - averaged median background .
stellar detections in both bands were matched by cross - correlating the two lists , assuming a @xmath19 uncertainty in the positions .
this procedure yielded a catalogue of @xmath20 stellar objects , the great majority of them being galactic foreground stars .
in fig . [ colmag ] the colour - magnitude diagram ( cmd ) of the entire observed field in @xmath14 and @xmath15 is shown .
two features are obvious : the large concentration of stars at @xmath21 and redder , and the plume with bluer colors at @xmath22 .
stars with @xmath21 are mainly galactic foreground stars , though it is expected that some bright red stars in ngc 6822 also contribute .
stars with @xmath23 are the blue stars which ( mostly ) belong to ngc 6822 .
completeness limits at the faint magnitude end were derived by plotting the histograms of the apparent @xmath14 and @xmath15 magnitudes .
these showed the expected steep increase of number of stars toward fainter apparent magnitudes , followed by a sharp drop - off due to decreasing completeness .
the peaks in the histograms occurred at @xmath24 and @xmath25 , and we take these as our fiducial completeness limits .
we select the blue stars in the cmd by choosing only those detections with @xmath26 .
this blue stellar population will be discussed in detail in this letter . using a distance modulus of @xmath27 @xcite , we find that the colours and luminosities of the blue stars associated with ngc 6822 are consistent with main sequence stars with spectral types between b5 @xmath28 and o5 @xmath29 .
there are possibly a few stars of even earlier type present .
the main sequence life - times of these stars vary from @xmath30 myr ( b5 ) to @xmath31 myr or less ( o5 ) .
stars of later spectral types must obviously be present in ngc 6822 as well , but can not be detected using the current data set .
the total observed magnitude of the 1549 detected blue stars is @xmath32 or @xmath33 after correction for extinction and distance ( the total blue luminosity of ngc 6822 is @xmath34 ) .
figure [ mommap]a shows the position of the blue stars compared to the approximate extent of the optical galaxy , where we distinguish between stars with @xmath35 ( `` faint '' ) ( grey dots ) and stars with @xmath36 ( `` bright '' ) ( black dots ) . also shown ( fig .
[ mommap]b ) are the positions of the blue stars in relation to the distribution of the atomic interstellar medium ( ism ) in ngc 6822 ( see dbw00 and @xcite ) .
it is immediately clear that _ the blue stars cover a much larger area than what is usually regarded as the optical component of ngc 6822 _ ; _ the distribution of the blue stars is as extended as that of the inner main hi disk_. additionally stars seem present in the nw hi cloud .
the only areas seemingly devoid of blue stars are the centre of the large hole and the tidal arm in the se .
note that the latter does not imply there are no stars present there .
it simply means that stars in this area must be fainter than @xmath37 in order not to be detected in the current data set .
deeper surveys may perhaps show a faint population there . in fig .
[ mommap]c we show the distribution of c stars , which trace the intermediate age ( @xmath8 35 gyr ) population ( from @xcite ) . _
the distribution of the young stars is in stark contrast with that of the older stellar population . _
the latter is distributed in an elliptical halo that does not trace the distribution of the hi , and shows no internal structure .
the young stars , on the other hand , have a clumpy , almost filamentary distribution .
two of the large concentrations coincide with the opposite ends of the bar - like optical structure .
there is a larger ratio of bright to faint blue stars inside the optical bar than outside , presumably indicating more recent star formation there .
blue star candidates are scattered throughout the field ; however , most of these are likely to be false candidates ( faint galactic blue stars , or , more probably , stars that are scattered into the blue bin due to magnitude uncertainties ) . nevertheless , as presumably the contribution of false candidates within the disk of ngc 6822 is as important as outside the disk , it is illustrative to quantify the significance of the detected overdensities in the presence of this background of false detections . to this end
we have used the observed distribution of detected objects to calculate the number density distribution by counting the number of detections in 30@xmath10 by 30@xmath10 boxes .
this size was chosen as a compromise between preserving spatial resolution and getting sufficient numbers of detections in each box .
the size of the box does not affect the conclusions .
the creation and detection of false candidates in the field is a random proces , so the presence or absence of candidates in each box can be regarded as a shot - noise process , and the surface density of the false candidates will be approximately constant across the field .
we have investigated the number of false candidates in each box outside the @xmath38 @xmath3 hi column density contour ( i.e. , outside the entire hi disk as shown in fig .
[ mommap ] )
. we will refer to this area as `` the field '' .
there are 4955 30@xmath39 boxes in the field , of which 4868 have no false candidates , 84 have 1 false candidate and 3 have 2 false candidates .
this distribution can be well described by a poisson distribution @xmath40 , with @xmath41 .
the cumulative poisson distribution then yields that the chance of getting 2 or more false candidates in a @xmath42 box due to shot - noise effects is less than 0.1 percent .
figure [ densmap ] shows the number density image , where we only show boxes with 2 or more candidates , which are thus statistically very unlikely to have resulted from random effects .
note that the presence of the few significant boxes outside the hi disk is entirely consistent with the derived poisson distribution . using this distribution
we can also compute that we expect @xmath43 false detections within the entire hi disk .
this is in sharp contrast with the @xmath44 stars actually found .
a repeat of this analysis with different sized boxes yields identical conclusions : the number density of false detections can be described by a poisson distribution and the features in the stellar distribution as described above are statistically significant .
figure [ coldensx ] shows a histogram of the surface densities @xmath45 at the position of the blue stars ( at @xmath46 [ @xmath47 pc ] resolution ) . of the @xmath44 stars detected in the hi disc , @xmath48
are found at surface densities of @xmath49 @xmath3 , close to or above the empirical star - formation threshold ( e.g. @xcite ) .
it is however worth noting that @xmath50 stars are found at surface densities significantly lower than this star formation threshold value .
( these numbers do not include the presumed false candidates outside the hi disc ) . from the poisson distributions derived above
we only expect @xmath43 false candidates in the hi disk , clearly not enough to explain the large number of stars at low surface densities . for comparison ,
a normalized histogram showing the relative distribution of surface densities of the entire hi disc is over - plotted in fig .
[ coldensx ] .
figure [ mommap]d shows the positions of the blue stars superimposed on a map of the velocity dispersion of the hi gas .
a histogram of the velocity dispersions at the positions of the blue stars is shown in fig .
[ coldensx ] .
the average velocity dispersion of the hi gas at the positions of the blue stars is @xmath51 @xmath52 , about 1 @xmath52 higher than the average velocity dispersion of the hi gas in the entire disc .
assuming an internal velocity dispersion of a few km s@xmath12 for a single star cluster ( or ob association ) after its formation out of a molecular cloud , we estimate that the observed young stars can travel @xmath8 100 pc away from their parent molecular cloud over their lifetimes .
it is thus unlikely that the young stars found at low surface densities have traveled there from high - density regions . *
the companion hi cloud . * a small concentration of ob stars at a galactocentric distance of @xmath53 kpc is detected to the nw of the galaxy .
it coincides with the position of the companion hi cloud described in dbw00 .
this companion is presumably interacting with ngc 6822 on timescales of 100 - 200 myr ( see dbw00 for details ) .
the presence of blue stars with ages less than @xmath8100 myr gives additional support to this scenario , the idea being that star formation was triggered in the smaller companion as a result of the interaction .
moreover , no tracers for an intermediate age population of stars are found in the hi cloud , implying that we may indeed be seeing the first major episode of star formation here . * the supergiant hi shell . * it is striking that no large population of bright blue stars is found inside the supergiant hi shell ( sgs ) . in dbw00
we derived an age of @xmath54 myr for the sgs .
the blue stars trace ages up to @xmath55 myr so we can place a lower limit on the age of the sgs of order @xmath56 myr . if the sgs was ` only ' 130 myr old , and if stars created the hole , we would expect to detect at least some of the remnant stars .
the non - detection of this population of stars implies that the sgs is probably older than previously derived .
as was noted by dbw00 , it is difficult to derive an accurate dynamical age for the sgs due to the lack of clear signs of expansion .
as ngc 6822 rotates almost like a solid body throughout the region of interest @xcite , large coherent structures such as the sgs can remain intact for many @xmath1 years , long after the remnant stellar population has faded ( see also @xcite ) . * the prominent concentration of young stars toward the south - east . * a prominent large concentration of blue stars outside @xmath57 is readily seen to the south - east of the bar .
it causes the asymmetric appearance of ngc 6822 in the @xmath14-band ( e.g.@xcite ) .
it is in fact the most prominent association of blue stars in the galaxy . while the velocity dispersion of the hi at the positions of the other large blue star concentrations hovers between @xmath58 and @xmath59 @xmath52 , at this position
the average dispersion is only @xmath60 to 7 @xmath52(cf
. figs .
[ mommap]d and [ coldensx ] ) .
there is a lack of prominent h@xmath61 emission in this region compared to the other concentrations ( see @xcite ) , suggesting that the sfr has declined in the past 10 myr .
the conspicuous lack of c stars indeed suggests that this region has undergone its first major episode of star formation some 10 - 100 myr ago .
we have presented a comparison of the distribution of blue stars and the atomic ism in ngc 6822 .
we find that the distribution of the blue stars is almost as extended as that of the main hi disk of ngc 6822 . only the large hole and the tidal arms are devoid of blue stars .
blue stars are typically found at hi surface densities @xmath62 @xmath3 , however over one - third of the blue stars is found at lower surface densities .
this could imply that _ a ) _ stars can form at surface densities below the canonical star formation threshold ; _ b ) _ the ism is very clumped at scales smaller than the current resolution of @xmath63 pc ; _ c ) _ star formation has destroyed or ionized the regions of higher surface densities in which these stars originally formed or _
d ) _ the stars traveled from higher density regions within their lifetimes .
however , due to their short lifespan , blue stars are only expected to travel some 100 pc from their places of birth and stellar cluster dispersion can not explain the similarity of the distributions of the blue stars and the atomic ism .
given the rather symmetrical distribution of the c - stars , it is clear that neither the sgs , nor the interaction with the nw cloud has affected the distribution of the old population ( and therefore the bulk of the mass in stars and stellar remnants ) .
the processes that are so dramatically visible in the hi , must have had hardly any impact on the overall mass distribution in ngc 6822 .
the presence of blue stars over the main body of ngc 6822 , as well as in the nw cloud implies that star formation is occurring not just in the optical disk , but that pollution of the ambient ism is taking place on a larger scale ( similar conclusions can be drawn for massive galaxies as well , see e.g. , @xcite ) .
this may partly explain the flat abundance patterns found in chemical abundances studies of low - mass galaxies ( e.g. , @xcite ) .
no large population of young blue stars has been detected in the supergiant hi shell , setting a lower limit on its age of @xmath64 myr ( again , this does not rule out the presence of a fainter population ) .
the fact that a population of blue stars with ages @xmath65 myr is found in the companion hi cloud , but no tracers of an older population are found , provides evidence that star formation has been triggered here only recently , very possibly for the first time . in summary , our results show that the distribution of the atomic ism in dwarf irregular galaxies is not necessarily much more extended than that of its stellar population .
edb acknowledges support from a pparc advanced fellowship .
battinelli , p. , demers , s. , 2000 , aj , 120 , 1801 cuillandre , j .- c . , lequeux , j. , allen , r.j . ,
mellier , y. , bertin , e. , 2001 , apj , 554 , 190 de blok , w.j.g . , walter , f. , 2000 , apj , 537 , l95 gallart , c. , aparicio , a. , vlchez , j.m .
1996a , aj , 112 , 1928 gallart , c. , aparicio , a. , bertelli , g. , chiosi , c. 1996b , aj , 112 , 1950 gallart , c. , aparicio , a. , bertelli , g. , chiosi , c. 1996c , aj , 112 , 2596 hodge , p.w . , 1980 , apj , 241 , 125 hodge , p.w .
, 1977 , apjs , 33 , 69 hodge , p. , smith , t. , eskridge , p. , macgillivray , h. , beard , s. , 1991 , apj , 379 , 621 hodge , p. , kennicutt , jr .
lee , m.g . , 1988 ,
pasp , 100 , 917 hodge , p. , lee , m.g . ,
kennicutt , jr .
, r.c . , 1989 , pasp , 101 , 32 israel , f.p . ,
bontekoe , tj.r . ,
kester , d.j.m .
, 1996 , a&a , 308 , 723 kennicutt , jr . , r.c . , 1989 , apj , 344 , 685 kobulnicky , h.a . , skillman , e.d . , 1997 , apj , 489 , 636 letarte , b. , demers , s. , battinelli , p. , kunkel , w.e . , 2002 , apj , 123 , 832 mateo , m.l . , 1998 , ara&a , 36 , 435 pagel , b.e.j , edmunds , m.g . , smith , g. , 1980 , mnras , 193 , 219 schechter , p.l . ,
mateo , m. , saha , a. , 1993 , pasp , 105 , 1342 skillman , e.d .
, 1987 , in star formation in galaxies , ed .
lonsdale persson , nasa , cp-2466 , p. 263 skillman , e.d .
, terlevich , r. , melnick , j. , 1989 , mnras , 240 , 563 venn , k.a . , lennon , d.j . ,
kaufer , a. , mccarthy , j.k . ,
pryzbilla , n. , kudritzki , r.p . , lemke , m. , skillman , e.d .
, smartt , s.j . , 2001 ,
apj , 547 , 765 walter f. , brinks e. , 1999 , aj , 118 , 273 weldrake , d.t.f . , de blok , w.j.g . , walter , f. , 2003 , mnras , 340 , 12 | we present wide - field optical imaging covering the entire neutral hydrogen disc of the local group dwarf galaxy ngc 6822 .
these observations reveal the presence of numerous blue , young stars at large galactocentric radii well beyond @xmath0 .
blue stars are also found that are associated with ngc 6822 s companion hi cloud , indicating that star formation was triggered in the companion in the last @xmath1 yr . in general ,
blue stars are present where the hi surface densities reach values @xmath2 @xmath3 . however , over one - third of the blue stars detected within the hi disk are found at lower surface densities .
the young stars trace the distribution of the neutral hydrogen in the inner disk , but seem to be avoiding the supergiant hi shell in ngc 6822 , setting a lower limit for its age of @xmath1 yr .
the extended distribution of young stars implies that stars can form at large galactocentric radii in dwarf galaxies ; the hi is therefore not necessarily much more extended than the stellar population .
this finding has important consequences for the chemical enrichment of the interstellar medium throughout ( dwarf ) galaxies .
galaxies : individual ( ngc 6822 ) - galaxies : dwarf - galaxies : evolution - galaxies : stellar content |
methanol is a complex molecule it consists of six atoms , and two of them are heavy .
it is surely the 1st prize winner by complexity among the molecules producing cosmic masers .
it is really surprising that such a molecule gives birth to numerous and extremely bright masers that are only matched by the 1.7 ghz hydroxyl ( oh ) and 22 ghz water vapor ( h@xmath1o ) lines .
this paper describes the methanol maser phenomenon on different spatial scales emphasizing the physical conditions that characterize methanol maser sites .
several aspects of this topic are covered in papers of other participants of this conference .
we concentrate on results obtained by researchers from the ural state university mostly in cooperation with scientists from other institutes .
methanol masers are traditionally divided into two classes according to the sets of transitions in which they emit .
historically , the first methanol masers ( later attributed to class i ) were discovered in orion in the @xmath2 e series of lines at about 25 ghz ( ( * ? ? ?
* barrett 1971 ) ) .
the existence of the second class of methanol masers ( eventually named class ii ) was justified in the paper by @xcite after the discovery of strong masers in the @xmath3 e line at 12.1 ghz .
the existence of an intermediate class of methanol masers is predicted but these masers are not likely to attain fluxes comparable to those of the strongest lines of the two major classes ( ( * ? ? ?
* voronkov 2005 ) ) .
methanol masers of both classes often co - exist in the same star - forming region , though there are bright and notorious examples of class - i - only ( omc-1 ) and class - ii - only ( w3(oh ) ) sources .
methanol maser lines can achieve extremely high brightness : class i maser lines at 44 ghz are reported to be brighter than @xmath4 k ( ( * ? ? ?
* kogan & slysh 1998 ) ) and class ii lines at 6.7 ghz achieve even higher brightness exceeding @xmath5 k ( ( * ? ? ?
* menten 1992 ) ) .
spectra of the bright methanol masers usually contain numerous components ( see , e.g. , ( * ? ? ?
* caswell 1995 ) ) spread in a range smaller than 20 km / s around the systemic velocity of the respective molecular core ( ( * ? ? ?
* malyshev & sobolev 2003 ) ) .
maser components are usually narrow ( about 0.5 km / s ) and the narrowest ever reported line ( @xmath60.1 km / s ) belongs to a class i methanol maser ( ( * ? ? ?
* voronkov 2006 ) ) .
methanol masers display time variability of various kinds including periodic , chaotic and episodic flaring ( ( * ? ? ?
* goedhart 2004 ) ) .
studies of the distribution of methanol masers in the galaxy can be reduced to class ii maser sources because most of the class i maser sites are found close to class ii ones ( see , e.g. , ( * ? ? ?
* slysh . 1994 ) and ( * ? ? ?
* ellingsen 2005 ) ) . at present there exist three unified catalogues of class ii methanol masers each containing information on more than 500 objects ( ( * ? ? ?
* malyshev & sobolev 2003 ) ; ( * ? ? ?
* xu , zheng & jiang 2003 ) and ( * ? ? ?
* pestalozzi 2005 ) ) .
the distinguishing feature of the @xcite catalogue is that it contains data on several maser transitions and cross - references to the data on molecular shock tracers .
this catalogue is updated on a monthly basis and is available at http://astroserver.astro.usu.ru .
most of the class ii maser sources were found toward positions of oh masers and iras sources with specific spectral characteristics .
a new way of targeting based on glimpse data is suggested in @xcite .
blind surveys , including a recent arecibo survey by @xcite led to the detection of many new maser sources in selected areas .
a real breakthrough is expected after the completion of the blind survey within the mmb ( methanol multibeam ) project ( see the paper by j. green in this volume ) .
information on the galactic distribution of class ii methanol masers was presented in several papers which established that these masers mostly reside in the molecular ring of our galaxy ( for one of the most recent articles , see ( * ? ? ?
* pestalozzi 2005 ) ) . at present , class
ii masers were found only in high mass star - forming regions ( ( * ? ? ?
* minier 2003 ) ) and it is natural to expect that these sources should be concentrated in the spiral arms . for the outer galaxy ,
the concentration of methanol maser sources to the perseus spiral arm is quite certain : the velocity - galactic latitude ( @xmath7 ) diagram shows that the masers mostly reside in spiral arms which are well outlined by co emission ( see figure [ fig : fg1 ] ) .
evidence for a concentration of the methanol masers in the spiral arms is more difficult to prove in the tangled regions of the inner galaxy and was first provided in @xcite .
new maser sources were reported since then , mainly in @xcite and @xcite .
figure [ fig : fg1 ] shows a @xmath8 diagram for the masers from the inner part of the galaxy along with lines of equal galactocentric distance , @xmath9 , and the distribution of the maser sources as a function of @xmath9 .
the updated @xcite catalogue and the @xcite rotation curve were used .
the concentration of masers to certain ranges of galactocentric distance is apparent .
the values of @xmath9 correlate with the nearest spiral arm sections of the inner galaxy .
it is , however , not possible to conclude that all class ii masers reside in the spiral arms .
for the bulk of the sources there are only kinematic distances which are not accurate enough .
moreover , most of the distance estimates are based on velocities of 6.7 ghz maser peaks which are significantly scattered with respect to the systemic velocities of the clouds ( ( * ? ? ?
* malyshev & sobolev 2003 ) and ( * ? ? ?
* van der walt , sobolev & butner 2007 ) ) .
so a sophisticated analysis of the galactic methanol maser distribution requires a drastic improvement in the precision of distance estimates .
diagram ( lines of equal galactocentric distances are shown , background : co data from ( * ? ? ?
* dame 2001 ) ) and distribution of the maser sources as a function of galactocentric distance.,width=480 ]
in this section we will discuss the association of methanol masers with the large scale constituents of star forming complexes in the perseus arm .
we confine ourselves to the perseus arm because these objects have relatively simple structure , most of them have no confusing and strong back- or foreground and the objects are relatively near .
we realize that the perseus arm complexes are somewhat different from those residing in the other parts of the galaxy but it is easier to start with the least complicated case .
28 sources were selected from the @xcite catalogue on the basis of their position in the @xmath8 diagram . using the skyview virtual observatory ( http://skyview.gsfc.nasa.gov/ ) we compiled an atlas of images of these sources displaying their association with developed hii regions seen at optical wavelengths ( r colour images ) , embedded star clusters revealed in near infrared ( k@xmath10 colour images ) and compact hii regions detected in the nrao vla sky survey ( nvss ) .
the numbers of maser sources associated with these types of objects are shown in figure [ fig : fg2 ] .
it is found that 57 % of the sources are associated with embedded clusters which means that these maser sites can be affected by several young stellar objects ( ysos ) .
this is especially true for class i masers which are found apart from the ysos .
the fraction of cluster associations can be substantially higher since k@xmath10 colour images do not show very deeply embedded clusters revealed at submm and mm wavelengths . for 43 % of the targets we find an association with vla sources which means that a considerable fraction of maser sources is associated with rather developed compact hii regions around luminous ysos .
29 % of the sources appear to be associated with optical hii regions .
we conclude that a considerable fraction of the methanol masers is triggered by the expansion of evolved hii regions . since the time duration of the methanol masing phase is quite short ( see , e.g. , ( * ? ? ?
* van der walt 2005 ) ) this means that contemporary massive star formation in the perseus arm is to a great extent controlled by local phenomena .
a significant fraction of the sources ( @xmath1140 % , mostly the distant ones ) did not show a clear association with the objects of our atlas .
the environment of class ii methanol masers was discussed in many papers ( see , e.g. , ( * ? ? ?
* ellingsen 2006 ) , ( * ? ? ?
* minier 2005 ) and ( * ? ? ?
* walsh 2001 ) ) and talks at this symposium ( review by v. fish , talks by j. de buizer , s. ellingsen , t. hill , s. longmore , l. moscadelli and others ) . here
we will instead emphasize two facts which are decisive for our understanding of the maser pumping : it is established that almost all of the maser sources are associated with mid - infrared ( mir ) emission ( ( * ? ? ?
* walsh 2001 ) ) and some of them are associated with submm sources without cm - ir counterparts ( ( * ? ? ? * hunter 2006 ) ) .
both observations and modelling of class ii masers show that the 6.7 ghz transition always manifests the highest brightness temperature ( ( * ? ? ?
* malyshev & sobolev 2003 ) ; ( * ? ? ?
* sobolev 1997 ) ) .
observations have not shown significant differences in the positions of the maser spots seen in different transitions ( see , e.g. , ( * ? ? ?
* menten 1992 ) ; ( * ? ? ?
* sutton 2001 ) ) .
so , there is basically only one known pumping regime of class ii masers .
however , ratios of the brightnesses of different maser lines show considerable dependence on the values of the physical parameters .
the hypothesis that the masers are pumped by mir emission of the warm dust allows us to explain observed brightness temperatures and to determine physical parameters of several sources on the basis of multi - transitional observations ( e.g. , ( * ? ? ?
* sutton 2001 ) , ( * ? ? ?
* cragg , sobolev & godfrey 2005 ) ) .
a model of the `` common '' class ii methanol maser source was constructed on the basis of extensive surveys ( e.g. , ( * ? ? ?
* cragg 2004 ) and ( * ? ? ?
* ellingsen 2004 ) ) .
the results of such studies are summarized in @xcite and @xcite and will not be presented here .
the response of the maser fluxes to changes in the physical parameters allows us to determine these parameters and their variation in time via detailed modelling .
for example , interferometric studies of variability in the source g9.62 + 0.20 shows that different parts of the source are synchronized , suggesting a radiative origin for the variability ( ( * ? ? ? * goedhart 2005 ) ) .
monitoring observations have shown that this source is almost strictly periodical ( see the papers by goedhart & gaylard and gaylard & goedhart in this volume ) .
periodicity is revealed in the variations of g9.62 + 0.20 in three maser transitions at 6.7 , 12.2 and 107 ghz ( van der walt in preparation ) . a possible cause for the synchronized variability might be a variation of the dust temperature in the region where maser is formed .
model calculations show that corresponding changes of fluxes can be reproduced by variations of the dust temperature by a few per cent which may follow variations in the brightness of the central yso reflecting the character of the accretion process .
recent observations have shown that the masers which were previously `` hanging in nowhere '' ( e.g. , the brightest 6.7 ghz masers in ngc6334f , ( * ? ? ?
* ellingsen 1996 ) ) are associated with submm emission ( ( * ? ? ?
* hunter 2006 ) ) and have high chances to be formed in the vicinity of hypercompact hii regions ( hchiis ) .
these objects manifest strong free - free emission which in some cases dominates the dust emission up to frequencies as high as 300 ghz ( ( * ? ? ?
* kurtz 2005 ) ) .
we have undertaken model calculations in order to explore whether methanol masers can be pumped by free - free emission with high turnover frequency ( i.e. , high emission measure ) without considerable dust content .
historically , first calculations of this kind were presented in the paper by ( ( * ? ? ? * slysh 2002 ) ) but these authors did not account for very powerful torsional transitions which are crucial for pumping , at least in terms of intensities .
our calculations are based on the most recent transition rates that were used in the model calculations presented in @xcite .
the model parameters are close to those presented by @xcite with the major difference that the warm dust emission is substituted by the free - free emission from the hchii .
we have chosen the following set of parameters : gas temperature @xmath12 k , hydrogen number density @xmath13@xmath14 , methanol specific column density @xmath15@xmath14s , beaming factor @xmath16 , hchii electron temperature @xmath17 k , turnover frequency @xmath18 ghz and hchii dilution factor @xmath19 .
the calculations show that bright class ii masers are indeed produced .
figure [ fig : fg3 ] presents results from two series of models .
the left panel reveals variations of brightness temperature when modifying the turnover frequency .
it is clearly seen that to pump bright class ii masers free - free emission of hchii should have @xmath20 ghz , i.e. it should posess a very high emission measure .
the right panel demonstrates that the pumping of class ii methanol masers takes place quite close to the hchii where the free - free emission is not greatly diluted ( @xmath21 ) .
it is noteworthy that , when the distance from the hchii increases ( @xmath22 ) , some dense , warm and elongated clumps created by interaction of the hchii with the ambient medium can give birth to rather strong class i methanol masers .
this is one of the possible explanations of the fact that masers of different classes can reside close to each other ( see , e.g. , ( * ? ? ?
* ellingsen 2005 ) ) .
it was shown that important constraints on the models can be obtained through an evaluation of upper limits for the fluxes in particular maser lines . within this context
it is very important to know whether the weak lines are masers or whether a given transition is thermalized . in order to find a solution to this problem we conducted a very sensitive search for emission in the @xmath23 and @xmath24 lines at 23.1 and 23.4 ghz using the 100-m telescope at effelsberg .
the transitions at 23.1 and 23.4 ghz highly differ in their response to factors producing departures from lte : e.g. , w3(oh ) is strongly masing at 23.1 ghz and shows absorption at 23.4 ghz ( ( * ? ? ? * menten 1986 ) ) .
in contrast to previous surveys at parkes and onsala the detection rate in the maser line at 23.1 ghz was rather high and we detected emission in 10 sources out of 20 .
the 23.1 ghz line in the sources w51 , w75n , dr21(oh ) , cep a and g12.910.26 showed linewidths comparable to those of corresponding `` quasi - thermal '' lines from these objects . in these sources , the 23.4 ghz line shows very similar intensities and line shapes ( see the w51 spectra in figure [ fig : fg4 ] ) which means that the transitions are thermalized . toward the sources g9.62 + 0.19 , g23.01 + 0.41 and g188.95 + 0.89 a clear detection of the 23.1 ghz line
was not accompanied by 23.4 ghz emission .
the detected lines are considerably narrower than `` quasi - thermal '' lines from these objects ( see figure [ fig : fg4 ] ) .
line narrowing is a property of unsaturated masers with considerable optical depth .
so , we can state that we detected three new weak but real masers and that the weakness of masers does not necessarily mean that their amplification factors are low .
the environment of class i masers is highly turbulent .
lines of the strong masers usually contain several spectral components and their spots form clusters .
general characteristics of the maser images and spectra can be reproduced by modelling the turbulent velocity field ( ( * ? ? ?
* sobolev 1998 ) ) .
the existence of the interferometric data on the images and time variability of the maser line series at about 25 ghz in omc-1 ( ( * ? ? ?
* johnston 1992 ) ) allow us to construct models constraining the characteristics of the turbulence in this source ( ( * ? ? ?
* sobolev 1998 ) ; ( * ? ? ?
* sobolev 2003 ) ) .
class i methanol masers are less strong and widespread than class ii ones .
most of them were found in the vicinity of class ii maser sources ( see , e.g. , ( * ? ? ?
* slysh 1994 ) and ( * ? ? ?
* ellingsen 2005 ) ) .
there were extensive searches ( e.g. , ( * ? ? ?
* slysh 1994 ) and ( * ? ? ?
* valtts 2000 ) ) for class i sources but no unified catalogues of these objects are published in the journals and no successful blind surveys were performed yet .
there are several lines of evidence that class i masers trace dense parts of molecular outflows , probably locations where new stars might form ( e.g. , ( * ? ? ?
* voronkov 2006 ) and ( * ? ? ?
* sutton 2004 ) ) but generally the environment of class i sources is not well established yet .
recently , sensitive and extensive surveys of the southern class i masers were conducted utilizing the unique capabilities of the australia telescope compact array ( see the paper by voronkov in these proceedings ) .
they allow us to draw some general conclusions on the class i maser pumping and identify four distinct regimes differing by the series of the brightest ( in terms of brightness temperature ) line .
the most frequently encountered maser regime shows the brightest lines which belong to the series .
the 4@xmath25e and 5@xmath26e transitions at 36.1 and 84.5 ghz are weak masers under normal conditions of massive star forming regions . usually , the maser nature of the lines in this regime is difficult to prove observationally . nevertheless , there are cases where the line profiles contain narrow spikes and the maser nature is proved interferometrically .
the sources sgr b2 , g30.8 - 0.1 , and g1.6 - 0.025 can be considered as representatives of this maser regime ( ( * ? ? ?
* sobolev 1996 ) ) . in the second class
i maser regime the lines of the series become prevalent .
numerous sources show definitely maser lines arising in the and transitions at 44.1 and 95.2 ghz , respectively .
masers in the sources dr21w , ngc2264 and omc-2 represent this regime ( @xcite ) .
a preliminary theoretical analysis of the pumping shows that the lines of the series become brightest in the models with rather high beaming ( @xmath27 ) and moderate column densities ( ( * ? ? ?
* sobolev 2005 ) ) .
the third maser regime is less widespread and is represented by the sources where the lines of the series at about 25 ghz are the brightest .
omc-1 is the prototypical source ( ( * ? ? ?
* johnston 1992 ) ) and some additional bright sources were found recently ( voronkov these proceedings ) .
lines of this series become brightest in the models with high specific column densities ( @xmath28 [ @xmath29s])@xmath30 ) and require relatively high temperatures ( @xmath31@xmath32 k ) and densities ( @xmath33 [ @xmath34)=5$]@xmath35 ) .
the existence of the forth maser regime with the brightest lines from the series was previously uncertain .
recent atca observations confirmed however , the existence of such a regime in the sources w33-met and g343.12 - 0.06 .
preliminary modelling shows that the line at 9.9 ghz becomes brightest in models with specific column densities which are as high as required for the previously discussed regime .
however , a smaller beaming angle and either lower densities or higher temperatures ( @xmath36 k ) are also required . | general characteristics of methanol ( ch@xmath0oh ) maser emission are summarized .
it is shown that methanol maser sources are concentrated in the spiral arms .
most of the methanol maser sources from the perseus arm are associated with embedded stellar clusters and a considerable portion is situated close to compact hii regions .
almost 1/3 of the perseus arm sources lie at the edges of optically identified hii regions which means that massive star formation in the perseus arm is to a great extent triggered by local phenomena .
a multiline analysis of the methanol masers allows us to determine the physical parameters in the regions of maser formation .
maser modelling shows that class ii methanol masers can be pumped by the radiation of the warm dust as well as by free - free emission of a hypercompact region ( hchii ) with a turnover frequency exceeding 100 ghz .
methanol masers of both classes can reside in the vicinity of hchiis .
modelling shows that periodic changes of maser fluxes can be reproduced by variations of the dust temperature by a few percent which may be caused by variations in the brightness of the central young stellar object reflecting the character of the accretion process .
sensitive observations have shown that the masers with low flux densities can still have considerable amplification factors .
the analysis of class i maser surveys allows us to identify four distinct regimes that differ by the series of their brightest lines . |
with a few free parameters i can fit an elephant . with a few more
i can make him wiggle his trunk wigner s response to questions about a particular theory he did not like was : i think this theory is wrong .
but the old bohr - sommerfeld quantum theory was also wrong .. could we have reached the right theory without it ? in @xmath7 annihilation
a created @xmath8 fragments into hadrons . @xmath9
meson ; @xmath1 baryon . but
@xmath10 triquark and @xmath11triquark @xmath12 pentaquark ..
bj asked whether quark model says such state.is bound or lives long enough to be observable as hadron resonance . listening to bj usually pays off . 1
. qcd motivated models@xcite show same color - electric interaction for large multiquark states and separated hadrons and no binding .
only short - range color - magnetic interaction produces binding .
jaffe@xcite extended dgg model@xcite with one - gluon - exchange color factor to multiquark sector in a single cluster or bag model , defined @xmath13 and @xmath14 interactions and explained why lowlying exotics not observed 3 .
hyperfine ineraction suggested search for @xmath15 dibaryon@xcite @xmath16 and anticharmed strange pentaquark@xcite @xmath17 ( 1987 )
the pauli principle requires flavor - symmetric quark pairs to be antisymmetric in color and spin at short distances .
thus the short - range color - magnetic interaction is always repulsive between flavor - symmetric pairs .
best candidates for multiquark binding have minimum number of same - flavor pairs 1 .
nucleon has only one same - flavor pair ; @xmath18 has three . 2 .
extra two same - flavor pairs costs 300 mev .
deuteron separates six same - flavor pairs into two nucleons + only two same - flavor pairs feel short range repulsion .
@xmath19 has three same - flavor pairs .
optimum for light quark dibaryon 5 .
the @xmath20 pentaquark has only one same - flavor pair 6 .
@xmath5 @xmath21 has two same - flavor pairs , more than @xmath20 .
quark model calculations told experimenters " look for @xmath22 not @xmath5 .
ashery s e791 search for @xmath23 found events@xcite ; not convincing enough . better searches for this pentaquark are needed ; e.g. searches with good vertex detectors and good particle id@xcite ...
any proton emitted from secondary vertex is interesting .
one gold - plated event not a known baryon is enough ; no statistical analysis needed .
sakharov and zeldovich noted that the @xmath24 and @xmath25 are made of same quarks and asked why their masses are different .
their answer was that a unified two - body hyperfine interaction not only answers this question but led to a unified mass formula for both meson and baryon ground states mesons and baryon masses and showed that all are made of same quarks@xcite @xmath26 using ( [ sakhzel ] ) sakharov and zeldovich noted that both the mass difference @xmath27 between strange and nonstrange quarks and the flavor dependence of their hyperfine splittings ( later related@xcite to the mass ratio @xmath28 ) have the same values when calculated from baryon masses and meson masses@xcite , along with the comment that the masses are of course effective masses@xcite : @xmath29 & { } \langle m_s - m_u \rangle_{mes } & = { { 3(m_{k^{\scriptstyle * } } -m_\rho ) + m_k - m_\pi}\over 4 } = 180\,{\rm mev } \\[4pt ] & \langle m_s - m_u \rangle_{bar}&= { { m_n+m_\delta}\over 6}\cdot \left({{m_{\delta}-m_n}\over { m_{\sigma^{\scriptstyle * } } -m_\sigma } } - 1 \right ) = 190\,{\rm mev } \nonumber \\[4pt ] & { } \langle m_s - m_u \rangle_{mes}&= { { 3 m_\rho + m_\pi}\over 8 } \cdot \left({{m_\rho - m_\pi}\over{m_{k^*}-m_k } } - 1 \right ) = 178\,{\rm mev } \ , , \label{ight : mass}\end{aligned}\ ] ] the same value @xmath30 for @xmath27 is obtained from four independent calculations .
the same approach for @xmath31 gives @xmath32 & & \langle m_b - m_c \rangle_{mes } = { { 3(m_{b^{\scriptstyle * } } -m_{d^{\scriptstyle * } } ) + m_b - m_d}\over 4 } = 3339 \,{\rm mev } \label{heavy : m}\end{aligned}\ ] ] the same value @xmath33 for the ratio @xmath34 is obtained from meson and baryon masses .
@xmath35 derujula , georgi and glashow@xcite in 1975 used qcd arguments to relate hyperfine splittings to quark masses and baryon magnetic moments .
this led to remarkable agreement with experiment including three magnetic moment predictions with no free parameters @xmath36 & & { } \mu_p+\mu_n= 0.88 \,{\rm n.m . }
= { m_{\scriptstyle p}\over 3m_u } = { 2m_{\scriptstyle p}\over m_n+m_\delta}=0.865 \,{\rm n.m . } \nonumber \\[4pt ] & & { } -1.46 = { \mu_p \over \mu_n } = -{3 \over 2}\ , , \label{mag : mom}\end{aligned}\ ] ]
the meson - baryon puzzle - the @xmath39 and @xmath40 forces must be peculiarly related to bind both mesons and baryons .
it can not be a vector interaction giving equal and opposite forces , nor a scalar or tensor giving equal attractions for both . 2 .
exotics puzzle - no low - lying hadrons with exotic quantum numbers have been observed ; e.g. no @xmath41 or @xmath42 bound states .
nambu solved both puzzles@xcite in 1966 by introducing the color degree of freedom and a two - body interaction from a non - abelian gauge theory with the color - factor of one - gluon exchange .
this both related mesons and baryons and eliminated exotics .
a unified treatment of @xmath39 and @xmath40 interactions binds both mesons and baryons with the same forces . only @xmath37 and @xmath38 are stable in any single - cluster model with color space factorization .
any color singlet cluster that can break up into two color singlet clusters loses no color electric energy and gains kinetic energy .
the nambu color factor does not imply dynamics of one - gluon exchange .
higher order diagrams can have same color factor looking beyond bag or single - cluster models for possible molecular bound states lipkin(1972 ) showed that the color - electric potential energy could be lowered in potential models by introducing color - space correlations ; e , g , @xmath43 at corners of a square , but not enough to compensate for the kinetic energy@xcite a large spin - dependent interaction @xmath44 300 mev but a very weak interaction @xmath44 2 mev binding normal hadrons
. @xmath45 the low - lying hadron spectrum is described by a linear effective mass term and a hyperfine interaction with a one - gluon exchange color factor .
the @xmath46 and @xmath47 states behave like neutral atoms with a strong color electric field inside hadrons and none outside .
no molecular bound states arise in the simplest cases .
a strong spin - dependent interaction is crucial to understanding the spectrum only color singlet and @xmath48 color factors arise in @xmath46 and @xmath47 .
the low - lying hadron spectrum provides no direct experimental information on @xmath49 and @xmath14 interactions needed for multiquark exotic configurations .
qcd is a great theory , but nobody knows how to connect it with experiment or which approximations are good . we need to construct instructive simplified models .
i often recall the response by yoshio yamaguchi at a seminar at the weizmann institute in 1960 when asked if there had been any thought at cern about a possible breakdown of qed at small distances : no .
. many calculations . no thought . " what can we learn from qed ; a great theory that everyone knows how to connect with experiment ? we know
how isolated free electrons behave and carry currents . but nobody could explain the fractional hall effect.until robert laughlin told us the hall current is not carried by single electrons !
it is carried by quasiparticles related to electrons by a complicated transformation .
nobody has ever seen an isolated free quark .
current quark fields appear in the standard model lagrangian .
but experiments tell us that baryons are @xmath37 and mesons are @xmath8 and these are not the quarks that appear in the qcd lagrangian .
nobody knows what these quarks are .
are they complicated quasiparticles related to current quarks by a complicated transformation ? .
is hadron spectroscopy waiting for laughlin ?
does qcd need another laughlin to tell us what constituent quarks are ?
one good wrong model that stays away from free parameters and may teach us something : a two - cluster @xmath50-wave @xmath51 diquark-@xmath52 triquark model@xcite for the @xmath5 that separates @xmath53 and @xmath54 pairs and eliminates their short range repulsive interaction ... its hidden - strangeness @xmath55 partner keeps the same triquark with the @xmath56 and @xmath57 @xmath58 partners of the @xmath51 diquark .
its mass is roughly@xcite @xmath59 \approx m(\theta^+ ) + m(\lambda ) - m(n ) + { 3\over 4}\cdot [ m(\sigma)-m(\lambda ) ] \approx 1775 \,{\rm mev}\ ] ] experimental search motivated by another wrong model .
skyrmion model has no simple connection with quarks except by another wrong model .
the @xmath60 expansion invented@xcite pre - qcd to explain absence of free quarks .
the binding energy of @xmath8 pairs into mesons @xmath61 . at large @xmath62 the cross section for meson - meson scattering breaking up a meson into its constituent quarks is @xmath63 \approx g^2 { { e_m}\over{n_c } } \approx 0\ ] ] but @xmath64 ; @xmath65 this is not a small parameter !
some experiments@xcite see the @xmath5 ; others@xcite definitely do not .
further analysis is needed to check presence of specific production mechanisms in experiments that see it and their absence in those that do not@xcite .
no theoretical model addresses this question .
comprehensive review@xcite analyzes different models .. the reported@xcite @xmath66 can be the @xmath67-wave excitation of the @xmath68.with a @xmath57 diquark in a @xmath67-wave with the same @xmath69 triquark .
its dominant decays would be @xmath70 via the diquark transition @xmath71.and @xmath72 via @xmath73 .
decays like @xmath74and @xmath75 would be suppressed by the centrifugal barrier forbidding a quark in the triquark from joining the diquark .
some experimental checks of this mechanism are : 1 .
experiments which see the @xmath5 and have sufficient energy for producing the @xmath66 should look for an accompanying @xmath76 or @xmath77 and examine the mass spectrum of the @xmath78 and @xmath79 systems .
2 . experiments should look for @xmath72 .
3 . experiments searching for the @xmath5 should check possible production of a @xmath78 or @xmath79 resonance in the 2.4 gev region . @xmath80-decay
modes suggested for pentaquark searches@xcite would not produce this 2.4 gev @xmath55 .
similar considerations should be applied to searches in @xmath0 and @xmath81 like those proposed in ref .
4 . the other @xmath66 decay modes . ; e.g. @xmath82 , @xmath83 , @xmath84 , @xmath85 , are suppressed by the centrifugal barrier in the d - wave diquark - triquark model but may be appreciable .. finding them would would give further evidence for this model for pentaquark production .
the relative branching ratios would also provide information about the structure of th @xmath66 . 1 .
the angular distribution of the kaon emitted with the @xmath5 in @xmath86 @xcite carries interesting information .
production from a cryptoexotic @xmath55 , gives no forward - backward kaon asymmetry .
meson exchange gives forward peaking .
baryon exchange gives backward peaking , produces the @xmath5 equally by photons on protons and neutrons . and the same baryon exchange should be seen@xcite in @xmath87 .. 2 . the more complicated angular distributions in @xmath88
@xcite may still carry interesting information . +
all the above discussion for @xmath89 applies to the angular distribution of a @xmath90.in @xmath91 .
models@xcite with a suppressed @xmath92 coupling relative to @xmath93 predict stronger @xmath5 production with a backward @xmath94 than with a backward kaon .. in @xmath95 @xcite , the pion goes forward and everything else is in the target fragmentation region .
@xcite .. 1 .
search for exotic positive - strangeness baryon exchange in normal nonexotic reactions .
the baryon exchange diagram@xcite for @xmath5 photoproduction with an outgoing kaon is simply related to backward @xmath96 charge - exchange @xcite .
the lower @xmath97 vertices are the same ; the upper vertex is also @xmath97 for @xmath96 charge - exchange but @xmath98 for @xmath5 photoproduction .
if this diagram contributes appreciably to @xmath5 photoproduction , the contribution of the @xmath97 vertex is appreciable and should also contribute appreciably to backward @xmath99 charge - exchange .
some previously ignored backward @xmath99 charge - exchange data may be available .
the baryon and @xmath100 constituents of the @xmath5 are already initially present in low - energy photoproduction experiments in the target baryon and the @xmath100 component of the photon . in experiments where baryon number and strangeness must be created from gluons ,
the cost of baryon antibaryon and @xmath101 production by gluons must be used to normalize the production cross section in comparison with the photoproduction cross sections ; e.g. from baryon - antibaryon production and @xmath102 production data in the same experiment that does not see the @xmath5 .
the original work reported in this talk was in collaboration with marek karliner .
this work was partially supported by the u.s .
department of energy , division of high energy physics , contract w-31 - 109-eng-38 t. nakano _ et al . _
[ leps coll .
] , phys .
lett . * 91 * , 012002 ( 2003 ) , hep - ex/0301020 .
v. kubarovsky _ et al . _ [ clas coll . ]
, [ phys .
lett . * 92 * , 032001 ( 2004 ) ] erratum ibid . *
92 * , 049902 ( 2004 ) , hep - ex/0311046 .
t. nakano , talk at nstar 2004 , march 24 - 27 , grenoble , france , .
t. e. browder , i. r. klebanov and d. r. marlow , hep - ph/0401115 . s. armstrong , b. mellado and s. l. wu , hep - ph/0312344 .
m. battaglieri [ presenting clas coll .
data ] , talk at pentaquark workshop , feb .
10 - 12 , 2004 , trento , italy , | is there a theory or good experimental evidence ? bj s question : pentaquark is created by @xmath0 . @xmath1 baryon @xmath2 triquark@xmath3triquark @xmath4pentaquark does it live long onough to be observable ?
basic physics of constituent quarks and flavor antisymmetry .
report of @xmath5 violating flavor antisymmetry indicates need for two - cluster model .. ball in experimental court - some experiments see @xmath5 ; others do nt .
possible production mechanisms present in some experiments , absent in others ; e.g. via @xmath6 ?
= 11.6pt |
in the standard theory of formation history , matter evolved from small perturbations in the primordial density field into a complex structure of sheets and filaments with galaxy clusters at the intersections of this filamentary structure .
galaxy surveys , such as the 2df - grs ( 2 degree field galaxy redshift survey , * ? ? ?
* ) and the sdss ( sloan digital sky survey , * ? ? ?
* ) , have revealed a complex network of filamentary nature , which has become necessary known as the cosmic web @xcite . in the evolution of the cosmic web , baryons in the diffuse intergalactic medium accelerate towards dark matter halos under the growing influence of gravity and
go through shocks that heat them .
these cosmological shocks are a ubiquitous consequence of cosmic structure formation .
they are tracers of the large - scale structure and contain information about its dynamical history .
gas in walls and filaments follows the gravitational potential towards clusters of galaxies , colliding with the intracluster medium ( icm ) at speeds of @xmath9 km s@xmath10 .
while cosmological shocks heat the icm to temperatures of @xmath11 k , filaments are heated to temperatures of @xmath12 k , making their detection at present time challenging , since at this temperature there are hardly any emission lines .
therefore , the study of shocks may provide an independent and complementary way to study the warm - hot intergalactic medium ( whim ) , the low density intergalactic medium of the cosmic web that is believed to host the majority of the fraction of the baryon density .
some cosmological shocks are associated with diffuse radio emission caused by electrons accelerated to relativistic velocities by fermi acceleration @xcite .
this diffuse radio emission , without galaxy counterpart , is usually divided into two classes , namely _ radio halos _ and _ radio relics_. radio halos are unpolarized and have diffuse morphologies that are similar to those of the thermal x - ray emission of the cluster gas @xcite .
they are usually found in the center of clusters with significant substructure ( see e.g. , * ? ? ?
radio relics , on the other hand , are typically located near the periphery of the cluster .
they have been observed in several merging galaxy clusters .
they often exhibit sharp emission edges and many of them show strong radio polarization .
once accelerated , the electrons are short - lived because of the inverse compton scattering and synchrotron energy losses , and their spectrum rapidly steepens from the shock edge @xcite the synchrotron nature of this radio emission indicates the presence of cluster magnetic fields of the order of @xmath13 @xmath8 g @xcite .
cosmic shocks in large - scale structures have been investigated in a number of semi - analytical @xcite and numerical work @xcite .
some studies have focused on the non - thermal emission from galaxy clusters by modeling discretized cosmic ray ( cr ) energy spectra .
a series of papers explored the dynamical impact of cosmic ray ( cr ) protons on hydrodynamics in cosmological sph simulations @xcite .
observationally , detecting shocks waves in large - scale structures is still challenging , since they usually develop in the external regions of galaxy clusters , where the x - ray emission is faint .
however , a few merger shocks have been detected in nearby x - ray bright galaxy clusters @xcite and may be possible associated with a single or double radio relics discovered in a number of galaxy clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in numerical studies ,
@xcite studied the quantitative properties of large - scale shocks produced by gas during the formation of cosmic structures by means of hydrodynamical simulations .
they showed that shocks form abundantly in the course of structure formation and their topology is very complex and highly connected .
they also stated that considering the large size and long lifetime of shocks , they are potentially interesting sites for cosmic - ray acceleration .
@xcite ( see also @xcite ) found that shocks form around sheets , filaments and knots of mass distribution when the gas in void regions accretes onto them .
@xcite studied the properties of structure formation shock waves in cosmological sph simulations , allowing them to study their role in the thermalization of the plasma as well as for the acceleration of relativistic crs through diffusive shock acceleration .
they find that most of the energy is dissipated in weak internal shocks , with mach number @xmath14 . on the other hand , collapsed cosmological structures are surrounded by external shocks with much higher mach numbers ( @xmath15 ) , but they play only a minor role in the energy balance of thermalization .
@xcite computed the production of crs in a cosmological simulation volume , finding that crs are dynamically important in galaxy clusters .
they also found that shocks with low mach number typically trace mergers and complex flows , while those with mild and high mach number ( @xmath16 ) generally follow accretion onto filaments . in a similar study
, @xcite analyzed the properties of large - scale shocks .
they find that the bulk of the energy in galaxy clusters is dissipated at weak shocks , with mach numbers @xmath17 , although slightly stronger shocks are found in the external regions of merging clusters .
@xcite used gadget simulations of a sample of galaxy clusters , implementing a formalism for cr physics on top of radiative hydrodynamics .
they modeled relativistic electrons that are accelerated at cosmic formation shocks and produced in hadronic interactions of crs with protons of the icm .
they found that the radio emission in clusters is dominated by secondary electrons . only at the location of strong shocks the contribution of primary electrons
may dominate .
later , @xcite studied the radio emission for a sample of 10 galaxy clusters extracted from zoomed cosmological simulations .
they determined the radio luminosity , spectral index and faraday rotation measure for regions of the relics and concluded that upcoming radio telescopes , such as lofar , mwa , lwa and ska will discover a substantially larger sample of radio relics .
ska should probe the macroscopic parameters of plasma physics in clusters .
@xcite investigated the diffuse radio emission from clusters in cosmological sph simulations .
they found that the maximum diffuse radio emission in clusters depends strongly on their x - ray temperature .
they also found that the so - called accretion shocks cause only very little radio emission .
they conclude that a moderate efficiency of shock acceleration , namely @xmath5 , together with moderate magnetic field , namely 0.07 - 0.8 @xmath8 g , in the region of radio relic are sufficient to reproduce the number density and luminosity of radio relics .
@xcite , using amr simulations , also studied radio emission in galaxy clusters .
they investigated scaling relations between cluster parameters such as synchrotron power , mass and x - ray luminosity . in this paper
we follow the approach of @xcite but , instead of focusing on galaxy clusters , we investigate whether it is possible to detect radio emission in the entire cosmic web . due to the low temperature of the accreting gas , the mach number of external shocks - filaments - is high , extending up to @xmath15 or higher @xcite
. therefore , the total radio power generated by infall along filaments will be significantly weaker than infall into a cluster .
we will try to determine this total radio power . to this end
, we use the approach for detecting and identifying shock fronts and their characteristic mach number developed by @xcite .
this shock finder is applied to a large @xmath18-body / sph simulation . for computing the radio emission
, we follow the model elaborated in @xcite .
they assumed that electrons are accelerated by diffuse shock acceleration and cool subsequently by synchrotron and inverse compton losses .
consequently , the radio emission can be expressed as a function of downstream plasma properties , mach number and surface area of the shock front . in order to correctly assign a surface area to the sph particles
, we calibrate the radio emission using shock tubes experiments . of primary importance
is the proper identification of the different structures that reside in the cosmic web . to this end , we apply the spineweb technique ( @xcite , see also @xcite ) to the entire simulation .
the spineweb procedure correctly identifies voids , walls , filaments and clusters present in the cosmic web .
it is a powerful method that deals with the topology of the underlying density field of the cosmic web .
this allows us to study shock properties and radio emission in the different structures that constitutes the cosmic web .
this goes beyond the traditionally internal versus external shock classification suggested by @xcite and the method of characterizing shocks by their pre - shock overdensity and temperature proposed by @xcite . applying the radio emission model to
the simulation leads to radio - loud shock fronts which we are able to locate within the cosmic web .
this paper is organized as follows . in section [ sec : computer_simulation ] we summarize the characteristics of the @xmath18-body / sph simulation .
section [ sec : spineweb ] describes the spineweb method . in section [ sec : hydroshocks ] we give a brief description of the physics involving non - radiative shocks and depict how we find shocks in sph simulations . in order to calibrate the radio emission model
, we use shock tube tests which are described in section [ sec : shocktubes ] .
the radio emission model is sensitive to the shock surface area . using the results of the shock tubes , in section [ sec : shockarea ]
we described the method used in order to precisely determine the area of the shock front . in section
[ sec : results ] we present our results for the shock fronts and radio emission in the marenostrum universe . in section [
sec : conclusions ] we summarize our findings .
in order to study cosmological shock waves in the large - scale structure of the universe , we use the _ marenostrum universe _
simulation @xcite , one of the largest cosmological simulations available .
it assumes a standard flat @xmath19cdm universe with cosmological parameters @xmath20 , @xmath21 , @xmath22 and @xmath23 , where the hubble parameter is given by @xmath24 km s@xmath10mpc@xmath10 .
the normalization of the power spectrum is @xmath25 .
the simulation box has a side length of 500@xmath26mpc and contains 1024@xmath27 gas and 1024@xmath27 dark matter particles , with masses of @xmath28m@xmath29 and @xmath30m@xmath29 .
the initial conditions are followed from a redshift of @xmath31 until the present time ( @xmath32 ) using the massively parallel tree @xmath18-body / smoothed particle hydrodynamics ( sph ) code gadget-2 @xcite .
the results presented here are at present time ( @xmath32 ) .
radiative processes or star formation are not included in the simulation .
the plummer - equivalent softening was set at @xmath33 @xmath26kpc in comoving units , and the sph smoothing length was set to the distance to the 40th nearest neighbor of each sph particle .
smoothing scales are not allowed to be smaller than the gravitational softening of the gas particles . to extract the groups and galaxy clusters present in the simulation , we use hop @xcite .
hop assigns a density to every particle by smoothing the density field with a spline cubic kernel using the @xmath34 nearest neighbors of a given particle .
particles are then linked by associating each particle to the densest particle from the list of its closest neighbor .
this process is repeated until it reaches a particle that is its own highest density neighbor .
all particles linked to a local density maximum are identified as a group . at this stage
, no distinction between a high - density region and its surrounding has been make . to identify halos above a certain density threshold ,
a regrouping merging procedure is performed .
the code first includes only particles that are above some density threshold .
then , it merges all groups for which the boundary density between them exceeds a certain density value . finally , all groups identified must have one particle that exceeds a density peak to be accepted as an independent group .
we associate this density peak with the virial density needed for a spherical region to be in virial equilibrium , @xmath35 .
the value of @xmath35 is obtained from the solution to the collapse of a spherical top - hat perturbation under the assumption that the object has just virialized .
for the cosmology described here , at present time its value is @xmath36 . in using hop , we use the dark matter particles of the marenostrum simulation .
once we identified the virialized halos of the sample , we take the position of the densest dark matter particle as a first estimate of the center of mass . subsequently , we add the gas particles , grow a sphere around the center of mass and begin iterating , shrinking the sphere around the ( new ) center of mass until we reach a minimum of 50 particles .
this ensures a correct identification of the center of mass taking into account gas and dark matter particles .
once we have the final center of mass , we grow a sphere around it that encloses a value of @xmath36 .
given that we want to study radio emission in the cosmic web , the nodes of the filaments will be consider as galaxy clusters with masses @xmath37m@xmath38 .
we find that the marenostrum universe contains 3865 clusters of galaxies , the most massive with a mass of @xmath39m@xmath38 and virial radius @xmath40mpc .
the characterization of the large - scale structure was done using the spineweb method ( for a thorough explanation see @xcite , see also @xcite ) .
this method classifies the cosmic web into four basic morphological and dynamical components : voids , walls , filaments and clusters .
the spineweb method is based on the watershed transform @xcite and its cosmological implementation : the watershed void finder ( wvf ) @xcite .
the idea behind the wvf is to segment the density field into individual basins by `` flooding '' it ( in analogy to a landscape ) .
the spineweb method takes this analogy and goes one step further by directly relating the topology of the density field and the boundaries between basins by observing the simple relation between number of adjacent basins and topology : walls correspond to the boundary between two voids .
filaments+clusters are found at the intersection of three or more voids , which translates to the intersection of three or more walls : @xmath41 the resulting morphological characterization is parameter - free .
the analysis presented here is also based on the most recent implementation of the spineweb method which extends the original formalism on a multiscale hierarchical fashion making it also scale - independent . in practice
we start our analysis by generating a low resolution version of the original simulation using the averaging procedure described in @xcite .
the lower particle resolution was chosen to correspond to a inter - particle separation of approx .
3@xmath26mpc at the initial conditions .
with this `` linear - regime '' filtering we avoid small - scale substructure at the present time while keeping the characteristic large - scale anisotropy of the cosmic web . from the particle distribution
we compute the density field using the dtfe method @xcite .
the field is sampled at 27 different points inside each voxel and its mean is used to estimate the density for that particular voxel . from the density field
we compute the watershed transform and identify the boundaries of the basins ( i.e. the watershed transform itself ) . finally for each voxel in the watershed
we apply the criteria outlined in eqn .
[ eq : neighbour_morphology ] .
the complete procedure takes a few minutes on a regular workstation .
the spineweb method provides a complete framework for the characterization of the cosmic web into its basic morphological constituents : voids , walls , filaments and clusters .
it has no free parameters and being fundamentally a topological measure it is highly robust against noise in the density field .
[ fig : spine ] shows an example of how the spineweb correctly identifies the different elements of the cosmic web . shown is a full 3d network of filaments ( red ) and walls ( blue ) .
it is clearly visible the three - dimensional nature of the filament - wall network .
filaments define an interconnected web and walls fill the spaces in between the filaments forming a closed `` watertight '' network of voids .
in the formation of structures in the universe , the bulk velocity of the gas flow often exceeds the local sound speed . as a result ,
the shock surface separates two regions : the _ upstream regime _ or pre - shock regime , and the _ downstream regime _ or post - shock regime .
the plasma in the upstream regime moves with a velocity @xmath42 towards the shock front , while the downstream plasma departs with velocity @xmath43 .
most of the incident kinetic energy flux is converted into thermal energy . as the plasma passes through the shock front , mass , momentum and energy fluxes
are conserved , which is expressed in the rankine - hugoniot relations : @xmath44 where @xmath45 denotes the mass density , @xmath46 is the pressure and @xmath47 is the specific internal energy , and the velocities are measured in the rest - frame of the shock surface .
the strength of a shock is given by the mach numbers @xmath48 , @xmath49 where @xmath50 denotes the upstream sound speed which depends on the the specific internal energy by @xmath51 .
combining eqns .
[ eq : rankine - hugoniot ] and [ eq : machnumber ] , and assuming that the plasma obeys the polytropic relation , we can write for the mach number @xmath52 where @xmath53 and @xmath54 are the compression and entropy ratios .
the difference between downstream and upstream velocity can be written as @xmath55 the mach number estimator , which will be explained in the next section , computes the entropy ratio and the ratio @xmath56 for each sph particle in the simulation . in order to obtain the mach number , we tabulate the relations between @xmath54 , @xmath53 , @xmath48 and @xmath57 , allowing us to simple read the mach number from a table . in order to calculate the radio emission present in the marenostrum simulation , the first step is to identify shock fronts , and from this , derive the mach number , downstream ( postshock ) temperature and electron density .
the method we employ is the one proposed and used by @xcite . for a detailed description of the method
, we refer to the mentioned paper . here
, we will give the basic details .
the first step consists of computing the entropy gradient , @xmath58 , for each sph particle .
the entropy gradient gives the direction of the shock normal pointing into the downstream direction .
we define an associated upstream and downstream position , @xmath59 where @xmath60 denotes upstream or downstream , with the downstream position in the opposite direction , @xmath61 is the position of the sph particle @xmath62 , @xmath63 is the smoothing length of particle @xmath62 and @xmath64 denotes the shock normal , @xmath65 .
the upstream and downstream velocity are computed , together with the internal energy and density , using the usual sph scheme .
the upstream velocity is given by @xmath66 , and the same for the downstream velocity , but using the downstream position .
@xmath67 is the velocity of the shock front in the rest - frame of the simulation .
it is important to notice that the velocity field can also show perpendicular components .
we compute these components in a similar way to the upstream and downstream velocities , @xmath68 , where the three vectors @xmath69 , @xmath70 and @xmath71 form an orthonormal base .
once this is done , for those particles that belong to a shock front we demand that the velocity has to be divergent in the direction of the shock normal , i.e. , @xmath72 .
also , the velocity difference in the directions perpendicular to the shock normal have to be smaller than that parallel to the shock , i.e. , @xmath73 .
finally , to compute the mach number , there are two options : it is possible to use the entropy ratio @xmath74 ( @xmath75 is the entropy ) or use the ratio @xmath76 , where @xmath77 is the sound speed in the upstream region ( see section [ sec : hydroshocks ] ) . in order to have a conservative estimate for the mach number
, the smaller value of these two options is used .
in this section , we will briefly describe the radio emission model . for a complete description ,
we refer to @xcite and @xcite .
the first assumption of the model is that radio emission is produced by electrons accelerated to ultra - relativistic speeds required for synchrotron and inverse compton radiation .
these electrons are accelerated to a power - law distribution that is related to the mach number from diffusive shock acceleration theory . as for the magnetic field in the downstream region , it is assumed that , on average , simply obeys flux conservation to follow : @xmath78 where @xmath79 is @xmath80 and @xmath81 is the number density in the downstream region .
this is a simple assumption since a detailed model for the generation and evolution of magnetic fields is complex and beyond the scope of this paper ( see , e.g. , @xcite for a recent study on magnetic fields in the coma cluster ) .
there are several indications that the magnetic eld intensity decreases going from the center to the periphery of a cluster .
the exact power with which it varies with density is unclear , however .
some simulations suggest a higher power than that given by flux conservation which is attributed to the work of a fluctuation dynamo ( see e.g. @xcite and @xcite ) .
faraday rotation measurements in the coma cluster @xcite give a best fit for @xmath82 .
such a power would decrease the luminosity of cluster shocks by about 20% . outside of clusters
there are even fewer observational constraints .
recent constraints from tev @xmath83-ray sources have placed lower limits on extragalactic magnetic fields ( egmfs ) .
the charged particles from pair cascades are deflected by egmfs , thereby reducing the observed point - like flux .
dolag et al .
( 2011 ) have calculated the fluence of 1es 0229 + 200 as seen by fermi - lat for different egmf .
the non - observation of 1es 0229 + 200 by fermi - lat suggests that the egmf fills at least 60% of space with fields stronger than @xmath84 g assuming that the source is stable for at least @xmath85 yr . the fields outside clusters , however , will have a negligible effect on the radio luminosity function since the by far the largest part of the radio luminosity originates from shocks in or near clusters .
the radio emissivity , @xmath86 , which depends on three quantities : i ) the energy of the electron energy ( @xmath87 ) , ii ) the observing frequency ( @xmath88 ) , and iii ) the magnetic field strength ( @xmath89 ) , is then given by : @xmath90 where @xmath91 is the electron density , @xmath92 is the fraction of energy dissipated by the electron at the shock front and @xmath93 is the compression ratio at the shock front .
the radiation spectral index is related to the electron spectral index by @xmath94 .
@xmath95 has a value of @xmath96 .
this magnetic field is defined as the magnetic field corresponding to the energy density of the cmb . as stated before , in order to calculate the radio emission , we need the mach number , @xmath48 , the downstream temperature , @xmath97 , the electron density , @xmath91 , the magnetic field , @xmath89 , and the shock surface area , @xmath98 .
the shock finder described in sect .
[ sec : shockfinder ] , besides locating shock discontinuities in a sph simulation , also provides estimates for the mach number and for the shock normal .
this normal vector gives a position which is sufficiently downstream to determine @xmath97 and @xmath91 .
as for the shock surface area @xmath98 , each particle represents an area of the shock , since a shock front is comprised of particle in a sph simulation .
the shock discontinuity is smoothed by the size of the sph kernel , @xmath99 , which contains @xmath100 particles , with a corresponding volume of @xmath101 .
therefore , one particle belonging to the shock front represents a shock area of @xmath102 where @xmath103 is a normalization constant which was set to be @xmath104 by @xcite using shock tube simulations . here , we will correct this factor also using shock tubes , as it was found that @xmath103 depends strongly on @xmath48 for low mach number shocks ( see next section ) . .
[ cols="^,^,^,^",options="header " , ] [ tab : number_radio_objects ] the use of the spineweb enable us to make predictions of the radio flux in filaments .
assuming an spectral index of @xmath105 , filaments at a redshift @xmath106 , in a frequency of 150 mhz , a radio flux of @xmath7 @xmath8jy .
we have analyzed one of the largest hydrodynamical simulations of cosmic structure formation , namely the marenostrum simulation , with the purpose of studying the possibility of detecting radio emission in the cosmic web . to that purpose ,
we use : i ) a novel method for estimating the radio emission of strong shocks that occur during the process of structure formation in the universe developed by @xcite and ii ) the spineweb technique the spineweb procedure deals with the topology of the underlying density field , correctly identifying voids , walls and filaments . while previous studies focused on the study of cosmological shocks in selected catalogs of galaxy clusters or by defining the different environments of the cosmic web by density and/or temperature cuts , we are able to correctly disentangle the complex filamentary network of the cosmic web . the computation of the radio emission is based on an estimate for the shock surface area .
the radio emission is computed per surface area element using the mach number of the shock and the downstream plasma properties . in order to properly calculate the radio emission ,
two corrections were made .
the first correction deals with the shock surface area .
the constant area factor ( @xmath107 ) of eqn .
[ eq : total_emission2 ] was corrected using shock tube tests and a convolution method described in appendix [ app : dsdm ] .
one deals with the temperature of the simulation .
given that the simulation was run with a low floor temperature and without uv background , expansion cooling leads to very low temperature , which leads to very high mach numbers .
we corrected this by scaling the temperature of cold , low density regions using @xmath108 @xcite , where @xmath109 is the overdensity .
the first step was to calculate the strength of cosmological shocks , characterized by their mach number using the method described in @xcite .
we find that each environment of the cosmic web has a distinct mach distribution : voids have a characteristic mach number of @xmath3 ; walls , @xmath2 ; filaments , @xmath1 ; and clusters , @xmath110 .
the strength of the shock , i.e. , the characteristic mach number , is closely related to the temperature and density of the medium .
voids are the coldest and less dense region in the universe , with a mean temperature @xmath111 k and @xmath112 , and they interact with walls and filaments , which are denser and hotter structures , with filaments reaching a mean temperature of @xmath113 k and @xmath114 . applying a radio emission model based on dsa to the simulation
we are able to reproduce the diffuse radio emission in galaxy clusters using a lower electron efficiency to that used by @xcite .
this diffuse radio emission is also present in filaments . in walls and voids the emission is very weak , resulting in almost no presence of radio objects . by identifying radio objects based on their density , we find that clusters hosts the majority of such objects .
assuming an electron efficiency of @xmath5 , we find that 30 galaxy clusters ( out of 3865 ) host 34 radio objects with @xmath115 ergs s@xmath10 hz@xmath10 .
if we consider a radio luminosity of @xmath116 ergs s@xmath10 hz@xmath10 , then 659 radio objects are found in 547 galaxy clusters , only @xmath117 of the total sample .
these radio objects are present in the entire mass range of the galaxy clusters , i.e. , both low mass and high mass clusters present radio objects .
filaments also host radio objects , although only 2 of them are very luminous ( @xmath115 ergs s@xmath10 hz@xmath10 ) .
however , the majority of low luminous radio objects ( @xmath118 ) are present in filaments .
the presence of radio objects in walls and voids is rather negligible .
we calibrated our simulated data against the observational nvss radio relic luminosity function , yielding an electron efficiency of @xmath4 .
this is a factor of 10 less than the one estimated by @xcite . when using this efficiency , the number of very luminous radio objects drop dramatically .
we only find 4 radio object with @xmath115 ergs s@xmath10 hz@xmath10 and 1 in filaments . in the range @xmath116 ergs s@xmath10 hz@xmath10
, we find 235 radio objects in 198 galaxy clusters ( 5% of the total sample ) and 43 in filaments .
an all - sky survey up to @xmath119 and assuming an efficiency of @xmath5 ( @xmath4 ) should result in the discovery of 510 ( 60 ) radio objects in 450 ( 45 ) clusters , and 30 ( 15 ) radio objects in filaments . with the increase of surface brightness sensitivity of the upcoming radio telescopes , the detection of radio objects should increase by a factor of 10 , which opens in the possibility of finding a considerable number of radio objects in filaments .
furthermore , we predict that the radio flux of filaments at redshift @xmath106 , and at a frequency of 150 mhz , should be @xmath120 @xmath8jy .
the authors would like to thank gustavo yepes for making the marenostrum simulation available .
paam gratefully acknowledges bernard jones for helpful discussions and support by the german federal ministry for education and research under the _
verbundforschung_. 99 aragn - calvo , m. a. , platen , e. , van de weygaert , r. , & szalay , a. s. 2010a , , 723 , 364 aragn - calvo , m. a. , van de weygaert , r. , & jones , b. j. t. 2010b , , 408 , 2163 aragn - calvo , m. a. , van de weygaert , r. , araya - melo , p. a. , platen , e. , & szalay , a. s. 2010c , , 404 , l89 bagchi , j. , enlin , t. a. , miniati , f. , stalin , c. s. , singh , m. , raychaudhury , s. , & humeshkar , n. b. , 2002 , new astronomy , 7 , 249 bagchi , j. , durret , f. , neto , g. b. l. , & paul , s. , 2006 , science , 314 , 791 battaglia , n. , pfrommer , c. , sievers , j. l. , bond , j. r. , & enlin , t. a. , 2009 , mnras , 393 , 1073 berrington , r. c. , & dermer , c. d. , 2003 , , 594 , 709 s. , lantuejoul c. , 1979 , in proceedings international workshop on image processing , ccett / irisa , rennes , france bonafede , a. , giovannini , g. , feretti , l. , govoni , f. , & murgia , m. , 2009 , , 494 , 429 bonafede , a. , feretti , l. , murgia , m. , govoni , f. , giovannini , g. , dallacasa , d. , dolag , k. , & taylor , g. b. 2010 , , 513 , a30 bond , j. r. , kofman , l. , & pogosyan , d. , 1996 , , 380 , 603 brggen , m. , ruszkowski , m. , simionescu , a. , hoeft , m. , & dalla vecchia , c. 2005 , , 631 , l21 burn , b. j. 1966 , , 133 , 67 carroll , s. m. , press , w. h. , & turner , e. l. 1992 , , 30 , 499 cassano , r. , ettori , s. , giacintucci , s. , brunetti , g. , markevitch , m. , venturi , t. , & gitti , m. 2010 , , 721 , l82 cen , r. , & ostriker , j. p. , 2006
, , 650 , 560 clarke , t. e. , & ensslin , t. a. 2006 , , 131 , 2900 colless , m. , et al . , 2003 , arxiv : astro - ph/0306581 condon , j. j. , cotton , w. d. , greisen , e. w. , yin , q. f. , perley , r. a. , taylor , g. b. , & broderick , j. j. 1998 , , 115 , 1693 dav , r. , et al . , 2001 , , 552 , 473 dolag , k. , bykov , a. m. , & diaferio , a. 2008 , , 134 , 311 dolag , k. , grasso , d. , springel , v. , & tkachev , i. 2005 , , 1 , 9 eisenstein d. j. , hut p. , 1998
, apj , 498 , 137 enlin , t. a. , pfrommer , c. , springel , v. , & jubelgas , m. , 2007 , , 473 , 41 ensslin , t. a. , biermann , p. l. , klein , u. , & kohle , s. 1998 , , 332 , 395 feretti , l. , fusco - femiano , r. , giovannini , g. , & govoni , f. 2001 , , 373 , 106 feretti , l. , burigana , c. , & enlin , t. a. 2004 , , 48 , 1137 feretti , l. , bacchi , m. , slee , o. b. , giovannini , g. , govoni , f. , andernach , h. , & tsarevsky , g. 2006 , , 368 , 544 feretti , l. , & giovannini , g. , 2008 , in _ a pan - chromatic view of clusters of galaxies and the large - scale structure _ , ed . m. plionis , o. lopez - cruz , d. hughes , lect .
notes phys .
, 740 , 143 ( springer ) gabici , s. , & blasi , p. , 2003
, , 583 , 695 giacintucci , s. , et al . , 2008 , , 486 , 347 giovannini , g. , tordi , m. , & feretti , l. 1999 , , 4 , 141 giovannini , g. , feretti , l. , govoni , f. , murgia , m. , & pizzo , r. , 2006 , astronomische nachrichten , 327 , 563 gottlber , s. , okas , e. l. , klypin , a. , & hoffman , y. 2003 , , 344 , 715 gottlber , s. , & yepes , g. , 2007 , apj , 664 , 117 govoni , f. , feretti , l. , giovannini , g. , bhringer , h. , reiprich , t. h. , & murgia , m. 2001 , , 376 , 803 hoeft , m. , & brggen , m. 2007 , , 375 , 77 hoeft , m. , brggen , m. , yepes , g. , gottlber , s. , & schwope , a. 2008 , , 391 , 1511 hui , l. , & gnedin , n. y. 1997 , , 292 , 27 jubelgas , m. , springel , v. , enlin , t. , & pfrommer , c. , 2008 , , 481 , 33 kang , h. , ryu , d. , cen , r. , & song , d. , 2005 , , 620 , 21 kardashev , n. s. 1962 , soviet astronomy , 6 , 317 keshet , u. , waxman , e. , loeb , a. , springel , v. , & hernquist , l. , 2003 , , 585 , 128 kim , k .-
, kronberg , p. p. , giovannini , g. , & venturi , t. 1989 , , 341 , 720 klypin a. , kravtsov a. v. , bullock j. s. , primack j. r. , 2001 , , 554 , 903 markevitch , m. , govoni , f. , brunetti , g. , & jerius , d. , 2005 , , 627 , 733 markevitch , m. , 2006 , the x - ray universe 2005 , 604 , 723 meli , a. , & biermann , l. p. , 2006
, nuclear physics b proceedings supplements , 151 , 501 miniati , f. , ryu , d. , kang , h. , jones , t. w. , cen , r. , & ostriker , j. p. , 2000 , , 542 , 608 miniati , f. , jones , t. w. , kang , h. , & ryu , d. , 2001 , , 562 , 233 miniati , f. , 2002 , navarro , j. f. , frenk , c. s. , & white , s. d. m. 1995 , , 275 , 720 platen , e. , van de weygaert , r. , & jones , b. j. t. 2007 , , 380 , 551 pfrommer , c. , springel , v. , enlin , t. a. , & jubelgas , m. , 2006 , , 367 , 113 pfrommer , c. , enlin , t. a. , springel , v. , jubelgas , m. , & dolag , k. , 2007 , , 378 , 385 pfrommer , c. , enlin , t. a. , & springel , v. , 2008 , , 385 , 1211 roettiger , k. , burns , j. o. , & stone , j. m. 1999 , , 518 , 603 rttgering , h. j. a. , wieringa , m. h. , hunstead , r. w. , & ekers , r. d. , 1997 , , 290 , 577 ryu , d. , kang , h. , hallman , e. , & jones , t. w. , 2003 , , 593 , 599 ryu , d. , kang , h. , cho , j. , & das , s. 2008 , science , 320 , 909 sarazin , c. l. , 1999 , , 520 , 529 schaap , w. e. , & van de weygaert , r. , 2000 , , 363 , l29 schaap , w. e. , 2007 , the delaunay tessellation field estimator , ph.d .
thesis , university of groningen sheth , r. k. , & van de weygaert , r. 2004 , , 350 , 517 sigl , g. , miniati , f. , & ensslin , t. a. 2003 , , 68 , 043002 skillman , s. w. , oshea , b. w. , hallman , e. j. , burns , j. o. , & norman , m. l. , 2008 , , 689 , 1063 skillman , s. w. , hallman , e. j. , oshea , b. w. , burns , j. o. , smith , b. d. , & turk , m. j. , 2010 , arxiv:1006.3559 sod , g. a. 1978 , journal of computational physics , 27 , 1 solovyeva , l. , anokhin , s. , feretti , l. , sauvageot , j. l. , teyssier , r. , giovannini , g. , govoni , f. , & neumann , d. , 2008 , , 484 , 621 springel v. , 2005 , mnras , 364 , 1105 tegmark , m. , et al . , 2004 , , 606 , 702 van de weygaert , r. , & van kampen , e. 1993 , , 263 , 481 van de weygaert , r. , & platen , e. 2009 , arxiv:0912.2997 van de weygaert , r. , & schaap , w. 2009 , data analysis in cosmology , 665 , 291 van weeren , r. j. , rttgering , h. j. a. , brggen , m. , & cohen , a. , 2009 , , 505 , 991 vazza , f. , brunetti , g. , & gheller , c. , 2009 , , 395 , 1333 venturi , t. , giacintucci , s. , brunetti , g. , cassano , r. , bardelli , s. , dallacasa , d. , & setti , g. 2007 , , 463 , 937 zeldovich , y. b. 1970 , , 5 , 84
convolution is a mathematical operation on two functions , @xmath121 and @xmath122 , which produces a third function that is a modified version of the one of the original functions .
usually , one of the two functions is taken to be a _ kernel _ function which acts on the other function , modifying it .
the standard convolution states that @xmath123 where @xmath124 is the convolution operand .
the inverse of this operation is called deconvolution . in our case , we have a measured distribution , namely @xmath125 , obtained from the simulation and we wish to obtain the _ real _ distribution , i.e. , the shock surface area distribution taking into account the particles that were assign a low mach number in strong shocks and the underestimation of the different mach numbers .
therefore , we need to find the shock tube _ distribution kernel _ taken from the shock tubes data in order to deconvolve it with our measure function .
however , instead of finding the inverse of the kernel function , we will iterate the convolution procedure , which will yield the real distribution .
the convolution we want to solve is @xmath126 where @xmath127 is the _ measured _ shock surface distribution , @xmath128 is the kernel function and @xmath129 is the _ real _ distribution we want to find . in eqn .
[ eq : our_convolution ] we have taken two things into account :
i ) there are not negative mach numbers , and ii ) the kernel function depends on the mach number .
in section [ sec : shocktubes ] we saw that a simple representation to the shock surface area of the shock tubes are a heaviside function and a gaussian .
therefore , our kernel has the form @xmath130 where @xmath131 is a gaussian with form @xmath132 where the standard deviation also depends on the measured mach numbers , i.e. , @xmath133 as seen in the previous section . after constructing the kernel with the fitting parameters of the various shock tubes , we give a guess function in order to start the iteration procedure . as a first guess , we choose @xmath134 , and convolve that with the kernel , giving as a result @xmath135 .
the new guess is then calculated by subtracting the previous guess with the difference between the result of the convolution of the first guess with the measured function , i.e. , @xmath136 .
we found that after iterating 5 times we are able to recover @xmath137 . | we explore the possibility of detecting radio emission in the _ cosmic web _ by analyzing shock waves in the marenostrum cosmological simulation .
this requires a careful calibration of shock finding algorithms in smoothed - particle hydrodynamics simulations , which we present here .
moreover , we identify the elements of the cosmic web , namely voids , walls , filaments and clusters with the use of the spineweb technique , a procedure that classifies the structure in terms of its topology . thus , we are able to study the mach number distribution as a function of its environment .
we find that the median mach number , for clusters is @xmath0 , for filaments is @xmath1 , for walls is @xmath2 , and for voids is @xmath3 .
we then estimate the radio emission in the cosmic web using the formalism derived in hoeft & brggen ( 2007 ) .
we also find that in order to match our simulations with observational data ( e.g. , nvss radio relic luminosity function ) , a fraction of energy dissipated at the shock of @xmath4 is needed , in contrast with the @xmath5 proposed by hoeft et al .
( 2008 ) .
we find that 41% of clusters with @xmath6 host diffuse radio emission in the form of radio relics .
moreover , we predict that the radio flux from filaments should be @xmath7 @xmath8jy at a frequency of 150 mhz .
[ firstpage ] cosmology : theory large - scale structure of the universe
hydrodynamics methods : numerical radiation mechanisms : nonthermal shock waves |
theories of anomalous diffusion and transport abound @xcite , including continuous time random walks ( ctrw)@xcite and related fractional fokker - planck equations @xcite , generalized langevin equations @xcite , anomalous kinetic equations of the ( linear ) boltzmann equation type @xcite , hopping and continuous dynamics in disordered and fractal media @xcite , etc .
often it is not clear from the experimental results which theory works better for the case of study . in particular , theory of anomalous response
is often based on ctrw with divergent mean residence times ( mrts ) .
often it is seen as a proper explanation of the physical origin of all major anomalous response functions , including cole - cole response @xcite featuring the so - called fast or @xmath1relaxation in complex liquids and glasses .
the latter one is relatively fast ( lasting from pico- to milliseconds , even if it follows to a power law ) , and the corresponding response quickly becomes stationary on the time scale of observations .
it is followed by @xmath2relaxation , which is often described by kohlrausch - williams - watts stretched exponential dependence , or even better @xcite by the cole - davidson relaxation law .
@xmath2relaxation extends up to @xmath3 seconds in a typical aging experiment @xcite , until a stationary response regime is reached @xcite .
stationary response requires to calculate stationary autocorrelation function ( acf ) .
if to do this properly for ctrw with divergent mean residence times , it becomes clear that such a response is absent , because the corresponding stationary acf does not decay at all .
this in turn indicates the breaking of ergodicity in accordance with the slutsky theorem @xcite .
non - ergodic ctrw can not respond to stationary perturbations in a very long run .
its response must die to zero , which invalidates a part of the ctrw based theories of anomalous stationary response , when applied to physical systems at thermal equilibrium .
we discuss these subtleties below within a sufficiently simple and generic model .
linear response theory has been put by kubo on the firm statistic - mechanical foundation @xcite . consider a physical system with dynamical variable @xmath4 , say coordinate of a particle moving in a multistable potential in a dissipative environment .
it exhibits stochastic dynamics ( here classical ) due to thermal agitations of the medium at temperature @xmath5 , with the ensemble average @xmath6 ( we bear in mind many independent particles ) .
the system is at thermal equilibrium .
one applies a periodic driving force @xmath7 with frequency @xmath8 and amplitude @xmath9 linearly coupled to the dynamical variable @xmath4 , with perturbation energy @xmath10 .
the mean deviation from equilibrium , @xmath11 , responds generally to the driving switched on at the time @xmath12 with some delay @xcite , @xmath13 where @xmath14 is the linear response function ( lrf ) , and the upper integration limit @xmath15 reflects causality . for any physical system at thermal equilibrium and
weakly perturbed by driving , lrf can depend only on the difference of time arguments .
stationary response to the periodic signal reads ( @xmath16 ) @xmath17 where @xmath18 is the ( laplace-)fourier transform of @xmath19 ( @xmath20 for @xmath21 due to causality ) , and @xmath22 is the phase lag .
kubo derived from microscopic hamiltonian dynamics under fairly general conditions that such stationary lrf is related to equilibrium autocorrelation function , or rather covariance @xmath23 of the variable @xmath4 as @xcite @xmath24 where @xmath25 is the heaviside step function .
this fundamental result in statistical physics is known as classical fluctuation - dissipation theorem ( fdt ) in the time domain .
stationary response is related to the equilibrium autocorrelation function which must be stationary .
a generalization of this result to a nonstationary , e.g. aging environment , or in the presence of steady state nonequilibrium fluxes in the unperturbed state ( non - equilibrium steady state
ness ) is not trivial .
especially , the systems exhibiting non - gaussian anomalous diffusion demonstrate profound deviations from equilibrium fdt behavior in ness @xcite .
non - equilibrium fdt was developed primarily for the systems with underlying markovian dynamics @xcite .
it allows , for example , to rationalize the ness response of anomalous 1d subdiffusive dynamics to a small additional tilt on a comb lattice @xcite .
this dynamics is embedded as 2d markovian dynamics @xcite . a generalization to intrinsically non - markovian dynamics is yet to be done . in the present work
, we are focusing on a standard response of unperturbed systems being at thermal equilibrium , without dissipative fluxes present .
specifically , our focus is on anomalous non - debye type response featuring many materials . as an important example ,
famous cole - cole response function @xcite reads ( setting the high - frequency component @xmath26 to zero ) @xmath27 in the frequency domain , with static @xmath28 and @xmath29 .
the cole - cole response corresponds to the mittag - leffler relaxation @xcite of the autocorrelation function @xmath30,\end{aligned}\ ] ] where @xmath31=\sum_{n=0}^{\infty}z^n/\gamma(\alpha n+1)$ ] is mittag - leffler function @xcite and @xmath32 is ( anomalous ) relaxation time , which in the range of from picoseconds to milliseconds for various materials @xcite .
if a constant signal @xmath33 is switched on at @xmath12 , then the time - dependent response in accordance with fdt is @xmath34\;.\end{aligned}\ ] ] it starts from zero and approaches the asymptotical value @xmath35 , if @xmath36 with @xmath16 .
we model further the thermal dynamics of @xmath4 by a non - markovian ( semi - markovian ) ctrw of the following type .
there are @xmath37 localized states @xmath38 which are populated with stationary probabilities @xmath39 .
scattering events due to a thermal agitation are not correlated and distributed in time with the waiting time density ( wtd ) @xmath40 , which is normalized and has a finite mean time @xmath41 between two subsequent events .
scattering can lead to any other state , but the particle can remain also in the same state .
the finiteness of @xmath41 is crucial for calculating the stationary acf .
then , we can study the response of our system in the limit @xmath42 , _ after _ the stationary acf @xmath43 is found and not _ before _ , overcoming thereby the common fallacy done in huge many previous works ( they are simply not based on the correct stationary acf ) . the transition probability from the state @xmath44 to the state @xmath45
is @xmath46 , i.e. it corresponds to the statistical weight of the next trapping state .
such a ctrw is a particular version of separable ctrw by montroll and weiss @xcite .
this is a toy model , but the main features are expected to be generic . to calculate the stationary and equilibrium averages the distribution of the first residence time
is required @xcite .
it is well - known to be @xcite @xmath47 ( see also below in sec .
[ aging_part ] for a derivation of this result ) , where @xmath48 is survival probability .
the survival probability for the first sojourn without scattering events is @xmath49 .
it plays the central role .
we show in appendix that @xmath50 within this fully decoupled ctrw model .
now we can ask the question if such a probability density @xmath40 exists within this model that we can obtain the cole - cole response exactly .
the answer is no .
indeed , the laplace transform of @xmath51 is related to the laplace transform of survival probability @xmath52 as @xmath53\;.\ ] ] for the mittag - leffler relaxation @xmath54 $ ] . from this and ( [ rel1 ] ) , @xmath55 , which does not correspond to any survival probability since @xmath56 , and not @xmath57 , as must be .
remarkably , within the framework of langevin dynamics with memory the cole - cole response is easily reproduced by fractional overdamped langevin dynamics in parabolic potential @xcite . from eqs .
( [ acf1 ] ) and ( [ rel1 ] ) it follows immediately in the formal limit @xmath58 that @xmath59 for any wtd , which means that @xmath60 it does not decay at all ! from this in ( [ fdttime ] ) it follows immediately that stationary response is absent exactly , @xmath61 the `` no stationary response '' theorem for a rather general class of processes with infinite mean residence time is thus proven .
it agrees with the death of response found earlier for other non - ergodic ctrw dynamics @xcite .
in fact , it follows at once by the fluctuation - dissipation theorem ( [ fdttime ] ) from nonergodicity of any process @xmath4 with divergent mean residence times @xcite .
the breaking of ergodicity in turn follows by the virtue of slutsky theorem from nondecaying character of stationary autocorrelation function @xcite .
non - ergodic system with @xmath62 can not respond stationary to a stationary ( e.g. periodic ) signal , and the corresponding theories of linear response at thermal equilibrium are deeply flawed .
however , it is more instructive to investigate the way how the stationary response becomes increasingly suppressed with increasing @xmath41 .
two particular models are interesting in this respect .
the first model corresponds to the mittag - leffler density tempered by introduction of an exponential cutoff @xcite .
the corresponding survival probability is @xmath63\exp(-r t)$ ] , where @xmath64 is fractional rate and @xmath65 is a cutoff rate .
all the moments then become finite .
the survival probability has laplace transform @xmath66 the mean time is @xmath67=\langle \tau \rangle$ ] , and @xmath68 clearly , it also exhibits the generic death of stationary response in the limit @xmath69 .
we shall not elaborated this model in more detail here , bur rather focus on a very different model with finite @xmath41 , but divergent higher moments .
it exhibits a number of other surprising and interesting features .
we assume that transition events can occur either with normal rate @xmath65 , or with some fractional rate @xmath70 ( @xmath71 ) intermittently through @xmath72 independent relaxation channels .
this yields the following survival probability @xcite , @xmath73 which can also be written as @xmath74 where @xmath75 is the mean relaxation time ( defined exclusively by the normal relaxation channel ) and the time constants @xmath76 provide the spectrum of anomalous relaxation times . specifying the generalized master equation ( gme ) by kenkre , montroll , and shlesinger @xcite for this model ( see e.g. appendix in ref .
@xcite for a general derivation from the trajectory perspective ) yields fractional relaxation equation @xmath77,\end{aligned}\ ] ] where @xmath78 defines the operator of fractional riemann - liouville derivative of the order @xmath79 @xcite .
( [ fme ] ) belongs to a general class of gmes with distributed fractional derivatives @xcite .
one relaxation channel must be normal , @xmath80 , @xmath81 for ergodic process .
evolution of state probabilities @xmath82 become completely decoupled because of our choice of the transition probabilities , @xmath46 . from eq .
( [ fme ] ) , @xmath39 are obviously the stationary probabilities , @xmath83 , which is not immediately obvious from the trajectory description . for normal relaxation , our choice of model corresponds to a very popular textbook model of single relaxation time @xmath41 , which does not depend on the particular state .
it is often used as the simplest approximation in studying various kinetic equations . by the same token , anomalous relaxation times
also do not depend on the particular state within the model considered .
notice , however , that gme ( [ fme ] ) can not be used to find the stationary acf ( [ acf1 ] )
. its direct use would be namely the typical fallacy which plagued many earlier works .
one must proceed differently , see in appendix for details . for the simplest nontrivial case @xmath84 with @xmath85 , @xmath86 , and in the parameter regime @xmath87 , @xmath88 , on the time scale @xmath89 .
it corresponds to the same wtd as in the basic ctrw model with divergent @xmath41 .
hence , for @xmath90 , @xmath91 .
this initial power law corresponds to the initially stretched exponential survival probability @xmath92 .
then , intermediate power law follows , @xmath93 , for @xmath94 within the range @xmath95^{1/\alpha } \tau_r \ll\tau\ll \tau_r \;,\end{aligned}\ ] ] i.e. over @xmath96^{1/\alpha}$ ] intermediate time decades . for @xmath97 , it further changes into the asymptotic power law @xmath98 @xcite .
interestingly enough , introduction of a finite @xmath41 results into a sufficiently strong power law decay for @xmath99 .
the parameter @xmath32 plays thus the role of a cutoff time , though the cutoff character is very different from the model of exponentially tempered distribution .
the latter one has an exponential cutoff for @xmath100 .
the important parameter regime , @xmath87 , is depicted in fig .
notice , however , that for @xmath101 , the intermediate power law first disappears , and for @xmath102 it transforms into intermediate exponential decay ( like one in fig . [
fig2],b ) .
on time ( in units of @xmath32 ) for @xmath103 and @xmath104 exhibits three different power laws , @xmath105 , @xmath106 , @xmath107 , in the parameter regime @xmath108 .
, width=283 ] the laplace - transformed stationary acf of this model reads @xmath109 from which it becomes immediately clear that in the limit @xmath110 , the limiting stationary acf is simply constant , as in eq .
( [ acf - ml ] ) .
the stationary response is absent .
furthermore , the response function in laplace domain follows as @xmath111 in frequency domain it is @xmath112 .
we have thus a nice and nontrivial generalization of cole - cole response function which includes one normal relaxation time @xmath113 and @xmath114 anomalous .
it provides a very rich model of anomalous response based on ergodic ctrw dynamics with finite mean waiting time .
we consider further the particular case of one anomalous relaxation channel , @xmath84 .
then , the cole - cole response with index @xmath79 instead of @xmath115 is reproduced in the limit @xmath116 at fixed anomalous relaxation time @xmath86 .
strikingly enough , this is the opposite limit with respect to one considered in the theory of ctrw with infinite @xmath117 .
indeed , for sufficiently small @xmath108 , @xmath118 $ ] , see in fig .
this corresponds to the cole - cole response with exponent @xmath79 providing a very important result : stationary cole - cole response emerges within the ctrw approach if there exists a very fast normal relaxation channel acting in parallel to the anomalous one and making the mean transition times finite . in striking contrast to the traditional ctrw models based on @xmath93 dependence with divergent mean residence times our approach yields the cole - cole response with exponent @xmath79 instead of @xmath115 .
furthermore , it is instructive to see how @xmath43 and the response function @xmath119 , behave for large but finite @xmath120 with @xmath32 kept constant .
autocorrelation function is plotted in fig .
[ fig2 ] for @xmath121 .
notice that it has a heavy power law tail @xmath122 , onset of which moves to larger @xmath15 with the increase of @xmath113 .
for very large @xmath123 , the initial decay is nearly exponential , @xmath124 . with this in eq .
( [ static ] ) , one can clearly see that on the time scale @xmath125 , the response of stationary equilibrium environment is not realized , being increasingly suppressed with the increase of @xmath113 .
time - dependent signals with frequencies smaller than a corner frequency @xmath126 should be regarded as slow .
however , the response to them ( even if it does exist asymptotically ! ) can not be detected in noisy background as we shall clarify soon .
the dependence on @xmath115 is very important and remarkable .
it is very different from one of ctrw theory with infinite mean time .
basically , we have @xmath115 instead of @xmath79 .
this means that with @xmath115 close to one , e.g. @xmath127 , the corresponding power noise spectrum @xmath128\nonumber \\ & & = 2\langle \delta x^2(0)\rangle_{\rm eq}\frac{\tau_r(\omega\tau_r)^\alpha \cos(\frac{\alpha \pi}{2 } ) + \langle \tau \rangle ( \omega\tau_r)^{2\alpha}}{a(\omega ) } , \end{aligned}\ ] ] where @xmath129 becomes very close to @xmath0 noise at small frequencies , @xmath130 , see in fig . [ fig3 ] .
notice that the closer the anomalous relaxation channel to the normal one the closer the power spectrum becomes to one of @xmath0 noise !
this provides a very nice and physically plausible explanation of the wide presence of @xmath0 noise in condensed matter @xcite .
the second relaxation channel must be most similar to the normal one and not mostly deviating . ) on time ( in units of @xmath32 ) for different @xmath41 and @xmath121 .
the tail is always @xmath131 , see in part a ) .
decay of correlations is dramatically delayed with the increase of @xmath41 .
initially , it is exponential for @xmath132 , @xmath124 , as part b ) shows .
the limit @xmath133 , corresponds to the mittag - leffler decay of correlations , @xmath134 $ ] and cole - cole response .
part b ) represents some of the curves in part a ) on semi - logarithmic plot revealing initial exponential decay for large @xmath41 .
+ , title="fig:",width=283 ] + ) on time ( in units of @xmath32 ) for different @xmath41 and @xmath121 .
the tail is always @xmath131 , see in part a ) .
decay of correlations is dramatically delayed with the increase of @xmath41 .
initially , it is exponential for @xmath132 , @xmath124 , as part b ) shows .
the limit @xmath133 , corresponds to the mittag - leffler decay of correlations , @xmath134 $ ] and cole - cole response .
part b ) represents some of the curves in part a ) on semi - logarithmic plot revealing initial exponential decay for large @xmath41 .
+ , title="fig:",width=283 ] ( in units of @xmath135 ) on circular frequency ( in units of @xmath136 ) for @xmath127 and @xmath137 .
, width=283 ] furthermore , real and imaginary parts of the response function are @xmath138&&=\chi_0\frac{(\omega\tau_r)^{1+\alpha } \sin(\frac{\alpha \pi}{2 } ) + ( \omega\tau_r)^{2\alpha}}{a(\omega ) } , \end{aligned}\ ] ] and @xmath139&&=\frac{1}{2k_bt}\omega s(\omega)\label{fdtf}\\ & & = \chi_0\frac{(\omega\tau_r)^{1+\alpha } \cos(\frac{\alpha \pi}{2 } ) + \frac{\langle \tau \rangle}{\tau_r}(\omega\tau_r)^{2\alpha+1}}{a(\omega ) } , \end{aligned}\ ] ] correspondingly .
notice that eq .
( [ fdtf ] ) is nothing else the classical fdt in the frequency domain .
this is an exact relation justifying the given name fdt , since @xmath140 $ ] is related to dissipation losses .
it can not be violated at thermal equilibrium for classical dynamics .
the absolute value of the response function measuring the output - to - input ratio of periodic signal amplitude is @xmath141 and the phase lag given by @xmath142 for slow signals with frequencies @xmath143 , @xmath144 .
this corresponds to quasi - static response .
the response to such slow periodic signals is present , as this result shows .
however , this does not mean that it can be detected .
the signal - to - noise ratio ( snr ) , @xmath145 which measures the ratio of spectral amplitude of signal to the spectral power of noise at the same frequency is important to determine whether the signal can in principle be detected in the noise background , or not @xcite . for the discussed model , @xmath146 where @xmath147 . for small frequencies snr
is strongly suppressed since then @xmath148 .
similar remarkable feature has been detected in the theory of non - markovian stochastic resonance @xcite .
for sufficiently large frequencies , snr becomes frequency - independent and attains the maximal value @xmath149 .
clearly , for @xmath69 , @xmath150 , and even slow signals can not be detected in the noise background , cf .
[ fig4 ] . besides
, the response itself to all signals with frequency @xmath151 becomes strongly suppressed .
stationary response is thus virtually absent in all systems at thermal equilibrium in the limit @xmath69 .
( in units of @xmath152 ) and signal - to - noise ratio @xmath153 ( in units of @xmath154 ) on circular frequency ( in units of @xmath136 ) for @xmath121 and different values of @xmath41 .
notice that both the frequency domain of quasi - static response and the maximal value of @xmath153 shrink as @xmath155 with increase of @xmath41 ( in units of @xmath32 ) .
, width=283 ] nonstationary response exists , however , even in non - ergodic systems @xcite . to describe nonstationary response also in ergodic systems driven initially out of equilibrium and left to relax or age to the new equilibrium ( e.g. after a temperature jump ) , one needs to find non - stationary or aging correlation function of the unperturbed system @xmath156 which depends on the system age @xmath157 . for the considered ctrw model
, it can also be found straightforwardly following to the derivation given in appendix .
one can repeat it with another first - time survival probability @xmath158 which depends on the age of system @xmath157 instead of @xmath51 .
this yields @xmath159 where averaging @xmath160 is done over initial distribution @xmath161 which generally is different from the stationary one , and @xmath162 .
the major problem is reduced to finding @xmath158 .
the central role is played indeed by the first time survival probability , as already well - established @xcite for symmetric two state process of the kind considered .
it features also our fully decoupled ctrw model with arbitrary number of states .
aging @xmath158 can be found from the exact relation similar to one discussed for forward recurrence - time in refs .
@xcite , @xmath163 where @xmath164 is the probability density to have @xmath165 scattering events .
it is the @xmath165-time convolution of the density @xmath40 , @xmath166^n$ ] in the laplace space .
indeed , let assume that our system was newly prepared at the time @xmath167 in the past relative to the starting point of observations @xmath12 .
then , the first term ( [ f1 ] ) is just the probability to stay from @xmath167 until @xmath15 without any scattering event .
however , @xmath165 such intermittent events can occur until any `` unseen '' time point @xmath168 within the time interval @xmath169 $ ] in the past and then no events occur until @xmath15 . integrating over @xmath170 and summing over all possible @xmath165 yields the above exact result . from it
, one can easily find the double laplace transform @xmath171 , where @xmath172 is laplace - conjugated to @xmath15 and @xmath173 to @xmath157 . some algebra yields simple result @xmath174 the laplace - transform of the fully aged @xmath175 , the first - time stationary survival probability ,
can be obtained now as @xmath176 . for @xmath177
this reproduces the well - known result in eq .
( [ rel1 ] ) and provides a very important consistency check . and for @xmath178 , @xmath179 , or @xmath180 .
this entails again the death of stationary response . in this case ,
nonstationary response to a periodic signal is dying to zero asymptotically , as found for a two - state nonergodic dynamics in ref
ergodic systems with finite @xmath41 will also exhibit aging response being prepared in a non - equilibrium state at some @xmath181 in the past , e.g. after a temperature jump .
their response but approaches non - zero stationary limiting value featuring new equilibrium state , which has been considered in this work .
this is what normally seen in most aging experiments : response to a periodic signal gradually dies out and approaches a stationary non - zero limit @xcite .
nonstationary aging response within the considered model will be studied in detail elsewhere . as an intelligent guess based on multi - state markovian dynamics @xcite in the absence of stationary fluxes in unperturbed dynamics ,
the two - time inhomogeneous response function should read @xmath182\;,\end{aligned}\ ] ] with @xmath183 , @xmath5 being the temperature after the temperature quench and @xmath184 with external field starting to act at @xmath185 .
as to the survival probability @xmath186 incorrectly used instead of @xmath51 in most ctrw theories of stationary response in thermally equilibrium environments , it corresponds to @xmath158 at zero age @xmath187 , @xmath188 , i.e. to the response of systems mostly deviating from thermal equilibrium , and not mostly close to it . clearly , it does not describe even nonstationary zero - age response in aging systems .
in this work , we showed within a multistate renewal model that nonergodic systems featured by infinite mean waiting times can not respond asymptotically to stationary signals .
this questions a good part of the anomalous response theory based on such processes . a common fallacy in the corresponding literature
consists in failure to find the correct stationary autocorrelation function which can be used as a phenomenological input in the fundamental microscopical theory of stationary linear response by kubo and others . in this respect ,
our critique is similar to one by tunaley @xcite earlier . more important
, we studied the way how the stationary response dies out with increasing mean waiting times in a model with one normal and one anomalous relaxation channel acting concurrently and in parallel .
the normal channel defines the mean waiting time .
this model reproduces approximately the cole - cole response in the limit where mean time is much less than the anomalous relaxation time defining the time of cole - cole response ( about inverse frequency at the maximum of absorption line defined by @xmath189\propto \omega s(\omega)$ ] ) .
paradoxically enough , such a response is more anomalous , with index @xmath79 , for less anomalous relaxation channels with index @xmath115 closer to one .
then , the resulting stochastic process becomes closer to @xmath0 noise .
this provides an elegant way to explain the origin of @xmath0 noise within ergodic dynamics featured by stationary response .
absence of stationary response for nonergodic dynamics with infinite mean waiting times does not mean of course that the response is totally absent .
it will be dying down to zero in the course of time as clarified earlier for two - state nonergodic dynamics @xcite .
however , it can be of primary importance for reacting on non - stationary signals only transiently present a common situation in many biological applications @xcite , or for other complex nonergodic input signals @xcite .
for example , response of neuronal systems to boring constant step signals should normally be damped out ( a healthy reaction ) , and this is indeed the case as shown in several experiments @xcite .
this is a sign of complexity @xcite .
a theory of such dying nonstationary response within ctrw approach has been initiated in refs .
@xcite and we refer interesting readers to those works .
they laid grounds for a continuing scientific exploration of the response of non - ergodic , non - stationary and aging systems which encompass also this author and readers .
this fascinating scientific exploration is now at the very beginning .
i am confident that it will attract ever more attention in the future .
the hospitality of kavli institute for theoretical physics at the chinese academy of sciences in beijing , where this work was partially done , is gratefully acknowledged .
stationary function @xmath190 can be found as @xmath191 , where @xmath192 is two - time stationary probability density of the process @xmath4 ( joint probability ) .
it depends only on the time difference of arguments and can be expressed through the corresponding stationary conditional probability density or propagator @xmath193 as @xmath194 , where @xmath195 is stationary single - time probability density . how to construct stationary propagator of such and similar ctrw processes
is explained in detail in refs .
@xcite . along similar lines
, we introduce the matrix of waiting time densities @xmath196 which in the present case is expressed through the only one density @xmath40 . in components , @xmath197 , where @xmath198 is kronecker symbol ( unity matrix @xmath199 ) .
next , we introduce the transition matrix @xmath200 with component @xmath201 being the transition probability from the state @xmath44 to the state @xmath45 as a result of scattering event . matrix of survival probabilities is @xmath202 .
moreover , we need the matrices of the first time densities , @xmath203 and the first time survival probabilities , @xmath204 , @xmath205 .
stationary propagator is is easy to find in the laplace - space , @xmath206 ( tilde denotes the corresponding laplace - transform for any function ) .
then , the conditional state probabilities to remain in the initial state are captured by @xmath207 .
contribution of the path with one scattering event is @xmath208 .
two scattering events contribute as @xmath209 , and so on .
all in all , @xmath210^n \mathbf t \tilde { \mathbf \psi}^{(0)}(s ) \\ & & = \tilde { \mathbf \phi}^{(0)}(s ) + \tilde { \mathbf \phi}(s ) \frac{1}{\mathbf i-\mathbf t
\tilde { \mathbf \psi}(s ) } \mathbf t \tilde { \mathbf \psi}^{(0)}(s ) . \nonumber\end{aligned}\ ] ] this expression is quite general and valid for other models of @xmath211 , and @xmath212 .
further calculations are straightforward for the considered model , @xmath46 , @xmath213 .
the first term in ( [ first ] ) contributes as @xmath214 to the correlation function .
the series also can be summed exactly by using that @xmath215 , or @xmath216 , and @xmath217 repeatedly , and standard relations between the densities and survival probabilities like @xmath218/s$ ] .
the series term contributes as @xmath219 , and after some algebra we obtain the simple exact result @xmath220 with @xmath221 .
this yields covariance ( [ acf ] ) upon identifying stationary averages with thermal equilibrium averages .
this result is valid for any number of states within the fully decoupled ctrw model .
the simplest stationary process of this kind is two - state semi - markov process considered by geisel _
@xcite for velocity variable , as a statistical model for chaotic dynamics .
it must be stressed that @xmath40 is not the time probability density to stay in the corresponding state because at each scattering event the particle can remain in the same state .
this is waiting time distribution between scattering events .
this process differs from the alternating symmetric two - state process considered by stratonovich , _ et al .
the last one alternates at each scattering event with the probability one , i.e. @xmath222 , @xmath223 .
in such a case , @xmath224 denoted for this model as @xmath225 is indeed the time density to reside in each state .
autocorrelation function of such a process looks less elegant : @xmath226 this result can also be easily derived using the corresponding transition matrix @xmath200
. it can not be immediately inverted to the time domain .
however , the relation between these two different two - state semi - markovian models is in fact simple : @xmath227 $ ] with @xmath228 . | linear theory of stationary response in systems at thermal equilibrium requires to find equilibrium correlation function of unperturbed responding system .
studies of the response of the systems exhibiting anomalously slow dynamics are often based on the continuous time random walk description ( ctrw ) with divergent mean waiting times .
the bulk of the literature on anomalous response contains linear response functions like one by cole - cole calculated from such a ctrw theory and applied to systems at thermal equilibrium .
here we show within a fairly simple and general model that for the systems with divergent mean waiting times the stationary response at thermal equilibrium is absent , in accordance with some recent studies .
the absence of such stationary response ( or dying to zero non - stationary response in aging experiments ) would confirm ctrw with divergent mean waiting times as underlying physical relaxation mechanism , but reject it otherwise .
we show that the absence of stationary response is closely related to the breaking of ergodicity of the corresponding dynamical variable . as an important new result , we derive a generalized cole - cole response within ergodic ctrw dynamics with finite waiting time .
moreover , we provide a physically reasonable explanation of the origin and wide presence of @xmath0 noise in condensed matter for ergodic dynamics close to normal , rather than strongly deviating . |
-5 mm converse theorems traditionally have provided a way to characterize dirichlet series associated to modular forms in terms of their analytic properties . most familiar are the converse theorems of hecke and weil .
hecke first proved that @xmath1-functions associated to modular forms enjoyed `` nice '' analytic properties and then proved `` conversely '' that these analytic properties in fact characterized modular @xmath1-functions .
weil extended this converse theorem to @xmath1-functions of modular forms with level . in their modern formulation ,
converse theorems are stated in terms of automorphic representations of @xmath2 instead of modular forms .
jacquet , piatetski - shapiro , and shalika have proved that the @xmath1-functions associated to automorphic representations of @xmath2 have nice analytic properties via integral representations similar to those of hecke .
the relevant `` nice '' properties are : analytic continuation , boundedness in vertical strips , and functional equation .
converse theorems in this context invert these integral representations .
they give a criterion for an irreducible admissible representation @xmath3 of @xmath2 to be automorphic and cuspidal in terms of the analytic properties of rankin - selberg convolution @xmath1-functions @xmath4 of @xmath3 twisted by cuspidal representations @xmath5 of @xmath6 of smaller rank groups .
to use converse theorems for applications , proving that certain objects are automorphic , one must be able to show that certain @xmath1-functions are `` nice '' . however , essentially the only way to show that an @xmath1-function is nice is to have it associated to an automorphic form .
hence the most natural applications of converse theorems are to functoriality , or the lifting of automorphic forms , to @xmath0 .
more explicit number theoretic applications then come as consequences of these liftings .
recently there have been several applications of converse theorems to establishing functorialities .
these have been possible thanks to the recent advances in the langlands - shahidi method of analysing the analytic properties of general automorphic @xmath1-functions , due to shahidi and his collaborators @xcite . by combining our converse theorems with their control of the analytic properties of @xmath1-functions many new examples of functorial liftings to @xmath0
have been established .
these are described in section 4 below .
as one number theoretic consequence of these liftings kim and shahidi have been able to establish the best general estimates over a number field towards the ramamujan - selberg conjectures for @xmath7 , which in turn have already had other applications .
-5 mm let @xmath8 be a global field , @xmath9 its adele ring , and @xmath10 a fixed non - trivial ( continuous ) additive character of @xmath9 which is trivial on @xmath8
. we will take @xmath11 to be an integer . to state these converse theorems , we begin with an irreducible admissible representation @xmath3 of @xmath12 .
it has a decomposition @xmath13 , where @xmath14 is an irreducible admissible representation of @xmath15 . by the local theory of jacquet , piatetski - shapiro , and shalika @xcite to each @xmath14
is associated a local @xmath1-function @xmath16 and a local @xmath17-factor @xmath18 .
hence formally we can form @xmath19 we will always assume the following two things about @xmath3 : 1 .
@xmath20 converges in some half plane @xmath21 , 2 .
the central character @xmath22 of @xmath3 is automorphic , that is , invariant under @xmath23 . under these assumptions
, @xmath24 is independent of our choice of @xmath10 @xcite . as in weil
s case , our converse theorems will involve twists but now by cuspidal automorphic representations of @xmath25 for certain @xmath26 . for convenience ,
let us set @xmath27 to be the set of automorphic representations of @xmath25 , @xmath28 the set of ( irreducible ) cuspidal automorphic representations of @xmath25 , and @xmath29 .
if @xmath30 is a finite set of places , we will let @xmath31 denote the subset of representations @xmath32 with local components
@xmath33 unramified at all places @xmath34 and let @xmath35 denote those @xmath36 which are unramified for all @xmath37 .
let @xmath38 be a cuspidal representation of @xmath25 with @xmath39
. then again we can formally define @xmath40 since the local factors make sense whether @xmath3 is automorphic or not .
a consequence of ( 1 ) and ( 2 ) above and the cuspidality of @xmath41 is that both @xmath42 and @xmath43 converge absolutely for @xmath21 , where @xmath44 and @xmath45 are the contragredient representations , and that @xmath46 is independent of the choice of @xmath10 .
we say that @xmath42 is _ nice _ if it satisfies the same analytic properties it would if @xmath3 were cuspidal , i.e. , 1 .
@xmath42 and @xmath43 have continuations to _ entire _ functions of @xmath47 , 2 .
these entire continuations are _ bounded in vertical strips _ of finite width , 3 .
they satisfy the standard _ functional equation _ @xmath48 the basic converse theorem for @xmath49 is the following . * theorem 1 .
* @xcite _ let @xmath3 be an irreducible admissible representation of @xmath12 as above .
let @xmath30 be a finite set of finite places .
suppose that @xmath42 is nice for all @xmath50 .
then @xmath3 is quasi - automorphic in the sense that there is an automorphic representation @xmath51 such that @xmath52 for all @xmath37 .
if @xmath30 is empty , then in fact @xmath3 is a cuspidal automorphic representation of @xmath2 .
_ it is this version of the converse theorem that has been used in conjunction with the langlands - shahidi method of controlling analytic properties of @xmath1-functions in the new examples of functoriality explained below . * theorem 2 .
* @xcite _ let @xmath3 be an irreducible admissible representation of @xmath12 as above .
let @xmath30 be a non - empty finite set of places , containing @xmath53 , such that the class number of the ring @xmath54 of @xmath30-integers is one .
suppose that @xmath42 is nice for all @xmath55 .
then @xmath3 is quasi - automorphic in the sense that there is an automorphic representation @xmath51 such that @xmath52 for all @xmath34 and all @xmath37 such that both @xmath14 and @xmath56 are unramified . _
this version of the converse theorem was specifically designed to investigate functoriality in the cases where one controls the @xmath1-functions by means of integral representations where it is expected to be more difficult to control twists .
the proof of theorem 1 with @xmath30 empty and @xmath57 replaced by @xmath58 essentially follows the lead of hecke , weil , and jacquet - langlands .
it is based on the integral representations of @xmath1-functions , fourier expansions , mellin inversion , and finally a use of the weak form of langlands spectral theory . for theorems 1 and 2 where we have restricted our twists either by ramification or rank
we must impose certain local conditions to compensate for our limited twists . for theorem 1 are a finite number of local conditions and for theorem 2 an infinite number of local conditions .
we must then work around these by using results on generation of congruence subgroups and either weak approximation ( theorem 1 ) or strong approximation ( theorem 2 ) .
as for our expectations of what form the converse theorem may take in the future , we refer the reader to the last section of @xcite .
-5 mm in order to apply these theorems , one must be able to control the analytic properties of the @xmath1-function .
however the only way we have of controlling global @xmath1-functions is to associate them to automorphic forms or representations .
a minute s thought will then convince one that the primary application of these results will be to the lifting of automorphic representations from some group @xmath59 to @xmath0 .
suppose that @xmath59 is a reductive group over @xmath8 . for simplicity of exposition we will assume throughout that @xmath59
is split and deal only with the connected component of its @xmath1-group , which we will ( by abuse of notation ) denote by @xmath60 @xcite .
let @xmath61 be a cuspidal automorphic representation of @xmath59 and @xmath62 a complex representation of @xmath60 .
to this situation langlands has associated an @xmath1-function @xmath63 @xcite .
let us assume that @xmath62 maps @xmath60 to @xmath64 .
then by langlands general principle of functoriality to @xmath36 should be associated an automorphic representation @xmath3 of @xmath2 satisfying @xmath65 , @xmath66 , with similar equalities locally and for the twisted versions @xcite . using
the converse theorem to establish such liftings involves three steps : construction of a candidate lift , verification that the twisted @xmath1-functions are `` nice '' , and application of the appropriate converse theorem .
_ construction of a candidate lift _ : we construct a candidate lift @xmath13 on @xmath2 place by place .
we can see what @xmath14 should be at almost all places . since we have the arithmetic langlands ( or hecke - frobenius ) parameterization of representations @xmath33 of @xmath67 for all archimedean places and those non - archimedean places where the representations are unramified @xcite , we can use these to associate to @xmath33 and the map @xmath68 a representation @xmath14 of @xmath69 .
this correspondence preserves local @xmath1- and @xmath17-factors @xmath70 along with the twisted versions .
if @xmath59 happens to be @xmath71 or a related group then we in principle know how to associate the representation @xmath14 at all places now that the local langlands conjecture has been solved for @xmath71 .
for other situations , we may not know what @xmath14 should be at the ramified places .
we will return to this difficulty momentarily and show how one can work around this with the use of a highly ramified twist .
but for now , let us assume we can finesse this local problem and arrive at a global representation @xmath13 such that @xmath72 and similarly @xmath66 with similar equalities for the twisted versions .
@xmath3 should then be the langlands lifting of @xmath36 to @xmath2 associated to @xmath62 .
_ analytic properties of global @xmath1-functions _
: for simplicity of exposition , let us now assume that @xmath62 is simply a standard embedding of @xmath73 into @xmath64 , such as will be the case if we consider @xmath59 to be a split classical group , so that @xmath74 is the standard @xmath1-function of @xmath36 .
we have our candidate @xmath3 for the lift of @xmath36 to @xmath0 from above . to be able to assert that the @xmath3 which we constructed place by place is automorphic
, we will apply a converse theorem . to do
so we must control the twisted @xmath1-functions @xmath75 for @xmath76 with an appropriate twisting set @xmath77 from theorem 1 or 2 . in the examples presented below , we have used theorem 1 above and the analytic control of @xmath78 achieved by the so - called langlands - shahidi method of analyzing the @xmath1-functions through the fourier coefficients of eisenstein series @xcite .
currently this requires us to take @xmath8 to be a number field . the _ functional equation _
@xmath79 has been proved in wide generality by shahidi @xcite .
the _ boundedness in vertical strips _ has been proved in close to the same generality by gelbart and shahidi @xcite . as for
the entire continuation of @xmath78 , a moments thought will tell you that one should not always expect a cuspidal representation of @xmath80 to necessarily lift to a cuspidal representation of @xmath2 .
hence it is unreasonable to expect all @xmath78 to be entire .
we had previously understood how to work around this difficulty from the point of view of integral representations by again using a highly ramified twist .
kim realized that one could also control the entirety of these twisted @xmath1-functions in the context of the langlands - shahidi method by using a highly ramified twist .
we will return to this below .
thus in a fairly general context one has that @xmath78 is _ entire _ for @xmath5 in a suitably modified twisting set @xmath81 .
_ application of the converse theorem _ : once we have that @xmath78 is nice for a suitable twisting set @xmath82 then from the equalities @xmath83 we see that the @xmath4 are nice and then we can apply our converse theorems to conclude that @xmath3 is either cuspidal automorphic or at least that there is an automorphic @xmath51 such that @xmath84 at almost all places .
this then effects the ( possibly weak ) automorphic lift of @xmath36 to @xmath3 or @xmath51 .
_ highly ramified twists _ : as we have indicated above , there are both local and global problems that can be finessed by an appropriate use of a highly ramified twist .
this is based on the following simple observation .
* observation . *
_ let @xmath3 be as in theorem 1 or 2 .
suppose that @xmath85 is a fixed character of @xmath86 .
suppose that @xmath42 is nice for all @xmath87 , where @xmath77 is either of the twisting sets of theorem 1 or 2 .
then @xmath3 is quasi - automorphic as in those theorems . _
the only thing to observe is that if @xmath88 then @xmath89 so that applying the converse theorem for @xmath3 with twisting set @xmath90 is equivalent to applying the converse theorem for @xmath91 with the twisting set @xmath77 .
so , by either theorem 1 or 2 , whichever is appropriate , @xmath91 is quasi - automorphic and hence @xmath3 is as well .
if we now begin with @xmath36 automorphic on @xmath80 , we will take @xmath92 to be the set of finite places where @xmath33 is ramified . for applying theorem 1
we want @xmath93 and for theorem 2 we would want @xmath94 .
we will now take @xmath85 to be highly ramified at all places @xmath95 , so that at @xmath95 our twisting representations are all locally of the form ( unramified principal series)@xmath96(highly ramified character ) . in order to finesse the lack of knowledge of an appropriate local lift
, we need to know the following two local facts about the local theory of @xmath1-functions for @xmath59 .
* multiplicativity of @xmath97-factors . * _ if @xmath98 , with @xmath99 and irreducible admissible representation of @xmath100 , then we have @xmath101 _ * stability of @xmath97-factors . * _ if @xmath102 and @xmath103 are two irreducible admissible representations of @xmath67 with the same central character , then for every sufficiently highly ramified character @xmath104 of @xmath105 we have @xmath106 _ both of these facts are known for @xmath0 , the multiplicativity being found in @xcite and the stability in @xcite .
multiplicativity in a fairly wide generality useful for applications has been established by shahidi @xcite .
stability is in a more primitive state at the moment , but shahidi has begun to establish the necessary results in a general context in @xcite . to utilize these local results , what one now does is the following . at the places where
@xmath33 is ramified , choose @xmath14 to be arbitrary , except that it should have the same central character as @xmath33 .
this is both to guarantee that the central character of @xmath3 is the same as that of @xmath36 and hence automorphic and to guarantee that the stable forms of the @xmath97-factors for @xmath33 and @xmath14 agree .
now form @xmath13 .
choose our character @xmath85 so that at the places @xmath95 we have that the @xmath1- and @xmath97-factors for both @xmath107 and @xmath108 are in their stable form and agree .
we then twist by @xmath109 for this _ fixed _
character @xmath85 .
if @xmath110 , then for @xmath95 , @xmath111 is of the form @xmath112 from which one deduces a similar equality for the @xmath1- and @xmath17-factors . from this
it will then follow that globally we will have @xmath113 for all @xmath110 with similar equalities for the @xmath17-factors .
this then completes step 1 . to complete our use of the highly ramified twist
, we must return to the question of whether @xmath78 can be made entire . in analysing @xmath1-functions via the langlands - shahidi method ,
the poles of the @xmath1-function are controlled by those of an eisenstein series . in general , the inducing data for the eisenstein series must satisfy a type of self - contragredience for there to be poles .
the important observation of kim is that one can use a highly ramified twist to destroy this self - contragredience at one place , which suffices , and hence eliminate poles .
the precise condition will depend on the individual construction .
a more detailed explanation of this can be found in shahidi s article @xcite .
this completes step 2 above .
-5 mm now take @xmath8 to be a number field .
there has been much progress recently in utilizing the method described above to establish global liftings from split groups @xmath59 over @xmath8 to an appropriate @xmath0 . among them
are the following .
_ classical groups_. take @xmath59 to be a split classical group over @xmath8 , more specifically , the split form of either @xmath114 , @xmath115 , or @xmath116 .
the the @xmath1-groups @xmath60 are then @xmath117 , @xmath118 , or @xmath119 and there are natural embeddings into the general linear group @xmath120 , @xmath121 , or @xmath120 respectively .
associated to each there should be a lifting of admissible or automorphic representations from @xmath80 to the appropriate @xmath122 .
the first lifting that resulted from the combination of the converse theorem and the langlands - shahidi method of controlling automorphic @xmath1-functions was the weak lift for generic cuspidal representations from @xmath114 to @xmath123 over a number field @xmath8 obtained with kim and shahidi @xcite .
we can now extend this to the following result .
* theorem .
* @xcite _ let @xmath59 be a split classical group over @xmath8 as above and @xmath36 a globally generic cuspidal representation of @xmath80 .
then there exists an automorphic representation @xmath3 of @xmath122 for the appropriate @xmath124 such that @xmath14 is the local langlands lift of @xmath33 for all archimedean places @xmath125 and almost all non - archimedean places @xmath125 where @xmath33 is unramified .
_ in these examples the local langlands correspondence is not understood at the places @xmath125 where @xmath33 is ramified and so we must use the technique of multiplicativity and stability of the local @xmath97-factors as outlined in section 3 .
multiplicativity has been established in generality by shahidi @xcite and in our first paper @xcite we relied on the stability of @xmath97-factors for @xmath114 from @xcite .
recently shahidi has established an expression for his local coefficients as mellin transforms of bessel functions in some generality , and in particular in the cases at hand one can combine this with the results of @xcite to obtain the necessary stability in the other cases , leading to the extension of the lifting to the other split classical groups @xcite .
_ tensor products_. let @xmath126
. then @xmath127 .
then there is a natural simple tensor product map from @xmath128 to @xmath129 .
the associated functoriality from @xmath130 to @xmath131 is the _ tensor product lifting_. now the associated local lifting is understood in principle since the local langlands conjecture for @xmath49 has been solved .
the question of global functoriality has been recently solved in the cases of @xmath132 to @xmath133 by ramakrishnan @xcite and @xmath134 to @xmath135 by kim and shahidi @xcite .
* theorem .
* @xcite _ let @xmath136 be a cuspidal representation of @xmath137 and @xmath138 a cuspidal representation of @xmath137 ( respectively @xmath139 ) .
then there is an automorphic representation @xmath3 of @xmath140 ( respectively @xmath141 ) such that @xmath14 is the local tensor product lift of @xmath142 at all places @xmath125 .
_ in both cases the authors are able to characterize when the lift is cuspidal
. in the case of ramakrishnan @xcite @xmath143 with each @xmath144 cuspidal representation of @xmath137 and @xmath3 is to be an automorphic representation of @xmath140 . to apply
the converse theorem ramakrishnan needs to control the analytic properties of @xmath4 for @xmath5 cuspidal representations of @xmath145 and @xmath137 , that is , the rankin triple product @xmath1-functions @xmath146 .
this he was able to do using a combination of results on the integral representation for this @xmath1-function due to garrett , rallis and piatetski - shapiro , and ikeda and the work of shahidi on the langlands - shahidi method . in the case of kim and shahidi @xcite @xmath138
is a cuspidal representation of @xmath139 . since the lifted representation @xmath3 is to be an automorphic representation of @xmath141 , to apply the converse theorem they must control the analytic properties of @xmath146 where now @xmath5 must run over appropriate cuspidal representations of @xmath25 with @xmath147 .
the control of these triple products is an application of the langlands - shahidi method of analysing @xmath1-functions and involves coefficients of eisenstein series on @xmath148 , @xmath149 , and simply connected @xmath150 and @xmath151 @xcite .
we should note that even though the complete local lifting theory is understood , they still use a highly ramified twist to control the global properties of the @xmath1-functions involved .
they then show that their lifting is correct at all local places by using a base change argument .
_ symmetric powers_. now take @xmath152 , so @xmath153 . for each @xmath154
there is the natural symmetric @xmath155-th power map @xmath156 .
the associated functoriality is the _ symmetric power lifting _ from representations of @xmath7 to representations of @xmath157 .
once again the local symmetric powers liftings are understood in principle thanks to the solution of the local langlands conjecture for @xmath49 .
the global symmetric square lifting , so @xmath7 to @xmath158 , is an old theorem of gelbart and jacquet .
recently , kim and shahidi have shown the existence of the global symmetric cube lifting from @xmath7 to @xmath159 @xcite and then kim followed with the global symmetric fourth power lifting from @xmath7 to @xmath148 @xcite .
* theorem .
* @xcite _ let @xmath36 be a cuspidal automorphic representation of @xmath137 .
then there exists an automorphic representation @xmath3 of @xmath140 ( resp .
@xmath160 ) such that @xmath14 is the local symmetric cube ( resp .
symmetric fourth power ) lifting of @xmath33 .
_ in either case , kim and shahidi have been able to give a very interesting characterization of when the image is in fact cuspidal @xcite .
the original symmetric square lifting of gelbart and jacquet indeed used the converse theorem for @xmath158 . for kim and shahidi ,
the symmetric cube was deduced from the functorial @xmath161 tensor product lift above @xcite and did not require a new use of the converse theorem . for the symmetric fourth power lift ,
kim first used the converse theorem to establish the _ exterior square _ lift from @xmath159 to @xmath135 by the method outlined above and then combined this with the symmetric cube lift to deduce the symmetric fourth power lift @xcite .
-5 mm _ 1 .
classical groups _ : the applications so far of the lifting from classical groups to @xmath0 have been `` internal '' to the theory of automorphic forms . in the case of the lifting from @xmath114 to @xmath123 ,
once the weak lift is established , then the theory of ginzburg , rallis , and soudry @xcite allows one to show that this weak lift is indeed a strong lift in the sense that the local components @xmath14 at those @xmath34 are completely determined and to completely characterize the image locally and globally .
this will be true for the liftings from the other classical groups as well .
once one knows that these lifts are rigid , then one can begin to define and analyse the local lift for ramified representations by setting the lift of @xmath33 to be the @xmath14 determined by the global lift .
this is the content of the papers of jiang and soudry @xcite for the case of @xmath162 .
in essence they show that this local lift satisfies the relations on @xmath1-functions that one expects from functoriality and then deduce the _ local langlands conjecture for @xmath114 _ from that for @xmath123 . we refer to their papers for more detail and precise statements .
_ 2 . tensor product lifts _ : ramakrishnan s original motivation for establishing the tensor product lifting from @xmath163 to @xmath159 was to prove the multiplicity one conjecture for @xmath164 of langlands and labesse .
the primary application of the tensor product lifting from @xmath168 to @xmath135 of kim and shahidi was in the establishment of the symmetric cube lifting and through this the symmetric fourth power lifting , so the applications of the symmetric power liftings outlined below are applications of this lifting as well .
_ 3 . symmetric powers _ : it was early observed that the existence of the symmetric power liftings of @xmath7 to @xmath157 for all @xmath155 would imply the ramanujan - petersson and selberg conjectures for modular forms . every time a symmetric power lift is obtained we obtain better bounds towards ramanujan .
the result which follows from the symmetric third and fourth power lifts of kim and shahidi is the following . * theorem .
* @xcite _ let @xmath36 be a cuspidal representation of @xmath137 such that the symmetric cube lift of @xmath36 is again cuspidal .
let @xmath169 be the satake parameter for an unramified local component . then @xmath170 .
if in addition the fourth symmetric power lift is not cuspidal , the full ramanujan conjecture is valid . _ the corresponding statement at infinite places , i.e.
, the analogue of the selberg conjecture on the eigenvalues of mass forms , is also valid @xcite .
estimates towards ramanujan are a staple of improving any analytic number theoretic estimates obtained through spectral methods .
both the @xmath171 non - archimedean and @xmath171 archimedean estimate towards ramanujan above were applied in obtaining the precise form of the exponent in our recent result with sarnak breaking the convexity bound for twisted hilbert modular @xmath1-series in the conductor aspect , which in turn was the key ingredient in our work on hilbert s eleventh problem for ternary quadratic forms .
similar in spirit are the applications by kim and shahidi to the hyperbolic circle problem and to estimates on sums of shifted fourier coefficients @xcite .
* theorem . *
@xcite _ let @xmath36 be a cuspidal representation of @xmath137 with trivial central character .
let @xmath169 be the satake parameter for an unramified local component and let @xmath172 .
assuming @xmath36 satisfies the ramanujan conjecture , there are sets @xmath173 of positive lower density for which @xmath174 for all @xmath175 and @xmath176 for all @xmath177 .
[ note : @xmath178 _ kim and shahidi have other conditional applications of their liftings such as the conditional existence of siegel modular cusp forms of weight @xmath179 ( assuming arthur s multiplicity formula for @xmath180 ) .
we refer the reader to @xcite for details on these applications and others .
aa a. borel , automorphic @xmath1-functions .
pure math .
_ * 33 * , part 2 , ( 1979 ) , 2761 . j.w .
cogdell , h. kim , i.i .
piatetski - shapiro , and f. shahidi , on lifting from classical groups to @xmath49 .
ihes _ * 93 * ( 2001 ) , 530 . j.w .
cogdell , h. kim , i.i .
piatetski - shapiro , and f. shahidi , on lifting from classical groups to @xmath49 , ii , in preparation .
cogdell and i.i .
piatetski - shapiro , converse theorems for @xmath49 , i. _ publ .
ihes _ * 79 * ( 1994 ) , 157214
cogdell and i.i .
piatetski - shapiro , stability of gamma factors for @xmath114 .
_ manuscripta math .
_ * 95 * ( 1998 ) 437461 .
cogdell and i.i .
piatetski - shapiro , converse theorems for @xmath49 , ii . _
j. reine angew .
_ * 507 * ( 1999 ) , 165188 .
s. gelbart and f. shahidi , boundedness of automorphic @xmath1-functions in vertical strips . _ journal of the ams _ * 14 * ( 2001 ) , 79107 .
d. ginzburg , s. rallis , and d. soudry , generic automorphic forms on @xmath181 : functorial lift to @xmath182 , endoscopy , and base change .
_ internat .
res . notices .
* 14*(2001 ) , 729764 .
h. jacquet , i.i .
piatetski - shapiro , and j. shalika , rankin - selberg convolutions , _ amer . j. math .
_ , * 105 * ( 1983 ) , 367464 .
h. jacquet and j. shalika , a lemma on highly ramified @xmath17-factors , _ math .
_ , * 271 * ( 1985 ) , 319332 .
h. jacquet and j. shalika , rankin - selberg convolutions : archimedean theory , in _ festschrift in honor of i.i .
piatetski - shapiro _ , part i , weizmann science press , jerusalem , 1990 , 125207 .
d. jiang and d. soudry , the local converse theorem for @xmath181 and applications .
_ ann . of math
_ , to appear .
d. jiang and d. soudry , generic representations and local langlands reciprocity law for @xmath183-adic @xmath181 .
_ preprint _
h. kim , functoriality for the exterior square of @xmath159 and the symmetric fourth of @xmath7 .
_ preprint _ ( 2000 ) .
h. kim and f. shahidi , functorial products for @xmath184 and symmetric cube for @xmath185 .
_ ann . of math .
_ , to appear .
h. kim and f. shahidi , cuspidality of symmetric powers with applications .
_ duke math .
j. _ , * 112 * ( 2002 ) , to appear .
d. ramakrishnan , modularity of the rankin - selberg @xmath1-series , and multiplicity one for @xmath186 .
_ ann . of math .
( 2 ) _ * 152 * ( 2000 ) , 45111 .
f. shahidi , a proof of langlands conjecture on plancherel measures ; complementary series for @xmath183-adic groups .
_ ann . of math .
( 2 ) _ * 132 * ( 1990 ) , 273330 .
f. shahidi , on multiplicativity of local factors .
_ festschrift in honor of i. i. piatetski - shapiro _ , part ii ( ramat aviv , 1989 ) , 279289 .
israel math .
proc . , 3 , weizmann , jerusalem , 1990 .
f. shahidi , local coefficients as mellin transforms of bessel functions ; towards a general stability .
_ preprint _ ( 2002 ) .
f. shahidi , automorphic @xmath1-functions and functoriality .
these _ proceedings_. | there has been a recent coming together of the converse theorem for @xmath0 and the langlands - shahidi method of controlling the analytic properties of automorphic @xmath1-functions which has allowed us to establish a number of new cases of functoriality , or the lifting of automorphic forms . in this article
we would like to present the current state of the converse theorem and outline the method one uses to apply the converse theorem to obtain liftings .
we will then turn to an exposition of the new liftings and some of their applications .
4.5 mm * 2000 mathematics subject classification : * 11f70 , 22e55 . * keywords and phrases : * automorphic forms , @xmath1-functions , converse theorems , functoriality . |
the kinetics of many chemical reactions is influenced by the transport properties of the reactants that they involve @xcite .
in fact , schematically , any chemical reaction requires first that a given reactant a meets a second reactant b. this first reaction step can be rephrased as a search process involving a searcher a looking for a target b. in a very dilute regime , exemplified by biochemical reactions in cells @xcite which sometimes involve only a few copies of reactants , the targets b are sparse and therefore hard to find in this search process language .
in such reactions , the first step of search for reactants b is therefore a limiting factor of the global reaction kinetics . in the general aim of enhancing the reactivity of chemical systems ,
it is therefore needed to optimize the efficiency of this first step of search .
recently , it has been shown that intermittent processes , combining slow diffusion phases with a faster transport , can significantly increase reactions rates @xcite .
a minimal model demonstrating the efficiency of this type of search , introduced to account for the fast search of target sequences on dna by proteins @xcite is as follows ( see also @xcite ) .
the pathway followed by the protein , considered as a point - like particle , is a succession of 1d diffusions along the dna strand ( called sliding phases ) with diffusion coefficient @xmath0 and 3d excursions in the surrounding solution .
the time spent by the protein on dna during each sliding phase is assumed to follow an exponential law with dissociation rate @xmath1 . in this minimal model ,
the 3d excursions are uncorrelated in space , which means that after dissociation from dna , the protein will rebind the dna at a random position independently of its starting position . assuming further that the mean duration of such 3d excursions @xmath2 is finite , it has been shown that the mean first - passage time at the target can be minimized as a function of @xmath3 , as soon as the mean time spent in bulk excursions is not too long .
quantitatively , this condition writes in orders of magnitude as @xmath4 , and the minimum of the search time is obtained for @xmath5 in the large @xmath6 limit .
note that in this minimal model , where the time @xmath2 is supposed to be a fixed exterior parameter , bulk phases are always beneficial in the large @xmath6 limit ( i.e. allow one to decrease the search time with respect to the situation corresponding to 1d diffusion only ) . in many practical situations however , the duration of the fast bulk excursions strongly depends on the geometrical properties of the system @xcite and can not be treated as an independent variable as assumed in the mean - field ( mf ) model introduced above .
an important generic situation concerns the case of confined systems @xcite , involving transport of reactive molecules both in the bulk of a confining domain and on its boundary , referred to as surface - mediated diffusion in what follows .
this type of problems is met in situations as varied as heterogeneous catalysis @xcite , or reactions in porous media and in vesicular systems @xcite . in all these examples ,
the duration of bulk excursions is controlled by the return statistics of the molecule to the confining surface , which crucially depends on the volume of the system .
this naturally induces strong correlations between the starting and ending points of bulk excursions , and makes the above mf assumption of uncorrelated excursions largely inapplicable in these examples . at the theoretical level
, the question of determining mean first - passage times in confinement has attracted a lot of attention in recent years for discrete random walks @xcite and continuous processes @xcite .
more precisely , the surface - mediated diffusion problem considered here generalizes the so - called narrow escape problem , which refers to the time needed for a simple brownian motion in absence of surface diffusion to escape through a small window of an otherwise reflecting domain .
this problem has been investigated both in the mathematical @xcite and physical @xcite literature , partly due to the challenge of taking into account mixed boundary conditions .
the case of surface - mediated diffusion brings the additional question of minimizing the search time with respect to the time spent in adsorption , in the same spirit as done for intermittent processes introduced above . the answer to
this question is _ a priori _ not clear , since the mean time spent in bulk excursions diverges for large confining domains , so that the condition of minimization mentioned previously can not be taken as granted , even in the large system limit . in this context
, first results have been obtained in @xcite where , surprisingly enough , it has been found that , even for bulk and surface diffusion coefficients of the same order of magnitude , the reaction time can be minimized , whereas mf treatments ( see for instance @xcite ) predict a monotonic behavior . here
, we extend the perturbative results of @xcite obtained in the small target size limit . relying on an integral equation approach
, we provide an exact solution for the mean fpt , both for 2d an 3d spherical domains , and for any spherical target size .
we also develop approximation schemes , numerically validated , that provide more tractable expressions of the mean fpt .
the surface - mediated process under study is defined as follows .
we consider a molecule diffusing in a spherical confining domain of radius @xmath7 ( see figure [ fig1 ] ) , alternating phases of boundary diffusion ( with diffusion coefficient @xmath0 ) and phases of bulk diffusion ( with diffusion coefficient @xmath8 ) . the time spent during each one - dimensional phase
is assumed to follow an exponential law with dissociation rate @xmath1 . at each desorption event , the molecule is assumed to be ejected at a distance @xmath9 from the frontier ( otherwise it is instantaneously readsorbed ) . although formulated for any value of this parameter @xmath9 smaller than @xmath7 , in most physical situations of real interest @xmath10 .
the target is perfectly absorbing and defined in @xmath11 by the arc @xmath12 $ ] , and in @xmath13 by the region of the sphere such that @xmath14 $ ] where @xmath15 is in this case the elevation angle in standard spherical coordinates .
note that as soon as @xmath16 , the target can be reached either by surface or bulk diffusion . in what follows
we calculate the mean first - passage time at the target for an arbitrary initial condition of the molecule .
model , scaledwidth=45.0% ]
in this section , the confining domain is a disk of radius @xmath7 and the target is defined by the arc @xmath12 $ ] . for the process defined above
, the mean first - passage time ( mfpt ) at the target satisfies the following backward equations @xmath17 & = & -1 \;\;{\rm for}\;\theta\in[\epsilon,2\pi-\epsilon ] , \\
\label{eq : t2 } d_2\left ( \frac{\partial ^2}{\partial r^2 } + \frac{1}{r } \frac{\partial } { \partial r } + \frac{1}{r^2 } \frac{\partial ^2}{\partial \theta^2}\right ) t_2(r,\theta ) & = & -1 , \end{aligned}\ ] ] where @xmath18 stands for the the mfpt starting from the circle at a position defined on the circumference by the angle @xmath15 , and @xmath19 for the mfpt starting from the point @xmath20 within the disk . in these two equations , the first term of the lhs accounts for the diffusion respectively on the circumference and in the bulk , while the second term of eq .
( [ eq : t1 ] ) describes desorption events .
they have to be completed by two boundary conditions @xmath21\cup[2\pi-\epsilon,2\pi ] , \end{aligned}\ ] ] which describe the adsorption events and the absorbing target respectively .
eq.([eq : t2 ] ) is easily shown to be satisfied by the following fourier series @xmath22 with unknown coefficients @xmath23 to be determined . in particular , we aim at determining the search time @xmath24 , defined as the mfpt , with an initial position uniformly distributed on the boundary of the confining domain .
taking eq.([eq : t2integre ] ) at @xmath25 , we have @xmath26 $ , } \\ 0 & \text{if $ \theta\in[0,\epsilon]\cup[2\pi-\epsilon,2\pi]$ , } \end{cases}\ ] ] so that @xmath27 in what follows we will make use of the following quantities : @xmath28 @xmath29 and @xmath30 as we proceed to show , two different approaches can be used to solve this problem .
( i ) the first approach , whose main results have been published in @xcite , uses the explicit form of the green function for the two - dimensional problem and relies on a small target size @xmath31 expansion .
we recall these perturbative results below for the sake of self - consistency and give details of the derivation in appendix [ sec : approach ] .
( ii ) the second approach presented next relies on an integral equation which can be derived for @xmath18 , and leads to an exact non - perturbative solution .
it is shown in appendix that the fourier coefficients of @xmath32 as defined in eq.([eq : t2integre ] ) satisfy an infinite hierarchy of linear equations , which lead to the following small @xmath31 expansion : @xmath33 note that eq.([eq : alphanperturbation0 ] ) gives in particular the first terms of the perturbative expansion of the search time @xmath34 defined in ( [ eq : deffourier ] ) and given in @xcite . it should be stressed that since the coefficients of @xmath35 of this expansion diverge with @xmath36 , in practice one finds that the range of applicability in @xmath31 of this expansion is wider for @xmath36 small . in this section
, we first show that the resolution of the coupled pdes ( [ eq : t1 ] , [ eq : t2 ] ) amounts to solving an integral equation for @xmath18 only . as we proceed to show , this integral equation can be solved exactly .
writing eq .
( [ eq : t1 ] ) as @xmath37,\ ] ] and expanding its right - hand side into a taylor series leads to @xmath38 substituting the fourier representation ( [ eq : t2integre ] ) for @xmath19 into this equation yields @xmath39 changing the order of summations over @xmath40 and @xmath41 , using the binomial formula and the expression ( [ eq : deffourier ] ) for @xmath23 give @xmath42 this integro - differential equation for @xmath18 can actually easily be transformed into an integral equation for @xmath18 , by integrating successively two times , which leads to @xmath43 or equivalently to @xmath44 where @xmath45 with @xmath46 defined in eq .
( [ deft ] ) and @xmath47 .
note that eq .
( [ eq : inteq_2d ] ) holds for @xmath48 $ ] . when there is no desorption ( i.e. , @xmath49 ) , only the first term in eq .
( [ eq : psi_2d ] ) survives , yielding the classical result @xcite @xmath50 the same result is obtained for @xmath51 , since @xmath52 .
the limit @xmath51 is in fact equivalent to the limit @xmath53 because , after desorption , the particle immediately returns onto the circle ( @xmath51 ) as if it was never desorbed ( @xmath54 ) . iterating the integral equation ( [ eq : inteq_2d ] )
shows that the solution @xmath55 writes for @xmath56 $ ] : @xmath57 , \ ] ] with the coefficients @xmath58 which satisfy @xmath59 = \omega \sum\limits_{n=1}^\infty \biggl(u_n + \sum\limits_{n'=1}^\infty q_{n , n ' } d_{n'}\biggr)\bigl[\cos(n\theta ) - \cos(n\epsilon)\bigr ] , \\ \end{split}\ ] ] where we introduced @xmath60 and @xmath61 with @xmath62 since eq .
( [ eq : subs_2d ] ) should be satisfied for any @xmath48 $ ] , one gets @xmath63 , from which @xmath64_n \hskip 5 mm ( n = 1,2, ... ).\ ] ] since @xmath65 eq .
( [ eq : psi_2d ] ) with the @xmath58 given by eq .
( [ eq : dn_2d ] ) can be seen as a series in powers of @xmath66 , whose @xmath40-th order coefficient is explicitly written in terms of the @xmath40-th power of the matrix @xmath67 .
once the @xmath58 determined , the search time @xmath24 is @xmath74 while the previous expression of @xmath18 is exact , it is not fully explicit , since it requires either the inversion of the matrix @xmath75 or the calculation of all the powers of @xmath67 .
we give here an approximation of @xmath76 , which in turn provides a convenient and fully explicit representation of @xmath18 .
as shown numerically ( see figs .
[ 2dtest1 ] , [ 2dtest2 ] , [ 2dtest3 ] and section [ numerics ] for more details about numerical methods ) , this approximation of @xmath18 proves to be in quantitative agreement with the exact expression for a wide range of parameters .
this approximation relies on the fact that , in the small target size limit @xmath77 , the matrix @xmath67 is diagonal , which mirrors the orthogonality of the @xmath78 on @xmath79 $ ] . more precisely , one has from eqs .
( [ eq : q_2d],[ieps ] ) : @xmath80 and keeping only the leading term of this expansion yields @xmath81 from which we obtain the desired approximation : @xmath82 this yields an approximation for the search time : @xmath83 in 2d : the exact solution ( [ eq : psi_2d ] , [ eq : dn_2d ] ) , the approximation ( [ eq : psi_part_2d ] ) and the perturbative formula ( [ eq : alphanperturbation ] ) , with @xmath84 , @xmath85 . in the * first row * ,
the other parameters are : @xmath86 , @xmath87 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath89 , and the series are truncated to @xmath90 for the exact and approximate solutions , and to @xmath91 for the perturbative solution .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is below @xmath92 .
in the * third row * , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is evidently not applicable . in the * last row * , the parameters are : @xmath93 , @xmath89 , and the series are truncated to @xmath88 . in this case , the approximate solution significantly deviates from the exact one ( providing mostly negative values ) .
the perturbative solution is completely invalid ( not shown ) .
, title="fig:",width=302 ] in 2d : the exact solution ( [ eq : psi_2d ] , [ eq : dn_2d ] ) , the approximation ( [ eq : psi_part_2d ] ) and the perturbative formula ( [ eq : alphanperturbation ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath87 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath89 , and the series are truncated to @xmath90 for the exact and approximate solutions , and to @xmath91 for the perturbative solution .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is below @xmath92 .
in the * third row * , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is evidently not applicable . in the * last row
* , the parameters are : @xmath93 , @xmath89 , and the series are truncated to @xmath88 . in this case , the approximate solution significantly deviates from the exact one ( providing mostly negative values ) .
the perturbative solution is completely invalid ( not shown ) . , title="fig:",width=302 ] in 2d : the exact solution ( [ eq : psi_2d ] , [ eq : dn_2d ] ) , the approximation ( [ eq : psi_part_2d ] ) and the perturbative formula ( [ eq : alphanperturbation ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath87 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath89 , and the series are truncated to @xmath90 for the exact and approximate solutions , and to @xmath91 for the perturbative solution .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is below @xmath92 . in the * third row * ,
the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is evidently not applicable . in the * last row
* , the parameters are : @xmath93 , @xmath89 , and the series are truncated to @xmath88 . in this case
, the approximate solution significantly deviates from the exact one ( providing mostly negative values ) .
the perturbative solution is completely invalid ( not shown ) .
, title="fig:",width=302 ] in 2d : the exact solution ( [ eq : psi_2d ] , [ eq : dn_2d ] ) , the approximation ( [ eq : psi_part_2d ] ) and the perturbative formula ( [ eq : alphanperturbation ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath87 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath89 , and the series are truncated to @xmath90 for the exact and approximate solutions , and to @xmath91 for the perturbative solution .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is below @xmath92 .
in the * third row * , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is evidently not applicable . in the * last row
* , the parameters are : @xmath93 , @xmath89 , and the series are truncated to @xmath88 . in this case , the approximate solution significantly deviates from the exact one ( providing mostly negative values ) .
the perturbative solution is completely invalid ( not shown ) .
, title="fig:",width=302 ] in 2d : the exact solution ( [ eq : psi_2d ] , [ eq : dn_2d ] ) , the approximation ( [ eq : psi_part_2d ] ) and the perturbative formula ( [ eq : alphanperturbation ] ) , with @xmath84 , @xmath85 . in the * first row * ,
the other parameters are : @xmath86 , @xmath87 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath89 , and the series are truncated to @xmath90 for the exact and approximate solutions , and to @xmath91 for the perturbative solution .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is below @xmath92 . in the * third row * ,
the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is evidently not applicable . in the * last row
* , the parameters are : @xmath93 , @xmath89 , and the series are truncated to @xmath88 . in this case , the approximate solution significantly deviates from the exact one ( providing mostly negative values ) .
the perturbative solution is completely invalid ( not shown ) . , title="fig:",width=302 ] in 2d : the exact solution ( [ eq : psi_2d ] , [ eq : dn_2d ] ) , the approximation ( [ eq : psi_part_2d ] ) and the perturbative formula ( [ eq : alphanperturbation ] ) , with @xmath84 , @xmath85 . in the * first row * ,
the other parameters are : @xmath86 , @xmath87 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath89 , and the series are truncated to @xmath90 for the exact and approximate solutions , and to @xmath91 for the perturbative solution .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is below @xmath92 . in the * third row * ,
the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is evidently not applicable . in the * last row
* , the parameters are : @xmath93 , @xmath89 , and the series are truncated to @xmath88 . in this case , the approximate solution significantly deviates from the exact one ( providing mostly negative values ) .
the perturbative solution is completely invalid ( not shown ) .
, title="fig:",width=302 ] in 2d : the exact solution ( [ eq : psi_2d ] , [ eq : dn_2d ] ) , the approximation ( [ eq : psi_part_2d ] ) and the perturbative formula ( [ eq : alphanperturbation ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath87 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath89 , and the series are truncated to @xmath90 for the exact and approximate solutions , and to @xmath91 for the perturbative solution .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is below @xmath92 . in the * third row * ,
the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is evidently not applicable . in the * last row
* , the parameters are : @xmath93 , @xmath89 , and the series are truncated to @xmath88 . in this case , the approximate solution significantly deviates from the exact one ( providing mostly negative values ) .
the perturbative solution is completely invalid ( not shown ) .
, title="fig:",width=302 ] computed through eq .
( [ eq : dn_2d ] , [ eq : t1mean_2d ] ) with @xmath88 as a function of the desorption rate @xmath1 for three values of @xmath8 : @xmath95 ( dot - dashed blue line ) , @xmath84 ( dashed green line ) , and @xmath96 ( solid red line ) .
the other parameters are : @xmath97 and @xmath98 . when @xmath99 ( the first case ) , @xmath100 monotonously increases with @xmath1 so that there is no optimal value . in two other cases ,
@xmath101 , and @xmath24 starts first to decrease with @xmath1 , passes through a minimum ( the optimal value ) and monotonously increases .
symbols show the approximate mean time computed through eq .
( [ eq : betaa_2d ] , [ eq : t1mean_2d ] ) .
one can see that the approximation accurate enough even for large values of @xmath1 .
* right * : the derivative @xmath102 defined by eq .
( [ eq : t1deriv_2d ] ) for the same parameters . , title="fig:",width=302 ] computed through eq .
( [ eq : dn_2d ] , [ eq : t1mean_2d ] ) with @xmath88 as a function of the desorption rate @xmath1 for three values of @xmath8 : @xmath95 ( dot - dashed blue line ) , @xmath84 ( dashed green line ) , and @xmath96 ( solid red line ) .
the other parameters are : @xmath97 and @xmath98 . when @xmath99 ( the first case ) , @xmath100 monotonously increases with @xmath1 so that there is no optimal value . in two other cases ,
@xmath101 , and @xmath24 starts first to decrease with @xmath1 , passes through a minimum ( the optimal value ) and monotonously increases .
symbols show the approximate mean time computed through eq .
( [ eq : betaa_2d ] , [ eq : t1mean_2d ] ) .
one can see that the approximation accurate enough even for large values of @xmath1 .
* right * : the derivative @xmath102 defined by eq .
( [ eq : t1deriv_2d ] ) for the same parameters .
, title="fig:",width=302 ] as a function of @xmath31 , with @xmath84 , @xmath85 , and @xmath94 ( left ) or @xmath89 ( right ) .
the exact computation through eq .
( [ eq : dn_2d ] , [ eq : t1mean_2d ] ) is compared to the approximation ( [ eq : betaa_2d ] , [ eq : t1mean_2d ] ) and to the perturbative approach . in all cases ,
the series are truncated to @xmath103 . for small @xmath1 ( @xmath94 ) ,
the approximate solution is very close to the exact one , while the perturbative solution is relatively close for @xmath31 up to @xmath104 . in turn , for large @xmath1 ( @xmath89 ) , the approximate solution shows significant deviations for the intermediate values of @xmath31 , while the perturbative solution is not applicable at all .
, title="fig:",width=302 ] as a function of @xmath31 , with @xmath84 , @xmath85 , and @xmath94 ( left ) or @xmath89 ( right ) .
the exact computation through eq .
( [ eq : dn_2d ] , [ eq : t1mean_2d ] ) is compared to the approximation ( [ eq : betaa_2d ] , [ eq : t1mean_2d ] ) and to the perturbative approach . in all cases ,
the series are truncated to @xmath103 . for small @xmath1 ( @xmath94 ) ,
the approximate solution is very close to the exact one , while the perturbative solution is relatively close for @xmath31 up to @xmath104 . in turn , for large @xmath1 ( @xmath89 ) , the approximate solution shows significant deviations for the intermediate values of @xmath31 , while the perturbative solution is not applicable at all .
, title="fig:",width=302 ] in this section , we answer two important questions .
when are bulk excursions favorable , meaning enabling to reduce the search time ( with respect to the situation with no bulk excursion corresponding to @xmath54 ) ?
if so , is there an optimal value of the desorption rate @xmath1 minimizing the search time ?
this question can be investigated by studying the sign of the derivative @xmath105 at @xmath54 .
the mean search time from eq .
( [ eq : t1mean_2d ] ) can also be written as @xmath106,\ ] ] where @xmath107 , @xmath108 .
the derivative of @xmath24 with respect to @xmath1 is then @xmath109 .\ ] ] if the derivative is negative at @xmath49 , i.e. @xmath110 bulk excursions are beneficial to the search .
this inequality determines the critical value for the bulk diffusion coefficient @xmath111 ( which enters through @xmath112 ) , above which bulk excursions are beneficial : @xmath113 ^ 2.\ ] ] two comments are in order : \(i ) interestingly , this ratio depends only on @xmath114 and @xmath31 . in the limit of @xmath77 ,
one gets @xmath115 taking next the limit @xmath116 finally yields : @xmath117 where @xmath118 stands for the riemann @xmath118-function .
\(ii ) the dependence of the rhs of eq.([eq : d2crit_2d ] ) with @xmath31 is not trivial ( fig .
[ fig : d2crit_2d ] ) .
indeed it can be proved to have a maximum with respect to @xmath31 , which can be understood intuitively as follows : in the vicinity of @xmath119 , increasing @xmath31 makes the constraint less stringent since the target can be reached directly from the bulk ; in the opposite limit @xmath120 , the constraint on @xmath121 has to tend to @xmath71 since the target is found immediately from the surface .
quantitatively , in the physical limit @xmath122 , one finds that , as soon as @xmath123 , bulk excursions can be beneficial . as a function of @xmath31
is computed from eq .
( [ eq : d2crit_2d ] ) in 2d for three values of @xmath114 : @xmath124 , @xmath125 , and @xmath104 . when @xmath31 approaches @xmath126 ( the whole surface becomes absorbing ) , @xmath111 diverges ( not shown ) .
in fact , in this limit , there is no need for a bulk excursion because the target will be found immediately by the surface diffusion .
, width=302 ] if the reaction time @xmath100 is a decreasing function of the desorption rate @xmath1 , the bulk excursions are `` too favorable '' , and the best search strategy is obtained for @xmath127 ( purely bulk search ) . for the reaction time to be an optimizable function of @xmath1 , the derivative @xmath128 has to be positive at some @xmath1 . this necessary and sufficient condition remains formal and requires numerical analysis of eq .
( [ eq : t1deriv_2d ] ) . a simple _ sufficient _ condition can be used instead by demanding that the search time at zero desorption rate is less than the search time at infinite desorption rate : @xmath129 this writes in the physically relevant limit @xmath10 ( using the result of @xcite ) : @xmath130 finally , combining eqs .
( [ eq : d2crit_2d ] , [ cond2 ] ) , the search time is found to be an optimizable function of @xmath1 in the limit @xmath10 if @xmath131 ^ 2.\ ] ] knowing that @xmath132 , eq .
( [ cond3 ] ) writes in the small @xmath31 limit : @xmath133 which summarizes the conditions for the search time to be an optimizable function of @xmath1 .
this case is illustrated in fig .
[ fig : optimizable ] .
in this section , the confining domain is a sphere of radius @xmath7 and the target is the region on the boundary defined by @xmath14 $ ] , where @xmath15 is the elevation angle .
the 3d analogs of eqs .
( [ eq : t1 ] , [ eq : t2 ] ) read as @xmath134 & = & -1 \;\;{\rm for}\;\theta\in[\epsilon,\pi ] , \\
d_2\left ( \frac{\partial ^2}{\partial r^2 } + \frac{2}{r } \frac{\partial } { \partial r } + \frac{1}{r^2 } \frac{\partial ^2}{\partial \theta^2}\right ) t_2(r,\theta ) & = & -1 .
\end{aligned}\ ] ] these equations have to be completed by two boundary conditions : @xmath135 , \end{aligned}\ ] ] which respectively describe the adsorption events and express that the target is an absorbing region of the sphere .
one can search for a solution in the following form @xmath136 where @xmath137 stands for the legendre polynomial of order @xmath40 . using the orthonormality of legendre polynomials ,
the projection of @xmath138 on @xmath139 writes @xmath140 knowing that @xmath141 $ , } \\ 0 & \text{if $ \theta\in[0,\epsilon]$ , } \end{cases}\ ] ] the @xmath23 can be written in terms of @xmath142 as @xmath143 taylor expanding the rhs of @xmath144\end{aligned}\ ] ] leads to @xmath145 using eq .
( [ eq : exp3d ] ) for @xmath19 yields @xmath146 changing the order of summations over @xmath40 and @xmath41 and using the binomial formula and eq .
( [ eq : alpha_n_3d ] ) for @xmath23 finally give @xmath147 where , as in previous section , @xmath148 . this integro - differential equation for @xmath18
can actually easily be transformed into an integral equation for @xmath18 , by integrating successively two times . indeed , multiplying first both members of eq .
( [ eq : integrodiff_3d ] ) by @xmath149 and integrating between @xmath126 and @xmath15 gives @xmath150(\cos\theta+1)\nonumber\\ & + & \frac{\omega^2}{2}\sum_{n=1}^\infty ( x^n-1 ) \left(p_{n+1}(\cos\theta)-p_{n-1}(\cos\theta)\right ) \int_\epsilon^\pi \sin\theta ' p_n(\cos\theta ' ) t_1(\theta ' ) { \rm d}\theta',\end{aligned}\ ] ] where we have used @xmath151 dividing eq .
( [ eq : int3d ] ) by @xmath149 and integrating between @xmath31 and @xmath15 finally leads to @xmath152 where we have again used eq .
( [ eq : prop_legendre ] ) , or equivalently to @xmath153 with the following definitions @xmath154 in this 3d case . iterating the integral equation eq .
( [ eq : psi_3d ] ) shows that the solution @xmath55 writes for @xmath155 $ ] : @xmath156 , \ ] ] with the coefficients @xmath58 which satisfy @xmath157 = \omega \sum\limits_{n=1}^\infty \biggl(u_n + \sum\limits_{n'=1}^\infty q_{n , n ' } d_{n'}\biggr ) \bigl[p_n(\cos \theta)-p_n(\cos\epsilon)\bigr ] , \\
\end{split}\ ] ] where we introduced the new definitions @xmath158 and @xmath159 with @xmath160 in appendix [ sec : a_i ] , we compute this integral explicitly . since eq .
( [ eq : psi_3d ] ) should be satisfied for any @xmath15 , one gets @xmath161 , from which @xmath162_n \hskip 5 mm (
n = 1,2, ... ).\ ] ] as in 2d , using the series expansion of @xmath163 , eq .
( [ eq : dn_3d ] ) can be seen as a series in powers of @xmath66 , whose @xmath40-th order coefficient can be explicitly written in terms of the @xmath40-th power of the matrix @xmath67 .
once the @xmath58 determined , the search time @xmath24 can be written as @xmath167 the first terms of a perturbative expansion with respect to @xmath31 can easily be obtained from the previous exact solution . at leading order in @xmath31
, we have @xmath168 from which @xmath169_n + o(\epsilon ) = \frac{2\omega}{n(n+1 ) } ~ \frac{(1-x^n)(2n+1)}{n(n+1 ) + 2\omega ( 1-x^n ) } + o(\epsilon ) .\ ] ] one finds therefore @xmath170 averaging over @xmath15 , it finally yields : @xmath171 this result was given in @xcite without derivation . as earlier for the 2d case ,
an approximate solution can be derived . as shown numerically ( see figs .
[ 3dtest1 ] , [ 3dtest2 ] , [ 3dtest3 ] and section [ numerics ] for more details about numerical methods ) , this approximation of @xmath18 proves to be in quantitative agreement with the exact expression for a wide range of parameters .
this approximation relies on the fact that , in the small target size limit @xmath77 , the matrix @xmath67 is diagonal , which in turn mirrors the orthogonality of @xmath172 on @xmath173 $ ] .
more precisely , one has @xmath174 and keeping only the leading term of this expansion yields @xmath175 from which @xmath176 the mean time @xmath24 is then approximated as @xmath177 ^ 2}{n(n+1 ) + \omega ( 1-x^n)(2n+1 ) i_\epsilon(n , n)}\biggr]\biggr\}. \\ \end{split}\ ] ] in 3d : the exact solution ( [ eq : psi_3d ] , [ eq : dn_3d ] ) , the approximation ( [ eq : psi_3d_part ] ) and the perturbative formula ( [ eq : psi_3d_perturb ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath94 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath178 , and the series are truncated to @xmath88 .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is still small . in the * third row * , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 . the perturbative solution is inaccurate as expected for large @xmath31 .
in the * last row * , the parameters are : @xmath93 , @xmath178 , and the series are truncated to @xmath88 . in this case , the approximate solution deviates from the exact one for .
the perturbative solution is negative and not shown .
, title="fig:",width=302 ] in 3d : the exact solution ( [ eq : psi_3d ] , [ eq : dn_3d ] ) , the approximation ( [ eq : psi_3d_part ] ) and the perturbative formula ( [ eq : psi_3d_perturb ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath94 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath178 , and the series are truncated to @xmath88 .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is still small .
in the * third row * , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is inaccurate as expected for large @xmath31 . in the * last row * , the parameters are : @xmath93 , @xmath178 , and the series are truncated to @xmath88 . in this case , the approximate solution deviates from the exact one for .
the perturbative solution is negative and not shown . , title="fig:",width=302 ] in 3d : the exact solution ( [ eq : psi_3d ] , [ eq : dn_3d ] ) , the approximation ( [ eq : psi_3d_part ] ) and the perturbative formula ( [ eq : psi_3d_perturb ] ) , with @xmath84 , @xmath85 .
in the * first row * , the other parameters are : @xmath86 , @xmath94 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row
* , the parameters are : @xmath86 , @xmath178 , and the series are truncated to @xmath88 .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is still small . in the * third row
* , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is inaccurate as expected for large @xmath31 . in the * last row * , the parameters are : @xmath93 , @xmath178 , and the series are truncated to @xmath88 . in this case
, the approximate solution deviates from the exact one for .
the perturbative solution is negative and not shown .
, title="fig:",width=302 ] in 3d : the exact solution ( [ eq : psi_3d ] , [ eq : dn_3d ] ) , the approximation ( [ eq : psi_3d_part ] ) and the perturbative formula ( [ eq : psi_3d_perturb ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath94 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath178 , and the series are truncated to @xmath88 .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is still small . in the * third row
* , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is inaccurate as expected for large @xmath31 . in the * last row * , the parameters are : @xmath93 , @xmath178 , and the series are truncated to @xmath88 . in this case , the approximate solution deviates from the exact one for .
the perturbative solution is negative and not shown .
, title="fig:",width=302 ] in 3d : the exact solution ( [ eq : psi_3d ] , [ eq : dn_3d ] ) , the approximation ( [ eq : psi_3d_part ] ) and the perturbative formula ( [ eq : psi_3d_perturb ] ) , with @xmath84 , @xmath85 . in the * first row * , the other parameters are : @xmath86 , @xmath94 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath178 , and the series are truncated to @xmath88 .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is still small . in the * third row
* , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is inaccurate as expected for large @xmath31 . in the * last row * , the parameters are : @xmath93 , @xmath178 , and the series are truncated to @xmath88 . in this case , the approximate solution deviates from the exact one for .
the perturbative solution is negative and not shown .
, title="fig:",width=302 ] in 3d : the exact solution ( [ eq : psi_3d ] , [ eq : dn_3d ] ) , the approximation ( [ eq : psi_3d_part ] ) and the perturbative formula ( [ eq : psi_3d_perturb ] ) , with @xmath84 , @xmath85 . in the * first row * ,
the other parameters are : @xmath86 , @xmath94 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row * ,
the parameters are : @xmath86 , @xmath178 , and the series are truncated to @xmath88 .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is still small .
in the * third row * , the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is inaccurate as expected for large @xmath31 . in the * last row * , the parameters are : @xmath93 , @xmath178 , and the series are truncated to @xmath88 . in this case , the approximate solution deviates from the exact one for .
the perturbative solution is negative and not shown . , title="fig:",width=302 ] in 3d : the exact solution ( [ eq : psi_3d ] , [ eq : dn_3d ] ) , the approximation ( [ eq : psi_3d_part ] ) and the perturbative formula ( [ eq : psi_3d_perturb ] ) , with @xmath84 , @xmath85 . in the * first row * ,
the other parameters are : @xmath86 , @xmath94 , and the series are truncated to @xmath88 . on the right , the absolute error between the exact solution and the approximation ( dashed blue curve ) and between the exact solution and the perturbative formula ( solid red curve ) .
the approximation is very accurate indeed . in the * second row
* , the parameters are : @xmath86 , @xmath178 , and the series are truncated to @xmath88 .
one can see that the perturbative solution is inaccurate for large values of @xmath1 , while the maximal relative error of the approximate solution is still small . in the * third row * ,
the parameters are : @xmath93 , @xmath94 , and the series are truncated to @xmath88 .
the perturbative solution is inaccurate as expected for large @xmath31 . in the * last row * , the parameters are : @xmath93 , @xmath178 , and the series are truncated to @xmath88 . in this case , the approximate solution deviates from the exact one for .
the perturbative solution is negative and not shown .
, title="fig:",width=302 ] computed through eq .
( [ eq : dn_3d ] , [ eq : t1mean_3d ] ) with @xmath88 as a function of the desorption rate @xmath1 for three values of @xmath8 : @xmath179 ( dot - dashed blue line ) , @xmath180 ( dashed green line ) , and @xmath181 ( solid red line ) .
the other parameters are : @xmath97 and @xmath98 . when @xmath182 ( the first case ) , @xmath24 monotonously increases with @xmath1 so that there is no optimal value . in two other cases ,
@xmath101 , and @xmath24 starts first to decrease with @xmath1 , passes through a minimum ( the optimal value ) and then increases .
symbols show the approximate mean time computed through eq .
( [ eq : psi_3d_av ] ) .
one can see that the approximation accurate enough even for large values of @xmath1 .
* right * : the derivative @xmath183 defined by eq .
( [ eq : t1_deriv_3d ] ) for the same parameters .
, title="fig:",width=302 ] computed through eq .
( [ eq : dn_3d ] , [ eq : t1mean_3d ] ) with @xmath88 as a function of the desorption rate @xmath1 for three values of @xmath8 : @xmath179 ( dot - dashed blue line ) , @xmath180 ( dashed green line ) , and @xmath181 ( solid red line ) .
the other parameters are : @xmath97 and @xmath98 . when @xmath182 ( the first case ) , @xmath24 monotonously increases with @xmath1 so that there is no optimal value . in two other cases ,
@xmath101 , and @xmath24 starts first to decrease with @xmath1 , passes through a minimum ( the optimal value ) and then increases .
symbols show the approximate mean time computed through eq .
( [ eq : psi_3d_av ] ) .
one can see that the approximation accurate enough even for large values of @xmath1 .
* right * : the derivative @xmath183 defined by eq .
( [ eq : t1_deriv_3d ] ) for the same parameters . ,
title="fig:",width=302 ] as a function of @xmath31 , with @xmath84 , @xmath85 , and @xmath94 ( left ) or @xmath89 ( right ) .
the exact computation through eq .
( [ eq : dn_3d ] , [ eq : t1mean_3d ] ) is compared to the approximation ( [ eq : psi_3d_av ] ) and to the perturbative approach . in all cases ,
the series are truncated to @xmath88 . for small @xmath1 ( @xmath94 ) ,
the approximate solution is very close to the exact one , while the perturbative solution is relatively close for @xmath31 up to @xmath104 . in turn , for large @xmath1 ( @xmath178 )
, the approximate solution shows significant deviations for the intermediate values of @xmath31 , while the perturbative solution is not applicable at all .
, title="fig:",width=302 ] as a function of @xmath31 , with @xmath84 , @xmath85 , and @xmath94 ( left ) or @xmath89 ( right ) .
the exact computation through eq .
( [ eq : dn_3d ] , [ eq : t1mean_3d ] ) is compared to the approximation ( [ eq : psi_3d_av ] ) and to the perturbative approach . in all cases ,
the series are truncated to @xmath88 . for small @xmath1 ( @xmath94 ) ,
the approximate solution is very close to the exact one , while the perturbative solution is relatively close for @xmath31 up to @xmath104 . in turn , for large @xmath1 ( @xmath178 )
, the approximate solution shows significant deviations for the intermediate values of @xmath31 , while the perturbative solution is not applicable at all .
, title="fig:",width=302 ] we investigate here as in the 2d case the dependence of @xmath184 on @xmath1 .
the sign of @xmath185 at @xmath54 is conveniently studied by rewriting eq .
( [ eq : t1mean_3d ] ) as @xmath186 - \lambda \bigl(\xi \cdot ( i+\lambda \tilde{q})^{-1 } u\bigr ) \biggr\ } , \ ] ] where @xmath187 and @xmath188 .
the derivative of @xmath24 with respect to @xmath1 is then @xmath189 - \biggl(\xi \cdot \frac{(\eta^{-1 } + 2\lambda)i + \lambda^2 \tilde{q}}{(i+\lambda\tilde{q})^2 } u\biggr)\biggr\ } .\ ] ] if the above derivative is negative at @xmath49 , i.e. @xmath190 bulk excursions are beneficial to the search .
this inequality determines the critical value for the bulk diffusion coefficient @xmath111 ( which enters through @xmath112 ) : @xmath191 ^ 2 \biggr)^{-1 } .
\\ \end{split}\ ] ] in the limit of @xmath77 , one gets @xmath192 in the physically relevant limit @xmath10 , one has @xmath193 there are similarities and differences between the behaviors of @xmath111 in 2d and 3d .
figure [ fig : d2crit_3d ] shows that @xmath111 from eq .
( [ eq : d2crit2_3d ] ) is not a monotonous function of @xmath31 , with the qualitative explanation which is the same as in the two - dimensional case . in contrast to the analogous eq .
( [ eq : d2crit2_2d ] ) in 2d , the rhs of eq .
( [ eq : d2crit2_3d ] ) diverges as @xmath77 .
this divergence reflects the fact that a poink - like target ( @xmath194 ) , which could be found within a finite time in 2d by one - dimensional surface diffusion on the circle , is not detectable in 3d neither by bulk excursions , nor by surface diffusion . as a function of @xmath31
is computed from eq .
( [ eq : d2crit_3d ] ) in 3d for three values of @xmath114 : @xmath124 , @xmath125 , and @xmath104 . when @xmath31 approaches @xmath126 ( the whole surface becomes absorbing ) , @xmath111 diverges ( not shown ) .
in fact , in this limit , there is no need for a bulk excursion because the target will be found immediately by the surface diffusion .
in addition , @xmath111 also diverges as @xmath77 because a point - like target can not be detected neither by bulk excursions , nor by surface diffusion in 3d .
, width=302 ] for the reaction time to be an optimizable function of the desorption rate @xmath1 , it is necessary to write an additional condition , requiring that the bulk excursions are not `` too favorable '' ( otherwise , the best strategy is obtained for @xmath127 ) .
a sufficient condition is given by demanding that the search time at zero desorption rate ( i.e. , without leaving the boundary ) is less than the search time at infinite desorption rate @xmath195 @xmath196 which writes in the physically relevant limit @xmath10 ( using the result of @xcite ) : @xmath197 finally , this conditions leads , for small @xmath31 , to @xmath198 combining the two conditions ( [ eq : d2crit2_3d ] , [ cond4 ] ) , the search time is found to be optimizable when @xmath10 and @xmath199 if @xmath200 and @xmath201 by eqs .
( [ eq : optim_region_2d ] , [ eq : optim_region_3d ] ) in 2d and 3d , respectively .
when the ratio @xmath121 lies between two curves , the search time @xmath100 is optimizable with respect to @xmath1 . above the upper bound ,
surface diffusion is preferred ( @xmath49 is the optimal solution ) , while below the lower bound , bulk excursions may be `` too favorable ''
( @xmath202 may give the optimal solution ) .
we recall that the lower bound was obtained from the sufficient condition ( [ eq : cond_suff ] ) meaning that the region below the dotted line may still be optimizable . , title="fig:",width=302 ] and @xmath201 by eqs .
( [ eq : optim_region_2d ] , [ eq : optim_region_3d ] ) in 2d and 3d , respectively .
when the ratio @xmath121 lies between two curves , the search time @xmath100 is optimizable with respect to @xmath1 . above the upper bound ,
surface diffusion is preferred ( @xmath49 is the optimal solution ) , while below the lower bound , bulk excursions may be `` too favorable '' ( @xmath202 may give the optimal solution ) .
we recall that the lower bound was obtained from the sufficient condition ( [ eq : cond_suff ] ) meaning that the region below the dotted line may still be optimizable .
, title="fig:",width=302 ]
in the previous sections , we derived the closed matrix forms ( [ eq : dn_2d ] , [ eq : dn_3d ] ) for the coefficients @xmath58 in 2d and 3d .
these coefficients determine the angular dependence of @xmath142 through the explicit representations ( [ eq : psi_2d ] , [ eq : psi_3d ] ) in 2d and 3d , respectively .
although the formulas ( [ eq : dn_2d ] , [ eq : dn_3d ] ) which are based on the inversion of an infinite - dimensional matrix @xmath203 remain implicit , a numerical resolution of the problem has become straightforward .
in fact , one needs to truncate the infinite - dimensional matrix @xmath67 and vectors @xmath204 and @xmath205 and to invert the truncated matrix @xmath206 numerically .
there are six parameters that determine the function @xmath142 : the radius @xmath7 of the disk ( sphere ) , the diffusion coefficients @xmath0 and @xmath8 , the desorption rate @xmath1 , the size @xmath31 of the absorbing region , and the distance @xmath9 . from now on , we set the units of length and time by setting @xmath207 and @xmath208 .
although the distance @xmath9 may take any value from @xmath71 to @xmath7 , the physically interesting case corresponds to the limit of small @xmath9 . as we mentioned previously
, the limit @xmath51 exists but trivially leads to searching on the surface , without intermediate bulk excursions . in order to reveal the role of @xmath9
, we consider several values of @xmath9 : @xmath209 , @xmath124 , @xmath125 and @xmath104 , the latter corresponding to the specific situation when search is always restarted from the center .
since the diffusion coefficient @xmath8 enters only through the prefactor @xmath46 from eq .
( [ deft ] ) , its influence onto the searching time @xmath18 is easy to examine . in
what follows , we take three values of @xmath8 : @xmath125 , @xmath104 and @xmath210 .
the dependence of @xmath18 on the desorption rate @xmath1 and the size @xmath31 is the most interesting issue which will be studied below . in the previous sections
, we derived several formulas for computing @xmath18 : * explicit representations ( [ eq : psi_2d ] , [ eq : psi_3d ] ) with the exact expressions ( [ eq : dn_2d ] , [ eq : dn_3d ] ) for the coefficients ; * approximations ( [ eq : psi_part_2d ] , [ eq : psi_3d_part ] ) which were derived by neglecting non - diagonal elements of the matrix @xmath67 ; * perturbative formulas ( [ eq : alphanperturbation0 ] , [ eq : psi_3d_perturb ] ) which are valid for small @xmath31 . for a numerical computation of the coefficients in eqs .
( [ eq : dn_2d ] , [ eq : dn_3d ] ) , we truncate the infinite - dimensional matrix @xmath67 to a finite size @xmath211 and invert the matrix @xmath206 . in order to check the accuracy of this scheme , we compute the coefficients by taking several values of @xmath212 from @xmath210 to @xmath213 . for @xmath84 , @xmath86 , @xmath85 and @xmath94 , the computed mean time @xmath24 rapidly converges to a limit .
even the computation with @xmath214 gives the result with four significant digits .
note that other sets of parameters ( e.g. , larger values of @xmath66 ) may require larger truncation sizes .
to conclude , we have presented an exact calculation of the mean first - passage time to a target on the surface of a 2d and 3d spherical domain , for a molecule performing surface - mediated diffusion .
the presented approach is based on an integral equation which can be solved analytically , and numerically validated approximation schemes , which provide more tractable expressions of the mean fpt .
this minimal model of surface - mediated reactions , which explicitly takes into account the combination of surface and bulk diffusions , shows the importance of correlations induced by the coupling of the switching dynamics to the geometry of the confinement . indeed
, standard mf treatments prove to substantially underestimate the reaction time in this case @xcite , and sometimes even fail to reproduce the proper monotonicity @xcite . in the context of interfacial systems in confinement , our results show that the reaction time can be minimized as a function of the desorption rate from the surface , which puts forward a general mechanism of enhancement and regulation of chemical reactivity .
in this appendix , we describe another theoretical approach which relies on the explicit form of the green function of the poisson equation in 2d case . in particular , the perturbative analysis for small @xmath31 becomes easier within this approach . considering @xmath19 as a source term in the poisson type equation ( [ eq : t1 ] ) with absorbing conditions at @xmath215 and @xmath216 whose green function is well known @xcite
, @xmath18 writes @xmath217{\rm d } \theta',\end{aligned}\ ] ] and the notations @xmath218 and @xmath219 . injecting eq .
( [ eq : t2integre ] ) into eq .
( [ eq : t1integre ] ) leads to @xmath220 where , for @xmath221 integer , @xmath222 so that @xmath223 substituting eq .
( [ eq : t1complet ] ) into eq .
( [ eq : deffourier ] ) gives @xmath224 and @xmath225 where @xmath226 eq . ( [ eq : mode0 ] ) can be rearranged into @xmath227 and eq .
( [ eq : moden ] ) into @xmath228 in this case , the previous equations can be solved exactly , leading to @xmath229 and @xmath230 we note that the particular case @xmath231 is also described by these expressions , although it does not seem to be clear from eqs .
( [ eq : mode0])-([eq : moden ] ) . here again ,
( [ eq : mode0])-([eq : moden ] ) can be solved exactly , and give : @xmath233 and @xmath234 expanding @xmath235 and @xmath23 in powers of @xmath31 @xmath236 eqs .
( [ eq : mode0])-([eq : moden ] ) lead , after lengthy calculations , to @xmath237
using eq . ( [ eq : deffourier ] ) for the fourier coefficients in eq .
( [ eq : t1complet ] ) leads to a second integral equation satisfied by @xmath18 @xmath238 where @xmath239 this equation is especially well adapted to local expansions of @xmath142 in the vicinity of @xmath240 , but it can also be rearranged into the following integral equation , useful when @xmath10 : @xmath241 where @xmath242
in this appendix , we provide the explicit formula for the matrix @xmath243 in 3d case .
although technical , this is an important result for a numerical computation because it allows one to avoid an approximate integration in eq .
( [ eq : i_3d ] ) which otherwise could be a significant source of numerical errors . the formula ( [ eq : pmpn ] ) for non - diagonal elements
is somewhat elementary , while the derivation for diagonal elements seems to be original .
we denote @xmath251 using the relation @xmath252 we obtain @xmath253 the second integral is given by eq .
( [ eq : pmpn ] ) . in order to compute the first one , we consider @xmath254 + n(n+1)p_n(x)\biggr ] \\ & - x p_n(x )
\biggl[\frac{d}{dx } \biggl[(1-x^2 ) \frac{d}{dx } p_{n-1}(x)\biggr ] + ( n-1)np_{n-1}(x)\biggr]\biggr\ } \\ & = 2n \int\limits_a^b dx x p_{n-1}(x ) p_n(x ) + \biggl[xp_{n-1}(x ) ( 1-x^2 ) p'_n(x ) - xp_n(x ) ( 1-x^2 ) p'_{n-1}(x)\biggr]_a^b \\ & - \int\limits_a^b dx ( 1-x^2 ) \bigl [ p'_n(x ) p_{n-1}(x ) - p'_{n-1}(x ) p_n(x)\bigr ]
. \\ \end{split}\ ] ] the last integral can be written as @xmath255_a^b \\ & - \int\limits_a^b dx ( p_{n-1}(x)/p_n(x ) ) \bigl[-2x p_n^2(x ) + 2(1-x^2 ) p'_n(x ) p_n(x)\bigr ] \\ & = \bigl[(1-x^2 ) p_{n-1}(x ) p_n(x ) \bigr]_a^b - 2\int\limits_a^b dx \bigl[-x p_{n-1}(x)p_n(x ) + ( 1-x^2 ) p'_n(x ) p_{n-1}(x)\bigr ] . \\ \end{split}\ ] ] in the last term , we substitute @xmath256 to get @xmath257_a^b + 2\int\limits_a^b dx x p_{n-1}(x)p_n(x ) - 2\int\limits_a^b dx \bigl[-nxp_n(x ) + np_{n-1}(x)\bigr ] p_{n-1}(x ) .\ ] ] bringing these results together , we get @xmath258_a^b \\ & + \bigl[(1-x^2 ) p_{n-1}(x ) p_n(x ) \bigr]_a^b + 2\int\limits_a^b dx x
p_{n-1}(x)p_n(x ) - 2\int\limits_a^b dx \bigl[-nxp_n(x ) + np_{n-1}(x)\bigr ] p_{n-1}(x ) \\ \end{split}\ ] ] so that @xmath259 + ( 1-x^2 ) p_{n-1}(x ) p_n(x)\biggr]_a^b + \frac{n}{2n+1 } k_{n-1 } .\ ] ] we obtain @xmath260 + ( 1-x^2 ) p_{n-1}(x ) p_n(x)\biggr]_a^b + \frac{2n-1}{2n+1 } k_{n-1 } \\ & - \frac{n-1}{n } \biggl[\frac{2xp_{n-2}(x)p_n(x ) - np_{n-1}(x)p_{n-2}(x ) + ( n-2)p_{n-3}(x)p_n(x)}{2(2n-1)}\biggr]_a^b . \\ \end{split}\ ] ] we can further simplify this expression by using the following identities @xmath261 we get @xmath262 + ( 1-x^2 ) p_{n-1}(x ) p_n(x)\biggr]_a^b + \frac{2n-1}{2n+1 } k_{n-1 } \\
& - \frac{n-1}{2n(2n-1 ) } \biggl[(2n-1)x p_n(x)p_{n-2}(x ) - np_{n-1}(x)p_{n-2}(x ) - ( n-1)p_{n-1}(x)p_n(x)\biggr]_a^b \\ % & = - \frac{2n-1}{2n(2n+1 ) } \biggl[nx \bigl[p_{n-1}^2(x ) + p_n^2(x ) - 2x p_n(x ) p_{n-1}(x ) ] + ( 1-x^2 ) p_{n-1}(x ) p_n(x)\biggr]_a^b + \frac{2n-1}{2n+1 } k_{n-1 } \\ & - \frac{1}{2n } \biggl[((2n-1)x^2 + 1 ) p_n(x)p_{n-1}(x ) - nx ( p_{n-1}^2(x ) + p_n^2(x))\biggr]_a^b \\ & = \frac{\bigl[x ( p_{n-1}^2(x ) + p_n^2(x ) ) - 2p_n(x ) p_{n-1}(x)\bigr]_a^b}{2n+1 } + \frac{2n-1}{2n+1 } k_{n-1 } \\
\end{split}\ ] ] and we know that @xmath263 .
applying this formula recursively , one finds @xmath264 where @xmath265 \\ & - 2p_n(x)p_{n-1}(x ) - 2p_{n-1}(x)p_{n-2}(x ) - ... - 2p_1(x)p_0(x ) + x
\\ & = \sum\limits_{k=1}^n \bigl[2(x-1)p_k^2(x ) + [ p_k(x ) - p_{k-1}(x)]^2\bigr ] - ( x-1)p_n^2(x ) + ( x-1)p_0 ^ 2(x ) + x
.\\ \end{split}\ ] ] one can check that this function satisfies the recurrent relation @xmath266 - 2p_n(x ) p_{n-1}(x ) , \hskip 5 mm f_0(x ) = x .\ ] ] note that @xmath267 . | we present an exact calculation of the mean first - passage time to a target on the surface of a 2d or 3d spherical domain , for a molecule alternating phases of surface diffusion on the domain boundary and phases of bulk diffusion .
the presented approach is based on an integral equation which can be solved analytically .
numerically validated approximation schemes , which provide more tractable expressions of the mean first - passage time are also proposed . in the framework of this minimal model of surface - mediated reactions , we show analytically that the mean reaction time can be minimized as a function of the desorption rate from the surface . |
recently , demand for large - scale complex optimization is increasing in computational science , engineering and many of other fields . in that kind of problems
, there are many difficulties caused by noise in function evaluation , many tuning parameters and high computation cost . in such cases ,
derivatives of the objective function are unavailable or computationally infeasible .
these problems can be treated by the derivative - free optimization ( dfo ) methods .
dfo is the tool for optimization without derivative information of the objective function and constraints , and it has been widely studied for decades @xcite .
dfo algorithms include gradient descent methods with a finite difference gradient estimation @xcite , some direct search methods using only function values @xcite , and trust - region methods @xcite .
there is , however , a more restricted setting in which not only derivatives but also values of the objective function are unavailable or computationally infeasible . in such a situation , the so - called pairwise comparison oracle , that tells us an order of function values on two evaluation points , is used instead of derivatives and function evaluation @xcite .
for example , the pairwise comparison is used in learning to rank to collect training samples to estimate the preference function of the ranking problems @xcite . in decision making , finding the most preferred feasible solution from among the set of many alternatives is an important application of ranking methods using the pairwise comparison .
also , other type of information such as stochastic gradient - sign oracle has been studied @xcite .
now , let us introduce two dfo methods , i.e. , the nelder - mead method @xcite and stochastic coordinate descent algorithm @xcite .
they are closely related to our work . in both methods ,
the pairwise comparison of function values is used as a building block in optimization algorithms .
nelder and mead s downhill simplex method @xcite was proposed in early study of algorithms based on the pairwise comparison of function values . in each iteration of the algorithm , a simplex that approximates the objective function
is constructed according to ranking of function values on sampled points .
then , the simplex receives four operations , namely , reflection , expansion , contraction and reduction in order to get close to the optimal solution .
unfortunately , the convergence of the nelder - mead algorithm is theoretically guaranteed only in low - dimension problems @xcite . in high dimensional problems ,
the nelder - mead algorithm works poorly as shown in @xcite . the stochastic coordinate descent algorithm using only the noisy pairwise comparison was proposed in @xcite .
lower and upper bounds of the convergence rate were also presented in terms of the number of pairwise comparison of function values , i.e. , query complexity .
the algorithm iteratively solves one dimensional optimization problems like the coordinate descent method .
however , practical performance of the optimization algorithm was not studied in that work . in this paper , we focus on optimization algorithms using the pairwise comparison oracle . in our algorithm ,
the convergence to the optimal solution is guaranteed , when the number of pairwise comparison tends to infinity .
our algorithm is regarded as a block coordinate descent method consisting of two steps : the direction estimate step and search step . in the direction estimate step , the search direction is determined . in the search step , the current solution is updated along the search direction with an appropriate step length . in our algorithm , the direction estimate step is easily parallelized .
therefore , it is expected that our algorithm effectively works even in large - scale optimization problems .
let us summarize the contributions presented in this paper . 1 .
we propose a block coordinate descent algorithm based on the pairwise comparison oracle , and point out that the algorithm is easily parallelized .
we derive an upper bound of the convergence rate in terms of the number of pairwise comparison of function values , i.e. , query complexity .
we show a practical efficiency of our algorithm through numerical experiments . the rest of the paper is organized as follows . in section [ pre ] , we explain the problem setup and give some definitions
. section [ mai ] is devoted to the main results .
the convergence properties and query complexity of our algorithm are shown in the section . in section [ num ] ,
numerical examples are reported . finally in section [ con ]
, we conclude the paper with the discussion on future works .
all proofs of theoretical results are found in appendix .
in this section , we introduce the problem setup and prepare some definitions and notations used throughout the paper .
a function @xmath0 is said to be @xmath1-strongly convex on @xmath2 for a positive constant @xmath1 , if for all @xmath3 , the inequality @xmath4 holds , where @xmath5 and @xmath6 denote the gradient of @xmath7 at @xmath8 and the euclidean norm , respectively .
the function @xmath7 is @xmath9-strongly smooth for a positive constant @xmath9 , if @xmath10 holds for all @xmath11 .
the gradient @xmath12 of the @xmath9-strongly smooth function @xmath7 is referred to as @xmath9-lipschitz gradient .
the class of @xmath1-strongly convex and @xmath9-strongly smooth functions on @xmath13 is denoted as @xmath14 . in the convergence analysis , mainly we focus on the optimization of objective functions in @xmath14 .
we consider the following pairwise comparison oracle defined in @xcite .
the stochastic pairwise comparison ( pc ) oracle is a binary valued random variable @xmath15 defined as @xmath16 \ge \frac{1}{2 } + \min \{\delta_0 , \mu |f(\y ) - f(\x)|^{\kappa-1 } \ } , \label{eqp3}\ ] ] where @xmath17 , @xmath18 and @xmath19 . for @xmath20 , without loss of generality @xmath21 is assumed .
when the equality @xmath22 = 1\ ] ] is satisfied for all @xmath8 and @xmath23 , we call @xmath24 the deterministic pc oracle . for @xmath20 ,
the probability in ( [ eqp3 ] ) is not affected by the difference @xmath25 , meaning that the probability for the output of the pc oracle is not changed under any monotone transformation of @xmath7 . in @xcite ,
jamieson et al . derived lower and upper bounds of convergence rate of an optimization algorithm using the stochastic pc oracle .
the algorithm is referred to as the original pc algorithm in the present paper . under above preparations ,
our purpose is to find the minimizer @xmath26 of the objective function @xmath27 in @xmath28 by using pc oracle . in the following section , we provide a dfo algorithm in order to solve the optimization problem and consider the convergence properties including query complexity .
initial point @xmath29 , and accuracy in line search @xmath30 . set @xmath31 .
choose @xmath32 coordinates @xmath33 out of @xmath34 coordinates according to the uniform distribution .
solve the one - dimension optimization problems @xmath35 within the accuracy @xmath36 using the pc - based line search algorithm shown in algorithm [ alg2 ] , where @xmath37 denotes the @xmath38-th unit basis vector .
then , obtain the numerical solutions @xmath39 .
set @xmath40 .
if @xmath41 is the zero vector , add @xmath36 to @xmath42 .
apply algorithm [ alg2 ] to obtain a numerical solution @xmath43 of @xmath44 within the accuracy @xmath45 . @xmath46 .
@xmath47 in algorithm [ alg1 ] , we propose a dfo algorithm based on the pc oracle . in our algorithm
, @xmath32 coordinates out of @xmath34 elements are updated in each iteration to efficiently cope with high dimensional problems .
algorithm [ alg1 ] is referred to as @xmath48 $ ] .
the original pc algorithm is recovered by setting @xmath49 .
the pc oracle is used in the line search algorithm to solve one - dimensional optimization problems ; the detailed line search algorithm is shown in algorithm [ alg2 ] . for @xmath50 , the search direction @xmath41 in algorithm [ alg1 ]
approximates that of a modified newton method ( * ? ? ?
* chap.10 ) , as shown below . in step
d-1 of the algorithm , one - dimensional optimization problems are solved .
let @xmath51 be the optimal solution of with @xmath52 . then , @xmath51 will be close to the numerical solution @xmath53 .
the taylor expansion of the objective function leads to @xmath54 where @xmath55 is the hessian matrix of @xmath7 at @xmath56 .
when the point @xmath56 is close to the optimal solution of @xmath27 , the optimal parameter @xmath57 will be close to zero , implying that the higher order term @xmath58 in the above is negligible .
hence , @xmath53 is approximated by the optimal solution of the quadratic approximation , i.e. , @xmath59 . as a result , the search direction in @xmath60 $ ] is approximated by @xmath61 , where @xmath62 denotes the diagonal matrix , the diagonal elements of which are those of the square matrix @xmath63 . in the modified newton method , the hessian matrix in the newton method is replaced with a positive definite matrix to reduce the computation cost . using only the diagonal part of the hessian matrix is a popular choice in the modified newton method .
current solution @xmath64 , search direction @xmath65 and accuracy in line search @xmath30 .
set @xmath66 , @xmath67 , @xmath68 , @xmath69 .
@xmath70 @xmath71 ( double - sign corresponds ) @xmath72 , @xmath73 @xmath74 , @xmath75 , @xmath76 @xmath77 , @xmath78 , @xmath79 ( double - sign corresponds ) @xmath80 , @xmath81 @xmath82 figure [ fg1 ] demonstrates an example of the optimization process of both the original pc algorithm and our algorithm .
the original pc algorithm updates the numerical solution along a randomly chosen coordinate in each iteration .
hence , many iterations are required to get close to the optimal solution . on the other hand , in our algorithm
, a solution can move along a oblique direction . therefore
, our algorithm can get close to the optimal solution with less iterations than the original pc algorithm . with same initialization .
left panel : jamieson et al.s original pc algorithm .
right panel : proposed algorithm . , title="fig : " ] with same initialization . left panel : jamieson et al.s original pc algorithm .
right panel : proposed algorithm . , title="fig : " ] we now provide an upper bound of the convergence rate of our algorithm using the deterministic pc oracle .
let us denote the minimizer of @xmath7 as @xmath26 .
[ upper ] suppose @xmath83 , and define @xmath84 and @xmath85 be @xmath86 let us define @xmath87 be @xmath88 for @xmath89 , we have @xmath90\leq{}\varepsilon$ ] , where the expectation is taken with respect to the random choice of coordinates @xmath91 to be updated in @xmath48 $ ] .
the proof of theorem [ upper ] is given in [ proof.upper ] .
note that any monotone transformation of the objective function does not affect the output of the deterministic pc oracle .
hence , the theorem above holds even for the function @xmath27 such that the composite function with a monotone function is included in @xmath14 .
let @xmath92 be the output of blockcd@xmath93 $ ] after @xmath94 pairwise comparison queries .
to solve the one dimension optimization problem within the accuracy @xmath36 , the sufficient number of the call of pc - oracle is @xmath95 as shown in @xcite .
hence , if the inequality @xmath96 holds , theorem [ upper ] assures that the numerical solution @xmath92 based on @xmath94 queries satisfies @xmath97\leq\varepsilon.\ ] ] when @xmath98 is sufficiently large , we have @xmath99 where @xmath100 is a constant depending on @xmath101 and @xmath102 .
the last inequality holds if @xmath103 and @xmath104 are both greater than @xmath105 .
eventually we have @xmath106\leq { } \exp\left\{-\frac{c}{\log{n}}\sqrt{\frac{q}{n}}\right\ } , \ ] ] where @xmath107 .
the above bound is of the same order of the convergence rate for the original pc algorithm up to polylog factors . on the other hand ,
a lower bound presented in @xcite is of order @xmath108 with a positive constant @xmath109 up to polylog factors , when the pc oracle with @xmath110 is used . in theorem
[ upper ] , it is assumed that the objective function is strongly convex and gradient lipschitz . in a realistic situation
, we usually do not have the knowledge of the class parameters @xmath1 and @xmath9 of the unknown objective function . moreover , strong convexity and gradient lipschitzness on the whole space @xmath111 is too strong . in the following corollary
, we relax the assumption in theorem [ upper ] and prove the convergence property of our algorithm without strong convexity and strong smoothness . [ convergence ]
let @xmath112 be a twice continuously differentiable convex function with non - degenerate hessian on @xmath111 and @xmath26 be a minimizer of @xmath7 .
then , there is a constant @xmath109 such that the output @xmath113 of blockcd[n , m ] satisfies ( [ eq : quecom ] ) .
the proof of corollary [ convergence ] is given in [ proof.convergence ] .
@xmath3 , @xmath114 $ ] , @xmath115 set @xmath116 and toss the coin with probability @xmath117 of heads once .
@xmath118 @xmath119 in stochastic pc oracle , one needs to ensure that the correct information is obtained in high probability . in algorithm
[ alg3 ] , the query @xmath120 is repeated under the stochastic pc oracle .
the reliability of line search algorithm based on stochastic pc oracle was investigated by @xcite .
[ @xcite ] [ sto_line ] for any @xmath3 with @xmath114 $ ] , the repeated querying subroutine in algorithm [ alg3 ] correctly identifies the sign of @xmath121 $ ] with probability @xmath122 , and requests no more than @xmath123 queries .
it should be noted here that , in this paper , @xmath124\}= { \rm sign } \{f(y)-f(x)\}$ ] always holds because @xmath125 from ( [ eqp3 ] ) . in @xcite , a modified line search algorithm using a ternary search instead of bisection search
was proposed to lower bound @xmath126 in lemma [ sto_line ] for arbitrary @xmath3 .
then , one can find that the total number of queries required by a repeated querying subroutine algorithm is at most @xmath127 , where @xmath45 is an accuracy of line search .
the query complexity of the stochastic pc oracle is obtained from that of the deterministic pc oracle .
suppose that one has @xmath128 responses from the deterministic pc oracle . to obtain the same responses from the stochastic pc oracle with probability more than @xmath129 , one needs more than @xmath130 queries . from the above discussion
, we have the following upper bounds for stochastic setting : @xmath131 \le \begin{cases } \displaystyle \exp \left \{-\frac{c_{1}}{\log{n } } \sqrt{\frac{q}{n } } \right \ } , & \kappa = 1 , \vspace*{2 mm } \\
\displaystyle c_{2 } \displaystyle \frac{n^2}{m } \left(\frac{n}{q } \right)^{1/(2\kappa - 2 ) } , & \kappa > 1 , \end{cases } \label{eq : sto_compl}\ ] ] where @xmath132 and @xmath133 are constant depending on @xmath101 and @xmath134 as well as @xmath135 poly - logarithmically .
if @xmath32 and @xmath34 are of the same order in the case of @xmath136 , the bound coincides with that shown in theorem 2 of @xcite .
in this section , we present numerical experiments in which the proposed method in algorithm [ alg1 ] was mainly compared with the nelder - mead algorithm @xcite and the original pc algorithm @xcite , i.e. , blockcd@xmath137 $ ] of algorithm [ alg1 ] . here , the pc oracle was used in all the optimization algorithms . in blockcd@xmath93 $ ] with @xmath138
, one can execute the line search algorithm to each axis separately .
hence , the parallel computation is directly available to find the components of the search direction @xmath41 . also , we investigated the computation efficiency of the parallel implementation of our method .
the numerical experiments were conducted on amd opteron processor 6176 ( 2.3ghz ) with 48 cores , running cent os linux release 6.4 .
we used the r language @xcite with snow library for parallel statistical computing .
it is well - known that the nelder - mead method efficiently works in low dimensional problems .
indeed , in our preliminary experiments for two dimensional optimization problems , the nelder - mead method showed a good convergence property compared to the other methods such as blockcd@xmath139 $ ] with @xmath140 .
numerical results are presented in figure [ fig:2_dim_problems ] .
we tested optimization methods on the quadratic function @xmath141 , and two - dimension rosenbrock function , @xmath142 , where the matrix @xmath63 was a randomly generated 2 by 2 positive definite matrix . in two dimension problems
, we do not use the parallel implementation of our method , since clearly the parallel computation is not efficient that much .
the efficiency of parallel computation is canceled by the communication overhead . in our method , the accuracy of the line search is fixed to a small positive number @xmath45 .
hence , the optimization process stops on the way to the optimal solution , as shown in the left panel of fig .
[ fig:2_dim_problems ] . on the other hand , the nelder - mead method tends to converge to the optimal solution in high accuracy . in terms of the convergence speed for the optimization of two - dimensional quadratic function , there is no difference between the nelder - mead method and blockcd method , until the latter stops due to the limitation of the numerical accuracy . even in non - convex rosenbrock function ,
the nelder - mead method works well compared to the pc - based blockcd algorithm . [ cols="^,^ " , ]
in this paper , we proposed a block coordinate descent algorithm for unconstrained optimization problems using the pairwise comparison of function values .
our algorithm consists of two steps : the direction estimate step and search step .
the direction estimate step can easily be parallelized . hence
, our algorithm is effectively applicable to large - scale optimization problems .
theoretically , we obtained an upper bound of the convergence rate and query complexity , when the deterministic and stochastic pairwise comparison oracles were used .
practically , our algorithm is simple and easy to implement . in addition
, numerical experiments showed that the parallel implementation of our algorithm outperformed the other methods .
an extension of our algorithm to constrained optimization problems is an important future work .
other interesting research directions include pursuing the relation between pairwise comparison oracle and other kind of oracles such as gradient - sign oracle @xcite .
10 charles audet and john e dennis jr .
analysis of generalized pattern searches .
, 13(3):889903 , 2002 .
stephen poythress boyd and lieven vandenberghe .
. cambridge university press , 2004 . a andrew r conn , katya scheinberg , and luis n vicente . ,
volume 8 .
siam , 2009 .
andrew r conn , nicholas i m gould , and ph l toint .
, volume 1 .
siam , 2000 .
a. d. flaxman , a. t. kalai , and h. b. mcmahan .
online convex optimization in the bandit setting : gradient descent without a gradient . in
_ proceedings of the sixteenth annual acm - siam symposium on discrete algorithms _ , soda 05 , pages 385394 , philadelphia , pa , usa , 2005 .
society for industrial and applied mathematics .
m. c. fu .
gradient estimation . in s. g. henderson and b. l. nelson , editors , _ handbooks in operations
research and management science : simulation _ , chapter 19 .
elservier amsterdam , 2006 .
fuchang gao and lixing han .
implementing the nelder - mead simplex algorithm with adaptive parameters .
, 51(1):259277 , 2012 .
k. g. jamieson , r. d. nowak , and b. recht .
query complexity of derivative - free optimization . in _ nips _ , pages 26812689 , 2012 .
matti kriinen .
active learning in the non - realizable case . in _ algorithmic learning theory _ , pages 6377 .
springer , 2006 .
jeffrey c. lagarias , james a. reeds , margaret h. wright , and paul e. wright .
convergence properties of the nelder - mead simplex method in low dimensions .
, 9:112147 , 1998 .
d. luenberger and y. ye . .
springer , 2008 .
mehryar mohri , afshin rostamizadeh , and ameet talwalkar . .
mit press , 2012 .
j. a. nelder and r. mead . a simplex method for function minimization . , 7(4):308313 , 1965 . .
r foundation for statistical computing , vienna , austria , 2014 .
aaditya ramdas and aarti singh .
algorithmic connections between active learning and stochastic convex optimization . in _ algorithmic learning theory _ , pages 339353 .
springer , 2013 .
aaditya ramdas and aarti singh .
algorithmic connections between active learning and stochastic convex optimization . in sanjay jain , rmi munos , frank stephan , and thomas zeugmann , editors , _ alt _ , volume 8139 of _ lecture notes in computer science _ , pages 339353 .
springer , 2013 .
luis miguel rios and nikolaos v sahinidis .
derivative - free optimization : a review of algorithms and comparison of software implementations .
, 56(3):12471293 , 2013 .
the optimal solution of @xmath7 is denoted as @xmath26 .
let us define @xmath143 be @xmath144 . if @xmath145 holds in the algorithm , we obtain @xmath146 , since the function value is non - increasing in each iteration of the algorithm is assured by a minor modification of pc - oracle in @xcite . ] .
next , we assume @xmath147 .
the assumption leads to @xmath148 in which the second inequality is derived from ( 9.9 ) in @xcite . in the following ,
we use the inequality @xmath149 that is proved in @xcite . for
the @xmath38-th coordinate , let us define the functions @xmath150 and @xmath151 as @xmath152 then , we have @xmath153 let @xmath154 and @xmath155 be the minimum solution of @xmath156 and @xmath157 , respectively . then , we obtain @xmath158 the inequality @xmath159 yields that @xmath155 lies between @xmath160 and @xmath161 , where @xmath100 and @xmath132 are defined as @xmath162 here , @xmath163 holds .
each component of the search direction @xmath164 in algorithm [ alg1 ] satisfies @xmath165 if @xmath166 and otherwise @xmath167 . for @xmath168 ,
let @xmath169 of the vector @xmath170 be @xmath171 .
then , the triangle inequality leads to @xmath172 the assumption @xmath147 and the inequalities @xmath173 lead to @xmath174 hence , we obtain @xmath175_+ , \end{aligned}\ ] ] where @xmath176_+=\max\{0,x\}$ ] for @xmath177 .
let @xmath178 be a non - negative valued random variable defined from the random set @xmath179 , and define the non - negative value @xmath180 as @xmath181 .
a lower bound of the expectation of @xmath182 with respect to the distribution of @xmath179 is given as @xmath183 & \geq \ebb_i\left [ \left ( \frac { \left [ c_0\|\nabla{f}(\x)\|_{i}^2-\frac{c_0}{2}\sqrt{\frac{m}{n}}\|\nabla{f}(\x_t)\|\|\nabla{f}(\x_t)\|_i \right]_+ } { c_1\|\nabla{f}(\x_t)\|_i + \frac{c_0}{2}\sqrt{\frac{m}{n}}\|\nabla{f(\x_t)}\| } \right)^2 \right ] \\ & = k^2\frac{m}{n}\|\nabla{f(\x_t)}\|^2\ebb_i\left[{z^2 } \frac{[z-1/2]_+^2}{(z+k/2)^2}\right]\\ & \geq k^2\frac{m}{n}\|\nabla{f(\x_t)}\|^2 \ebb_i\left[{z^2 } \frac{[z-1/2]_+^2}{(z+1/2)^2}\right ] .
\end{aligned}\ ] ] the random variable @xmath184 is non - negative , and @xmath185=1 $ ] holds .
thus , lemma [ eqn : lemma_bound_expz ] in the below leads to @xmath183 & \geq \frac{k^2}{53}\frac{m}{n}\|\nabla{f}(\x_t)\|^2 . \end{aligned}\ ] ] eventually , if @xmath147 , the conditional expectation of @xmath186 for given @xmath187 is given as @xmath188 & \leq f(\x_t)-f(\x^*)- \frac{k^2}{106l}\frac{m}{n}\|\nabla{f}(\x_t)\|^2 + \frac{l\eta^2}{2}\\ & \leq \left(1-\frac{m}{n}\gamma\right)(f(\x_t)-f(\x^ * ) ) + \frac{l\eta^2}{2}. \end{aligned}\ ] ] combining the above inequality with the case of @xmath145 , we obtain @xmath189\\ & \leq \1[f(\x_{t})-f(\x^*)\geq\varepsilon']\cdot \left [ \left(1-\frac{m}{n}\gamma\right)(f(\x_{t})-f(\x^ * ) ) + \frac{l\eta^2}{2}\right ] + \1[f(\x_{t})-f(\x^*)<\varepsilon ' ] \cdot\varepsilon ' . \end{aligned}\ ] ] the expectation with respect to all @xmath190 yields @xmath191 & \leq \left(1-\frac{m}{n}\gamma\right ) \ebb[\1[f(\x_{t})-f(\x^*)\geq\varepsilon'](f(\x_{t})-f(\x^ * ) ) ] \\ & \phantom{\leq } + \ebb[\1[f(\x_{t})-f(\x^*)\geq\varepsilon']]\frac{l\eta^2}{2 } + \ebb[\1[f(\x_{t})-f(\x^*)<\varepsilon ' ] ] \varepsilon ' \\ &
\leq \left(1-\frac{m}{n}\gamma\right ) \ebb[f(\x_{t})-f(\x^ * ) ] + \max\left\ { \frac{l\eta^2}{2},\,\varepsilon ' \right\}.\end{aligned}\ ] ] since @xmath192 and @xmath193 hold , for @xmath194 $ ] we have @xmath195 when @xmath196 is greater than @xmath87 in , we obtain @xmath197 and @xmath198 let us consider the accuracy of the numerical solution @xmath199 . as shown in ( * ? ? ? * chap .
9 ) , the inequality @xmath200 holds .
thus , for @xmath89 , we have @xmath201 ^ 2\leq \mathbb{e}[\|\x_t-\x^*\|^2]\leq\frac{8l}{\sigma^2}\varepsilon = 64n\left(\frac{l}{\sigma}\right)^3\left(1+\frac{n}{m\gamma}\right)\eta^2 . \end{aligned}\ ] ] [ eqn : lemma_bound_expz ] let @xmath184 be a non - negative random variable satisfying @xmath202=1 $ ] .
then , we have @xmath203_+^2}{(z+1/2)^2}\right]\geq\frac{1}{53}. \end{aligned}\ ] ] for @xmath204 and @xmath205 , we have the inequality @xmath206_+^2}{(z+1/2)^2 } \geq \frac{\delta^2}{(1+\delta)^2}\1[z\geq1/2+\delta ] . \end{aligned}\ ] ] then ,
we get @xmath207_+^2}{(z+1/2)^2}\right ] & \geq \frac{\delta^2}{(1+\delta)^2}\ebb[z^2 \1[z\geq1/2+\delta]]\\ & = \frac{\delta^2}{(1+\delta)^2}\ebb[z^2(1-\1[z<1/2+\delta])]\\ & = \frac{\delta^2}{(1+\delta)^2 } \left ( 1-\ebb[z^2\1[z<1/2+\delta ] ] \right)\\ & \geq \frac{\delta^2}{(1+\delta)^2 } \left ( 1-(1/2+\delta)^2\pr(z<1/2+\delta ) \right)\\ & \geq \frac{\delta^2}{(1+\delta)^2 } \left ( 1-(1/2+\delta)^2 \right ) . \end{aligned}\ ] ] by setting @xmath208 appropriately , we obtain @xmath209_+^2}{(z+1/2)^2}\right]\geq\frac{1}{53}.\end{aligned}\ ] ]
for the output @xmath210 of blockcd@xmath93 $ ] , @xmath211 holds , and thus , the sequence @xmath212 is included in @xmath213 since @xmath7 is convex and continuous , @xmath214 is convex and closed . moreover , since @xmath7 is convex and it has non - degenerate hessian , the hessian is positive definite , and thus , @xmath7 is strictly convex .
then @xmath214 is bounded as follows .
we set the minimul directional derivative along the radial direction from @xmath26 over the unit sphere around @xmath26 as @xmath215 then , @xmath216 is strictly positive and the following holds for any @xmath217 such that @xmath218 , @xmath219 thus we have @xmath220 since the right hand side of ( [ include ] ) is a bounded ball , @xmath214 is also bounded .
thus , @xmath214 is a convex compact set . since @xmath7 is twice continuously differentiable , the hessian matrix @xmath221 is continuous with respect to @xmath222 . by the positive definiteness of the hessian matrix ,
the minimum and maximum eigenvalues @xmath223 and @xmath224 of @xmath221 are continuous and positive
. therefore , there are the positive minimum value @xmath1 of @xmath223 and maximum value @xmath9 of @xmath224 on the compact set @xmath214 .
it means that @xmath7 is @xmath1-strongly convex and @xmath9-lipschitz on @xmath214 .
thus , the same argument to obtain ( [ eq : quecom ] ) can be applied for @xmath7 . | this paper provides a block coordinate descent algorithm to solve unconstrained optimization problems . in our algorithm ,
computation of function values or gradients is not required . instead , pairwise comparison of function values is used .
our algorithm consists of two steps ; one is the direction estimate step and the other is the search step .
both steps require only pairwise comparison of function values , which tells us only the order of function values over two points . in the direction estimate step ,
a newton type search direction is estimated . a computation method like block
coordinate descent methods is used with the pairwise comparison . in the search step , a numerical solution is updated along the estimated direction . the computation in the direction estimate step can be easily parallelized , and thus , the algorithm works efficiently to find the minimizer of the objective function . also , we show an upper bound of the convergence rate . in numerical experiments
, we show that our method efficiently finds the optimal solution compared to some existing methods based on the pairwise comparison . |
in the standard unified model of active galactic nuclei ( agn ) , the observed differences between agn sub - classes are solely explained as the result of an orientation effect with respect to an optically and geometrically thick distribution of molecular gas and dust on tens of parsecs scales , frequently assumed to have a toroidal geometry ( @xcite ; @xcite ) . this material , hereafter referred to as the dusty torus , absorbs the agn bolometric luminosity output produced at rest - frame uv / optical and x - ray wavelengths in the accretion disk / corona and re - emits it mainly in the mid - infrared ( mid - ir ) regime ( rest - frame wavelengths @xmath75 - 30@xmath0 m ) .
thus , it is clear that the mid - ir spectral range offers a unique opportunity to study the obscured accretion phenomenon .
surveys conducted using data from spitzer - irac have identified a population of highly obscured agn whose number density is comparable to that observed for unobscured agn .
many of these objects might be hidden at optical and x - ray wavelengths by compton - thick material ( line - of - sight rest - frame gas column densities @xmath10 ; @xcite ) .
uncovering the mid - ir thermal continuum associated with the agn dusty torus emission can be a difficult task , due to contamination from the agn host galaxies .
high ( subarcsecond ) spatial resolution observations can isolate the agn emission from other contaminants on nearby objects with favourable geometries .
alternatively , mid - ir spectral decomposition can provide reliable estimates of the agn dusty torus contribution to the infrared light ( e.g. @xcite ) .
studies such as these have found a tight correlation between the intrinsic luminosity , traced by hard ( 2 - 10 kev ) x - rays , and reprocessed mid - infrared emission of agn over more than three orders of magnitude in luminosity ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
the relationship between hard x - ray and mid - ir luminosities seems to be identical for agn optically classified as type 1 and type 2 , even for compton - thick agn .
if the hard x - ray emission is a good tracer of both the intrinsic power of agn and the heating source of the dusty torus , then these studies provide strong observational support for the mid - ir light being a reliable proxy for the agn intrinsic power , regardless of any obscuration .
unfortunately , the agn samples involved in these studies are highly incomplete and biased towards infrared - bright objects . recent studies of local agn detected with the swift burst alert telescope and akari all - sky surveys have shown a tight correlation between the hard x - ray ( @xmath1110 kev energies ) and monochromatic mid - ir ( 9 and 18@xmath0 m ) luminosities and that unabsorbed and absorbed agn follow the same correlation ( @xcite ; @xcite ) . the sensitive all - sky infrared survey conducted with the wide - field infrared survey explorer ( @xmath12 ; @xcite ) at 3.4 , 4.6 , 12 and 22@xmath0 m will dramatically increase our agn census over a significant fraction of the age of the universe improving greatly our understanding of the obscured accretion phenomenon in luminous systems .
it has already been demonstrated that mid - ir colour - based selection techniques with @xmath12 , targeting the characteristic red power - law thermal continuum emission from the agn dusty torus ( @xmath13@xmath4@xmath14 with @xmath15@xmath16 - 0.5 ; @xcite ) , are highly reliable and effective at uncovering both unobscured and obscured luminous agn ( e.g. @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . as demonstrated in @xcite
, @xmath12 can potentially uncover many compton - thick agn missed at other wavelengths , at least up to @xmath17@xmath71 . despite such progress , the physical properties of the agn revealed with @xmath12 remain poorly constrained . in this paper
we investigate how well we can isolate the agn emission associated to the dusty torus in the mid - ir regime using the @xmath12 broad band photometric data and , more importantly , whether the mid - ir continuum luminosities derived in this way are a reliable isotropic proxy for the intrinsic luminosity of agn .
to do so we have conducted a uv - to - mid - ir spectral energy distribution decomposition into agn and galaxy components .
then we have studied the correspondence between the agn luminosities at rest - frame 6@xmath0 m , corrected for the accretion disk and host galaxy emission , and at 2 - 10 kev energies over more than three orders of magnitude in agn luminosity .
we have used the final data release of the @xmath12 survey ( allwise ; @xcite ) and a large uniform complete sample of 232 x - ray selected agn drawn from the bright ultra - hard xmm - newton survey .
the objects have @xmath17 in the range 0.05 - 2.8 , and intrinsic ( absorption - corrected ) 2 - 10 kev x - ray luminosities between @xmath1 and @xmath18 .
the x - ray luminosities , corrected for absorption , were computed from a detailed analysis of the _ xmm - newton _ x - ray spectroscopic data available for all objects .
this paper is structured as follows .
section2 describes the agn sample and the multiwavelength data used in this study . in section3 we describe the sed fitting approach used to isolate the agn mid - ir emission , the adopted technique to compute rest - frame 6@xmath0 m luminosities and discuss the effects associated with contamination from the agn host galaxies . in section4
we present the relationship between 6@xmath0 m and 2 - 10 kev continuum luminosities .
the main results are discussed in section5 and summarized in section6 . throughout this paper errors
are 68 per cent confidence for a single parameter , and we assume @xmath19 , @xmath20 and @xmath21 . [ cols="<,^,^,^,^,^,^,^,^ " , ] the distributions of log(@xmath3/@xmath22 ) for absorbed and unabsorbed objects are indistinguishable ( 47.0 per cent probability to reject the null hypothesis ) . as pointed out before , the optical classification and x - ray absorption of our agn
are highly correlated , but the correspondence is not one - to - one .
based on the f - test probability at a 95 per cent confidence level , 37/137 ( 27 per cent ) type 1 agn are x - ray absorbed while 10/95 ( 10.5 per cent ) type 2 agn are x - ray unabsorbed .
if we adopt a column density threshold of @xmath23=@xmath24 to distinguish between x - ray unobscured and obscured agn , as in recent studies , the numbers change to 22/137 type 1 agn ( 16 per cent ) and 15/95 type 2 agn ( 15.8 per cent ) .
this effect smooths out the already small difference in the @xmath3/@xmath22 luminosity ratio distributions between type 1 and type 2 agn
. moreover , we find that the mid - ir / x - ray luminosity ratio is largely independent of the line - of - sight x - ray absorption ( linear correlation coefficient @[email protected] ) .
median values of log(@xmath27/@xmath22 ) are @xmath28 for @xmath29 , @xmath30 for @xmath31 and @xmath32 for @xmath33 ( for similar results see also @xcite ; @xcite ; @xcite ) .
this is an expected result . on the one hand , in x - rays we are measuring the amount of absorption along our line - of - sight to each agn .
clearly , the measured absorption values could change significantly if we were observing the objects along different lines - of - sight . on the other hand , at mid - ir wavelengths we are measuring the integrated reprocessed emission from all the material in the dusty torus in the plane of the sky .
if this material is absorbing the x - rays and it is re - emitting the radiation almost isotropically , we would expect to observe approximately the same log(@xmath3/@xmath34 ) value independently of orientation .
we have shown that the agn emission at rest - frame 6@xmath0 m is not free of dust extinction effects .
nevertheless , there is certainly a large overlap between the log(@xmath3/@xmath35 ) distributions for the type 1 and type 2 agn populations ( @xmath70.1 dex of shift ; see fig .
[ fig8 ] ) and the scatter in the luminosity ratio distributions is significant .
this makes orientation - dependent effects , such as a correlation with the line - of - sight x - ray absorption , difficult to detect , at least for agn with x - ray absorption in the compton - thin regime ( for similar results see also @xcite ; @xcite ) . to summarize ,
our results imply that the agn emission at rest - frame 6@xmath0 m is not completely isotropic , due to self - absorption in the dusty torus .
such interpretation is consistent with agn torus models with both smooth and clumpy dust distributions as they all predict anisotropic dust emission at rest - frame wavelengths of 6@xmath0m(e.g .
@xcite ; @xcite ; @xcite ) . this effect could be enhanced if , as suggested by recent results in the literature for local seyfert galaxies , the properties of the dusty tori in type 1 and type 2 objects are different ( e.g. @xcite ; @xcite ) . in either case ,
our results imply that agn surveys at rest - frame @xmath76@xmath0 m are subject to modest dust obscuration biases .
we investigate whether rest - frame 6@xmath0 m mid - ir continuum luminosities associated to the dusty torus emission of agn determined with data from the wide - field infrared survey explorer ( @xmath12 ) are a reliable and isotropic proxy of the agn intrinsic luminosity .
to do so we have studied the correspondence between the rest - frame 6@xmath0 m and intrinsic 2 - 10 kev continuum luminosities of agn using a complete , x - ray flux limited sample of 137 type 1 agn and 95 type 2 agn drawn from the bright ultra - hard xmm - newton survey .
the agn in this study have 2 - 10 kev x - ray luminosities between @xmath1 and @xmath2 , @xmath17 in the range 0.05 - 2.8 with a median @xmath17 of 0.558 and compton - thin x - ray absorption .
the x - ray luminosities were computed from a detailed analysis of the good quality _ xmm - newton _ spectra available for all objects and are corrected for absorption . to compute the mid - ir continuum emission associated with the agn dusty torus at rest - frame 6@xmath0 m we decomposed the rest - frame uv - to - mid - ir seds of our agn into agn and host galaxy components . to determine the existing relationship between @xmath36 and @xmath37 luminosities ,
we have conducted a linear regression analysis in log - log space using the bayesian maximum likelihood technique proposed by @xcite .
the main results of our study can be summarized as follows : 1 .
we find that at agn luminosities @xmath34@xmath11@xmath38 , the mid - ir emission detected with @xmath12 at rest - frame @xmath76@xmath0 m is dominated by hot dust heated by the agn .
the contamination from the agn host galaxies and the accretion disk at these wavelengths is not negligible , at a @xmath730 per cent level for agn with @xmath1@xmath9@xmath39@xmath9@xmath40 and at a @xmath712 - 18 per cent level for agn with @xmath41@xmath9@xmath22@xmath9@xmath18 .
we also found a relationship for the full agn sample that is consistent with being linear , @xmath3@xmath4@xmath42 , but has significant intrinsic scatter , @xmath6log@[email protected] . at a given x - ray luminosity , both the intrinsic scatter in 6@xmath0 m luminosities and the slopes of the linear regression relations
are compatible , within the uncertainties , for both type 1 and type 2 agn .
the main contributors to the intrinsic scatter in the @xmath36 versus @xmath37 correlation are likely the well known large dispersion in the individual sed shapes associated with the varying properties of the material responsible for the x - ray and mid - ir emission and agn variability .
if the agn dust emission at rest - frame wavelengths @xmath912@xmath0 m is not isotropic , as predicted by models of the agn dusty tori , such effect should also contribute to the measured intrinsic dispersion .
4 . assuming both that the 2 - 10 kev intrinsic luminosity is a reliable proxy for the agn bolometric radiative power and a constant x - ray bolometric correction , the fraction of agn bolometric luminosity that is reprocessed in the mid - ir decreases weakly , if at all , with the agn luminosity over the three decades in x - ray luminosity covered with our sample .
this finding is at odds with simple receding torus models .
the log(@xmath3/@xmath22 ) luminosity ratio is largely independent of the line - of - sight x - ray absorption .
if the material that absorbs the x - rays re - emits the radiation almost isotropically , we should observe approximately the same log(@xmath43/@xmath22 ) value regardless of orientation , yet the x - ray absorption along the line - of - sight to the agn could differ significantly .
we have shown that the agn emission at rest - frame 6@xmath0 m is not free of dust extinction effects .
nevertheless , taking into account the large overlap between the log(@xmath3/@xmath35 ) distributions for the type 1 and type 2 agn populations ( @xmath70.1 dex of shift ) and the significant scatter in the luminosity ratio distributions , any orientation - dependent effects , such as a correlation with the line - of - sight x - ray absorption , will certainly be difficult to detect , at least for agn with x - ray absorption in the compton - thin regime .
type 2 agn have on average redder mid - ir continua and , at a given x - ray luminosity , are @xmath71.3 - 2 times fainter at rest - frame 6@xmath0 m than type 1 agn , although with a considerable overlap between both populations . if the agn x - ray emission is weakly anisotropic , as recent results suggest , a correction for such effect would increase the differences in mid - ir luminosities at a given x - ray luminosity between type 1 and type 2 agn .
regardless of whether all agn have the same or different nuclear dusty toroidal structures , our results imply that the agn emission at rest - frame 6@xmath0 m is not completely isotropic due to self - absorption in the dusty torus .
thus our results imply that agn surveys at rest - frame @xmath76@xmath0 m are subject to modest dust obscuration biases .
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 . based on data from the wide - field infrared survey explorer , which is a joint project of the university of california , los angeles , and the jet propulsion laboratory / california institute of technology , funded by the national aeronautics and space administration .
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/. based on observations collected at the european organization for astronomical research in the southern hemisphere , chile , programme ids 084.a-0828 , 086.a-0612 , 087.a-0447 and 088.a-0628 .
based on observations made with the william herschel telescope and its service programme ( programme ids : sw2009a17 , sw2009b15 , sw2010b20)- operated by the isaac newton group , the telescopio nazionale galileo ( programme ids : tng12 - 09b , tng15 - 10a , tng11 - 10b , tng12 - 11a ) - operated by the centro galileo galilei and the gran telescopio de canarias ( programme ids : gtc38 - 09b , gtc32 - 10a , gtc18 - 10b , gtc44 - 11a , gtc41 - 12a , gtc25 - 13a ) installed in the spanish observatorio del roque de los muchachos of the instituto de astrofsica de canarias , in the island of la palma .
sm , fjc , xb , a.h .- c . and
acknowledge financial support by the spanish ministry of economy and competitiveness through grant aya2012 - 31447 .
sm , fjc and a.a .- h .
acknowledge financial support from the arches project ( 7th framework of the european union , no .
aa - h acknowledges support from the universidad de cantabria through the augusto g. linares program .
ac , rdc and ps acknowledge financial support from the italian ministry of education , universities and research ( prin2010 - 2011 , grant n. 2010nhbsbe ) and from asi ( grant n. i/088/06/0 ) . the authors wish to thank the anonymous referee for constructive comments . 200 abazajian k. n. , et al . , 2009 ,
apjs , 182 , 543 alonso - herrero a. , et al . , 2011 , apj , 736 , 82 alonso - herrero a. , et al . , 2006 , apj , 640 , 167 alonso - herrero a. , quillen a. c. , rieke g. h. , ivanov v. d. , efstathiou a. , 2003 , aj , 126 , 81 alonso - herrero a. , ward m. j. , kotilainen j. k. , 1996 , mnras , 278 , 902 antonucci r. , 1993 , ara&a , 31 , 473 arnaud k. a. , 1996 , aspc , 101 , 17 asmus d. , hnig s. f. , gandhi p. , smette a. , duschl w. j. , 2014 , mnras , 439 , 1648 asmus d. , gandhi p. , smette a. , hnig s. f. , duschl w. j. , 2011 , a&a , 536 , a36 assef r. j. , et al . , 2013 , apj , 772 , 26 assef r. j. , et al . , 2010 , apj , 713 , 970 ballo l. , severgnini p. , della ceca r. , caccianiga a. , vignali c. , carrera f. j. , corral a. , mateos s. , 2014 , mnras , 444 , 2580 barcons x. , franceschini a. , de zotti g. , danese l. , miyaji t. , 1995 , apj , 455 , 480 bianchi s. , bonilla n. f. , guainazzi m. , matt g. , ponti g. , 2009 , a&a , 501 , 915 bruzual g. , charlot s. , 2003 , mnras , 344 , 1000 caccianiga a. , et al .
, 2008 , a&a , 477 , 735 caccianiga a. , severgnini p. , della ceca r. , maccacaro t. , carrera f. j. , page m. j. , 2007 , a&a , 470 , 557 calderone g. , sbarrato t. , ghisellini g. , 2012 , mnras , 425 , l41 calzetti d. , armus l. , bohlin r. c. , kinney a. l. , koornneef j. , storchi - bergmann t. , 2000 , apj , 533 , 682 cardelli j. a. , clayton g. c. , mathis j. s. , 1989 , apj , 345 , 245 chabrier g. , 2003 , apj , 586 , l133 cohen m. , wheaton w. a. , megeath s. t. , 2003 , aj , 126 , 1090 corral a. , della ceca r. , caccianiga a. , severgnini p. , brunner h. , carrera f. j. , page m. j. , schwope a. d. , 2011 , a&a , 530 , a42 cutri r. m. , et al . , 2013 , technical report , explanatory supplement to the allwise data release products cutri r. m. , et al . , 2003 , the irsa 2mass all - sky point source catalog , nasa / ipac infrared science archive , available at : http://irsa.ipac.caltech.edu/applications/gator/ della ceca r. , et al . , 2008 , a&a , 487 , 119 diamond - stanic a. m. , rieke g. h. , rigby j. r. , 2009 , apj , 698 , 623 dicken d. , et al .
, 2014 , apj , 788 , 98 dickey j. m. , lockman f. j. , 1990 , ara&a , 28 , 215 elbaz d. , et al . , 2011 ,
a&a , 533 , a119 elvis m. , et al . , 1994 , apjs , 95 , 1 fanali r. , caccianiga a. , severgnini p. , della ceca r. , marchese e. , carrera f. j. , corral a. , mateos s. , 2013 , mnras , 433 , 648 feltre a. , hatziminaoglou e. , fritz j. , franceschini a. , 2012 , mnras , 426 , 120 fiore f. , et al . , 2009 , apj , 693 , 447 fritz j. , franceschini a. , hatziminaoglou e. , 2006 , mnras , 366 , 767 gandhi p. , horst h. , smette a. , hnig s. , comastri a. , gilli r. , vignali c. , duschl w. , 2009 , a&a , 502 , 457 gordon k. d. , clayton g. c. , 1998 , apj , 500 , 816 granato g. l. , danese l. , 1994 , mnras , 268 , 235 guainazzi m. , matt g. , perola g. c. , 2005 , a&a , 444 , 119 hernn - caballero a. , et al . , 2009 , mnras , 395 , 1695 hewett p. c. , warren s. j. , leggett s. k. , hodgkin s. t. , 2006 , mnras , 367 , 454 hnig s. f. , leipski c. , antonucci r. , haas m. , 2011 , apj , 736 , 26 hopkins p. f. , richards g. t. , hernquist l. , 2007 , apj , 654 , 731 hopkins p. f. , et al . , 2004 ,
aj , 128 , 1112 horst h. , gandhi p. , smette a. , duschl w. j. , 2008 , a&a , 479 , 389 ichikawa k. , ueda y. , terashima y. , oyabu s. , gandhi p. , matsuta k. , nakagawa t. , 2012 , apj , 754 , 45 isobe t. , feigelson e. d. , akritas m. g. , babu g. j. , 1990 , apj , 364 , 104 jarrett t. h. , chester t. , cutri r. , schneider s. , skrutskie m. , huchra j. p. , 2000 , aj , 119 , 2498 kelly b. c. , 2007 , apj , 665 , 1489 kennicutt r. c. , jr .
, 1998 , ara&a , 36 , 189 kotilainen j. k. , ward m. j. , boisson c. , depoy d. l. , smith m. g. , bryant l. r. , 1992 , mnras , 256 , 125 krabbe a. , bker t. , maiolino r. , 2001 , apj , 557 , 626 lacy m. , et al . , 2013 , apjs , 208 , 24 lamassa s. m. , heckman t. m. , ptak a. , martins l. , wild v. , sonnentrucker p. , 2010 ,
apj , 720 , 786 lawrence a. , et al . , 2007 , mnras , 379 , 1599 lawrence a. , 1991 , mnras , 252 , 586 levenson n. a. , radomski j. t. , packham c. , mason r. e. , schaefer j. j. , telesco c. m. , 2009 , apj , 703 , 390 liu t. , wang j .- x . , yang h. , zhu f .- f . ,
zhou y .- y . , 2014 , apj , 783 , 106 lusso e. , et al . , 2013 ,
apj , 777 , 86 lutz d. , maiolino r. , spoon h. w. w. , moorwood a. f. m. , 2004 , a&a , 418 , 465 mainieri v. , bergeron j. , hasinger g. , lehmann i. , rosati p. , schmidt m. , szokoly g. , della ceca r. , 2002 , a&a , 393 , 425 maiolino r. , shemmer o. , imanishi m. , netzer h. , oliva e. , lutz d. , sturm e. , 2007 , a&a , 468 , 979 maiolino r. , marconi a. , salvati m. , risaliti g. , severgnini p. , oliva e. , la franca f. , vanzi l. , 2001 , a&a , 365 , 28 marchese e. , della ceca r. , caccianiga a. , severgnini p. , corral a. , fanali r. , 2012 , a&a , 539 , a48 marconi a. , risaliti g. , gilli r. , hunt l. k. , maiolino r. , salvati m. , 2004 , mnras , 351 , 169 martnez - sansigre a. , rawlings s. , lacy m. , fadda d. , marleau f. r. , simpson c. , willott c. j. , jarvis m. j. , 2005 , natur , 436 , 666 mason r. e. , et al . , 2012 , aj , 144 , 11 mateos s. , alonso - herrero a. , carrera f. j. , blain a. , severgnini p. , caccianiga a. , ruiz a. , 2013 , mnras , 434 , 941 mateos s. , et al . , 2012 , mnras , 426 , 3271 mateos s. , et al . , 2010 ,
a&a , 510 , a35 mateos s. , et al . , 2008 ,
a&a , 492 , 51 mateos s. , barcons x. , carrera f. j. , ceballos m. t. , hasinger g. , lehmann i. , fabian a. c. , streblyanska a. , 2005b , a&a , 444 , 79 mateos s. , et al .
, 2005a , a&a , 433 , 855 matsuta k. , et al . , 2012 , apj , 753 , 104 mckernan b. , ford k. e. s. , chang n. , reynolds c. s. , 2009 , mnras , 394 , 491 merloni a. , et al . , 2014 , mnras , 437 , 3550 merloni a. , krding e. , heinz s. , markoff s. , di matteo t. , falcke h. , 2006 , newa , 11 , 567 mor r. , netzer h. , elitzur m. , 2009 , apj , 705 , 298 nenkova m. , sirocky m. m. , nikutta r. , ivezi . , elitzur m. , 2008b , apj , 685 , 160 nenkova m. , sirocky m. m. , ivezi .
, elitzur m. , 2008a , apj , 685 , 147 neugebauer g. , oke j. b. , becklin e. e. , matthews k. , 1979 , apj , 230 , 79 peeters e. , spoon h. w. w. , tielens a. g. g. m. , 2004 , apj , 613 , 986 piconcelli e. , jimenez - bailn e. , guainazzi m. , schartel n. , rodrguez - pascual p. m. , santos - lle m. , 2005 , a&a , 432 , 15 pier e. a. , krolik j. h. , 1992 , apj , 401 , 99 pineau f .- x . , motch c. , carrera f. , della ceca r. , derrire s. , michel l. , schwope a. , watson m. g. , 2011 , a&a , 527 , a126 porquet d. , reeves j. n. , obrien p. , brinkmann w. , 2004 , a&a , 422 , 85 press w. h. , teukolsky s. a. , vetterling w. t. , flannery b. p. , 1992
, nrfa.book , ramos almeida c. , et al . , 2011 , apj , 731 , 92 ramos almeida c. , prez garca a. m. , acosta - pulido j. a. , rodrguez espinosa j. m. , 2007 , aj , 134 , 2006 richards g. t. , et al . , 2006 , aj , 131 , 2766 roseboom i. g. , lawrence a. , elvis m. , petty s. , shen y. , hao h. , 2013 , mnras , 429 , 1494 rovilos e. , et al . , 2014 , mnras , 438 , 494 scott a. e. , stewart g. c. , mateos s. , alexander d. m. , hutton s. , ward m. j. , 2011 , mnras , 417 , 992 scott a. e. , stewart g. c. , mateos s. , 2012 , mnras , 423 , 2633 silva l. , maiolino r. , granato g. l. , 2004 , mnras , 355 , 973 simpson c. , 2005 , mnras , 360 , 565 stern d. , et al . , 2012 , apj , 753 , 30 tozzi p. , et al . , 2006 , a&a , 451 , 457 treister e. , krolik j. h. , dullemond c. , 2008 , apj , 679 , 140 turner t. j. , george i. m. , nandra k. , mushotzky r. f. , 1997 , apjs , 113 , 23 urry c. m. , padovani p. , 1995 ,
pasp , 107 , 803 vasudevan r. v. , fabian a. c. , 2007 , mnras , 381 , 1235 winter l. m. , veilleux s. , mckernan b. , kallman t. r. , 2012 , apj , 745 , 107 winter l. m. , mushotzky r. f. , reynolds c. s. , tueller j. , 2009 , apj , 690 , 1322 wright e. l. , eisenhardt p. r. m. , mainzer a. k. , et al .
2010 , aj , 140 , 1868 yan l. , et al . , 2013 , aj , 145 , 55 yang h. , wang j. , liu t. , 2014 , arxiv , arxiv:1411.4585
@lcccc + name & type & @xmath44 & @xmath45 + [ 0.5ex ] & & @xmath46 & @xmath46 + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) + 2xmmij000441.2 + 00071 & 1.9 & [email protected] & [email protected] + 2xmmij001130.3 + 00575 & 1 & [email protected] & [email protected] + 2xmmj002146.3 - 084711 & 1 & [email protected] & [email protected] + 2xmmij002244.5 + 00182 & 1 & [email protected] & [email protected] + 2xmmj003430.8 - 213351 & 1 & [email protected] & [email protected] + 2xmmj004003.8 + 005853 & 1 & [email protected] & [email protected] + 2xmmj004341.4 + 005610 & 1 & [email protected] & [email protected] + 2xmmj004404.6 + 010153 & 2 & [email protected] & [email protected] + 2xmmij011829.6 + 00454 & 1 & [email protected] & [email protected] + 2xmmj012447.7 + 320727 & 1 & [email protected] & [email protected] + 2xmmj013943.0 + 061254 & 1.5 & [email protected] & [email protected] + 2xmmj014804.3 + 055055 & 1 & [email protected] & [email protected] + 2xmmj020011.5 - 093125 & 1 & [email protected] & [email protected] + 2xmmj020112.5 - 092016 & 2 & [email protected] & [email protected] + 2xmmj022430.4 + 185315 & 2 & [email protected] & [email protected] + 2xmmj023234.3 - 073102 & 1 & [email protected] & [email protected] + 2xmmj024200.8 + 000021 & 1 & [email protected] & [email protected] + 2xmmj024325.4 - 000412 & 1.5 & [email protected] & [email protected] + 2xmmj030419.5 - 010911 & 2 & [email protected] & [email protected] + 2xmmj030742.6 - 000121 & 1.5 & [email protected] & [email protected] + 2xmmj033729.5 + 004227 & 1 & [email protected] & [email protected] + 2xmmij044710.5 - 06255 & 1 & [email protected] & [email protected] + 2xmmj073534.9 + 435414 & 1.9 & [email protected] & [email protected] + 2xmmij075734.4 + 39260 & 2 & [email protected] & [email protected] + 2xmmj080411.1 + 650906 & 2 & [email protected] & [email protected] + 2xmmj080608.0 + 244421 & 1 & [email protected] & [email protected] + 2xmmj081014.5 + 280337 & 1 & [email protected] & [email protected] + 2xmmj082042.4 + 205715 & 2 & [email protected] & [email protected] + 2xmmj082053.8 + 210735 & 2 & [email protected] & [email protected] + 2xmmj083049.6 + 524910 & 1 & [email protected] & [email protected] + 2xmmj083139.1 + 524205 & 2 & [email protected] & [email protected] + 2xmmij084012.7 + 51124 & 1 & [email protected] & [email protected] + 2xmmj084117.5 + 003448 & 2 & [email protected] & [email protected] + 2xmmj084927.7 + 445458 & 1 & [email protected] & [email protected] + 2xmmj085228.6 + 163008 & 2 & [email protected] & [email protected] + 2xmmj085835.2 + 275543 & 1 & [email protected] & [email protected] + 2xmmj085841.4 + 140944 & 1 & [email protected] & [email protected] + 2xmmj090053.8 + 385616 & 2 & [email protected] & [email protected] + 2xmmij091557.3 + 29261 & 1 & [email protected] & [email protected] + 2xmmij091624.3 + 29391 & 1 & [email protected] & [email protected] + 2xmmj091636.5 + 301749 & 2 & [email protected] & [email protected] + 2xmmj091645.4 + 514146 & 2 & [email protected] & [email protected] + 2xmmj092129.2 + 370103 & 1 & [email protected] & [email protected] + 2xmmj092201.2 + 301411 & 2 & [email protected] & [email protected] + 2xmmj092313.0 + 511742 & 1 & [email protected] & [email protected] + 2xmmij092619.6 + 36271 & 1 & [email protected] & [email protected] + 2xmmj093347.9 + 551846 & 1 & [email protected] & [email protected]
+ 2xmmj093458.2 + 611234 & 1.9 & [email protected] & [email protected] + 2xmmj094057.1 + 032401 & 1.5 & [email protected] & [email protected]
+ 2xmmj094350.2 + 035913 & 1 & [email protected] & [email protected] + 2xmmj094404.3 + 480647 & 1 & [email protected] & [email protected] + 2xmmj094439.8 + 034940 & 1 & [email protected] & [email protected] + 2xmmj094509.0 + 040817 & 2 & [email protected] & [email protected] + 2xmmij095630.7 - 00343 & 1 & [email protected] & [email protected] + 2xmmj095732.0 + 024302 & 1 & [email protected] & [email protected] + 2xmmj095815.5 + 014922 & 1 & [email protected] & [email protected]
+ 2xmmj095857.3 + 021314 & 1 & [email protected] & [email protected] + 2xmmj095908.3 + 024309 & 1 & [email protected] & [email protected] + 2xmmj095918.7 + 020951 & 1 & [email protected] & [email protected] + 2xmmj100015.3 + 013147 & 2 & [email protected] & [email protected] + 2xmmj100025.2 + 015852 & 1.5 & [email protected] & [email protected] + 2xmmj100032.1 + 553630 & 2 & [email protected] & [email protected] + 2xmmj100035.4 + 052428 & 1.5 & [email protected] & [email protected] + 2xmmj100057.4 + 684230 & 1 & [email protected] & [email protected] + 2xmmj100120.7 + 555351 & 1 & [email protected] & [email protected] + 2xmmj100129.3 + 013633 & 2 & [email protected] & [email protected] + 2xmmj100205.0 + 023731 & 1 & [email protected] & [email protected] + 2xmmj100237.8 + 024701 & 1.5 & [email protected] & [email protected] + 2xmmj101616.7 + 391143 & 1 & [email protected] & [email protected] + 2xmmij101733.1 - 00014 & 1.9 & [email protected] & [email protected] + 2xmmij101811.5 + 00101 & 1 & [email protected] & [email protected] + 2xmmij101830.7 + 00050 & 2 & [email protected] & [email protected] + 2xmmj101922.6 + 412050 & 1 & [email protected] & [email protected] + 2xmmj102147.4 + 130850 & 1 & [email protected] & [email protected] + 2xmmj102147.8 + 131227 & 1.8 & [email protected] & [email protected] + 2xmmj102551.1 + 384008 & 1.5 & [email protected] & [email protected] + 2xmmj103739.4 + 414149 & 1 & [email protected] & [email protected] + 2xmmj104048.4 + 061819 & 1 & [email protected] & [email protected] + 2xmmj104451.4 - 012226 & 2 & [email protected] & [email protected] + 2xmmj104522.0 - 012844 & 1 & [email protected] & [email protected]
+ 2xmmj104522.1 + 212614 & 1 & [email protected] & [email protected] + 2xmmj104912.6 + 330501 & 1.9 & [email protected] & [email protected] + 2xmmj105250.0 + 335505 & 1 & [email protected] & [email protected] + 2xmmj105932.0 + 242939 & 1 & [email protected] & [email protected] + 2xmmij111006.8 + 61252 & 1 & [email protected] & [email protected]
+ 2xmmj111121.2 + 482333 & 1 & [email protected] & [email protected] + 2xmmj111121.6 + 482047 & 1 & [email protected] & [email protected] + 2xmmj111135.6 + 482945 & 1 & [email protected] & [email protected] + 2xmmj111559.0 + 425321 & 1 & [email protected] & [email protected] + 2xmmj111606.9 + 423645 & 1 & [email protected] & [email protected] + 2xmmj111750.7 + 075710 & 1.5 & [email protected] & [email protected] + 2xmmj111832.4 + 130732 & 1 & [email protected] & [email protected] + 2xmmj111909.2 + 130950 & 2 & [email protected] & [email protected] + 2xmmj112026.6 + 431519 & 2 & [email protected] & [email protected] + 2xmmj112328.0 + 052823 & 1 & [email protected] & [email protected] + 2xmmj112338.0 + 052038 & 1 & [email protected] & [email protected] + 2xmmj113121.7 + 310254 & 1.9 & [email protected] & [email protected]
+ 2xmmj113129.2 + 310944 & 1 & [email protected] & [email protected] + 2xmmij114654.6 + 20254 & 1 & [email protected] & [email protected] + 2xmmj115754.9 + 434753 & 2 & [email protected] & [email protected] + 2xmmj120518.6 + 443926 & 1.9 & [email protected] & [email protected] + 2xmmj120529.5 + 442106 & 1 & [email protected] & [email protected] + 2xmmj120952.6 + 393143 & 1 & [email protected] & [email protected]
+ 2xmmj121118.8 + 503653 & 1.9 & [email protected] & [email protected] + 2xmmj121122.4 + 130936 & 1 & [email protected] & [email protected] + 2xmmj121356.1 + 140431 & 1 & [email protected] & [email protected] + 2xmmj121422.9 + 024252 & 1 & [email protected] & [email protected] + 2xmmj121509.4 + 135450 & 1 & [email protected] & [email protected] + 2xmmj121732.7 + 465829 & 1 & [email protected] & [email protected] + 2xmmj121808.5 + 471613 & 1 & [email protected] & [email protected]
+ 2xmmj121839.4 + 470627 & 2 & [email protected] & [email protected] + 2xmmj121920.6 + 470323 & 2 & [email protected] & [email protected] + 2xmmj121930.9 + 064334 & 1 & [email protected] & [email protected]
+ 2xmmj121952.2 + 472058 & 1 & [email protected] & [email protected] + 2xmmj122132.3 + 043557 & 1 & [email protected] & [email protected] + 2xmmj122137.8 + 043025 & 1.5 & [email protected] & [email protected] + 2xmmj122330.7 + 154507 & 1 & [email protected] & [email protected] + 2xmmj122532.4 + 332532 & 1 & [email protected] & [email protected] + 2xmmj122649.5 + 311735 & 1.9 & [email protected] & [email protected] + 2xmmj122656.4 + 013124 & 1.8 & [email protected] & [email protected] + 2xmmj123149.0 + 214749 & 2 & [email protected] & [email protected]
+ 2xmmj123204.9 + 215254 & 2 & [email protected] & [email protected] + 2xmmj123305.8 + 001438 & 1 & [email protected] & [email protected] + 2xmmj123356.1 + 074755 & 1.5 & [email protected] & [email protected] + 2xmmj123412.7 + 372734 & 2 & [email protected] & [email protected] + 2xmmij123602.1 + 26182 & 2 & [email protected] & [email protected] + 2xmmij123604.0 + 26413 & 1 & [email protected] & [email protected] + 2xmmj123625.4 + 125844 & 2 & [email protected] & [email protected] + 2xmmj123725.2 + 114158 & 1 & [email protected] & [email protected] + 2xmmj123759.5 + 621102 & 1 & [email protected] & [email protected] + 2xmmj124127.1 + 331201 & 2 & [email protected] & [email protected] + 2xmmj124135.8 + 332624 & 2 & [email protected] & [email protected] + 2xmmj124205.1 + 323634 & 2 & [email protected] & [email protected] + 2xmmj124213.8 - 112510 & 1 & [email protected] & [email protected] + 2xmmj124301.0 + 131217 & 2 & [email protected] & [email protected] + 2xmmj124408.9 + 113334 & 1.5 & [email protected] & [email protected] + 2xmmj124540.9 - 002744 & 1 & [email protected] & [email protected] + 2xmmj125304.6 + 101239 & 2 & [email protected] & [email protected] + 2xmmj125357.1 + 154314 & 2 & [email protected] & [email protected] + 2xmmj125414.5 + 101605 & 2 & [email protected] & [email protected] + 2xmmj125453.1 + 272008 & 2 & [email protected] & [email protected] + 2xmmj125553.0 + 272405 & 1 & [email protected] & [email protected] + 2xmmj125610.4 + 260103 & 1 & [email protected] & [email protected] + 2xmmj130237.6 - 024055 & 2 & [email protected] & [email protected] + 2xmmj130619.0 + 672421 & 2 & [email protected] & [email protected] + 2xmmij130906.2 + 11330 & 2 & [email protected] & [email protected] + 2xmmj130936.2 + 082815 & 1 & [email protected] & [email protected] + 2xmmj131046.7 + 271645 & 1 & [email protected] & [email protected] + 2xmmij131213.6 + 23195 & 1 & [email protected] & [email protected] + 2xmmj132037.8 + 341126 & 1 & [email protected] & [email protected] + 2xmmj132101.4 + 340658 & 1 & [email protected] & [email protected] + 2xmmj132105.4 + 341500 & 1 & [email protected] & [email protected] + 2xmmj132349.6 + 654148 & 1 & [email protected] & [email protected] + 2xmmj132419.0 + 300042 & 2 & [email protected] & [email protected] + 2xmmj132447.4 + 300900 & 2 & [email protected] & [email protected] + 2xmmj132826.3 + 583420 & 2 & [email protected] & [email protected] + 2xmmj132958.6 + 242435 & 1 & [email protected] & [email protected] + 2xmmj133120.3 + 242304 & 2 & [email protected] & [email protected] + 2xmmj133614.8 + 520224 & 2 & [email protected] & [email protected] + 2xmmj133811.2 + 515736 & 2 & [email protected] & [email protected] + 2xmmj134044.5 - 004516 & 1 & [email protected] & [email protected] + 2xmmj134113.9 - 005314 & 1 & [email protected] & [email protected] + 2xmmj134133.0 - 004033 & 2 & [email protected] & [email protected] + 2xmmj134133.1 + 353252 & 1 & [email protected] & [email protected] + 2xmmj134245.8 + 403913 & 2 & [email protected] & [email protected] + 2xmmj134252.9 + 403202 & 1 & [email protected] & [email protected] + 2xmmj134256.5 + 000057 & 1 & [email protected] & [email protected] + 2xmmj134323.6 + 001223 & 1 & [email protected] & [email protected] + 2xmmj134511.9 + 554759 & 1 & [email protected] & [email protected] + 2xmmj134656.6 + 580316 & 2 & [email protected] & [email protected] + 2xmmj134749.8 + 582109 & 1 & [email protected] & [email protected] + 2xmmj134803.8 - 035619 & 2 & [email protected] & [email protected] + 2xmmij135436.3 + 05152 & 2 & [email protected] & [email protected] + 2xmmij135453.5 + 05071 & 2 & [email protected] & [email protected] + 2xmmj135527.6 + 181639 & 2 & [email protected] & [email protected] + 2xmmj135628.7 + 052144 & 2 & [email protected] & [email protected] + 2xmmj140127.6 + 025606 & 1 & [email protected] & [email protected] + 2xmmj140145.0 + 025332 & 2 & [email protected] & [email protected] + 2xmmj140248.0 + 541350 & 2 & [email protected] & [email protected] + 2xmmj140353.8 + 540939 & 1.9 & [email protected] & [email protected] + 2xmmj140515.4 + 542459 & 1.9 & [email protected] & [email protected] + 2xmmj140614.2 + 282340 & 2 & [email protected] & [email protected] + 2xmmj140716.7 + 281653 & 2 & [email protected] & [email protected] + 2xmmj140719.3 + 281814 & 1 & [email protected] & [email protected] + 2xmmj140745.3 + 283028 & 1 & [email protected] & [email protected] + 2xmmj140921.1 + 261337 & 1 & [email protected] & [email protected] + 2xmmj141449.5 + 361239 & 1.5 & [email protected] & [email protected] + 2xmmj141512.6 + 360813 & 1.9 & [email protected] & [email protected] + 2xmmj141531.4 + 113156 & 1 & [email protected] & [email protected] + 2xmmj141622.4 + 265631 & 2 & [email protected] & [email protected] + 2xmmj141731.0 + 265622 & 2 & [email protected] & [email protected] + 2xmmij142656.4 + 60290 & 1 & [email protected] & [email protected] + 2xmmj142759.5 + 262150 & 1 & [email protected] & [email protected] + 2xmmj143025.8 + 415957 & 1 & [email protected] & [email protected] + 2xmmj143623.8 + 631726 & 2 & [email protected] & [email protected] + 2xmmj144336.1 + 062802 & 1 & [email protected] & [email protected] + 2xmmj144404.5 + 291412 & 1 & [email protected] & [email protected]
+ 2xmmj144411.3 + 291508 & 1 & [email protected] & [email protected] + 2xmmj144545.5 + 292312 & 2 & [email protected] & [email protected] + 2xmmj144936.5 + 090829 & 1 & [email protected] & [email protected] + 2xmmj145426.6 + 182956 & 1 & [email protected] & [email protected] + 2xmmj145442.2 + 182937 & 2 & [email protected] & [email protected] + 2xmmj145459.4 + 184452 & 1 & [email protected] & [email protected] + 2xmmj145717.5 + 223332 & 1 & [email protected] & [email protected]
+ 2xmmj150431.2 + 474151 & 1 & [email protected] & [email protected] + 2xmmj150558.3 + 014104 & 2 & [email protected] & [email protected] + 2xmmij150743.3 + 01132 & 1 & [email protected] & [email protected] + 2xmmj150930.9 + 565454 & 1 & [email protected] & [email protected] + 2xmmj151536.7 + 000347 & 2 & [email protected] & [email protected] + 2xmmj151703.6 + 562338 & 1.5 & [email protected] & [email protected]
+ 2xmmj152721.7 + 360016 & 2 & [email protected] & [email protected] + 2xmmj153202.2 + 301629 & 2 & [email protected] & [email protected] + 2xmmj153228.8 + 045358 & 1.5 & [email protected] & [email protected]
+ 2xmmj153304.0 + 302508 & 1.8 & [email protected] & [email protected] + 2xmmj154424.1 + 535546 & 2 & [email protected] & [email protected] + 2xmmj154930.6 + 213422 & 2 & [email protected] & [email protected] + 2xmmij155850.7 + 02353 & 1 & [email protected] & [email protected] + 2xmmj163331.8 + 570520 & 1 & [email protected] & [email protected] + 2xmmj164119.4 + 385407 & 1 & [email protected] & [email protected] + 2xmmj164203.4 + 385300 & 1 & [email protected] & [email protected] + 2xmmij164723.9 + 27054 & 2 & [email protected] & [email protected] + 2xmmij173104.7 + 36580 & 1.9 & [email protected] & [email protected] + 2xmmj200812.3 - 111658 & 2 & [email protected] & [email protected] + 2xmmj204043.2 - 004548 & 2 & [email protected] & [email protected] + 2xmmj204956.3 - 053457 & 2 & [email protected] & [email protected] + 2xmmj205038.3 - 053134 & 2 & [email protected] & [email protected] + 2xmmj211516.5 + 060840 & 1 & [email protected] & [email protected] + 2xmmj212932.8 + 001044 & 1.9 & [email protected] & [email protected] + 2xmmij215017.3 - 05571 & 2 & [email protected] & [email protected] + 2xmmij215138.7 - 05224 & 1 & [email protected] & [email protected] + 2xmmj221813.9 + 001624 & 2 & [email protected] & [email protected] + 2xmmij232807.6 + 14420 & 1 & [email protected] & [email protected] + @xmath47 .
the targets are listed in order of increasing right ascention .
column 1 : x - ray source name as listed in the second incremental version of the second xmm - newton serendipitous source catalogue ( 2xmm - dr3 ; http://xmmssc-www.star.le.ac.uk/catalogue/xcat_public_2xmmi-dr3.html ) ; column 2 : optical spectroscopic classification .
@xmath48for objects with a galaxy uv / optical spectrum without emission lines ; column 3 : 2 - 10 kev x - ray luminosity and 1@xmath49 uncertainty ( in logarithmic units ) .
the luminosities have been corrected for x - ray absorption .
column 4 : rest - frame 6@xmath0 m agn luminosity associated to the dusty torus emission and 1@xmath49 uncertainty ( in logarithmic units ) .
the infrared luminosities have been corrected for both the accretion disk and host galaxy emission . | we have determined the relation between the agn luminosities at rest - frame 6@xmath0 m associated to the dusty torus emission and at 2 - 10 kev energies using a complete , x - ray flux limited sample of 232 agn drawn from the bright ultra - hard xmm - newton survey .
the objects have intrinsic x - ray luminosities between @xmath1 and @xmath2 and redshifts from 0.05 to 2.8 .
the rest - frame 6@xmath0 m luminosities were computed using data from the wide - field infrared survey explorer and are based on a spectral energy distribution decomposition into agn and galaxy emission . the best - fit relationship for the full sample
is consistent with being linear , @xmath3@xmath4@xmath5 , with intrinsic scatter , @xmath6log@[email protected] dex .
the @xmath8 luminosity ratio is largely independent on the line - of - sight x - ray absorption . assuming a constant x - ray bolometric correction , the fraction of agn bolometric luminosity reprocessed in the mid - ir decreases weakly , if at all , with the agn luminosity , a finding at odds with simple receding torus models .
type 2 agn have redder mid - ir continua at rest - frame wavelengths @xmath912@xmath0 m and are overall @xmath71.3 - 2 times fainter at 6 @xmath0 m than type 1 agn at a given x - ray luminosity . regardless of whether type 1 and type 2 agn have the same or different nuclear dusty toroidal structures ,
our results imply that the agn emission at rest - frame 6@xmath0 m is not isotropic due to self - absorption in the dusty torus , as predicted by agn torus models .
thus , agn surveys at rest - frame @xmath76@xmath0 m are subject to modest dust obscuration biases .
[ firstpage ] galaxies : active - quasars : general - infrared : galaxies |
the physical process by which a heavy quark - antiquark pair is created in a hadronic collision and then emerges bound as a colorless hadronic state has been the subject of intense discussion within the framework of qcd for many years .
the change of color configurations by soft gluon emission was first described in the color singlet model ( csm ) by baier and ruckl @xcite in 1983 .
subsequent cross section measurements at cdf i @xcite in 1994 showed that the predictions underestimated the data by factors of 10 - 50 . an innovative effective field theory approach ( nrqcd ) by bodwin , braaten , and lepage @xcite recast the problem in terms of a perturbative treatment at the quark level with long - range process - independent parameters to describe the hadronization .
the free parameters in the problem were set from the differential cross section data but result in a definite statement about the contributing matrix elements in the process .
therefore , the polarization state of the resulting vector meson is determined from the theory .
the prediction is that at large @xmath9 the quarkonium state should be transversely polarized .
+ the prediction refers to quarkonium states that are directly produced in the hard collision .
experimentally , one measures quarkonium states that are _ prompt _ , i.e. , that come from the primary production vertex .
therefore , the experimental data include not only directly - produced quarkonium states but also those that come from the decay of higher - lying @xmath10 excitations , either s - wave or p - wave , a process known as _ feed - down decays_. these extra events will almost certainly have different polarizations than the directly produced states .
this presents a fundamental difficulty in interpreting the results from all experiments thus far .
some theoretical calculations attempt to estimate the feed - down depolarization effects , but there are large uncertainties in these estimates .
+ recent phenomenological approaches have taken different viewpoints than the nrqcd approach to describe the quarkonium production process . in the @xmath11 factorization approach of baranov and zotov @xcite , the color - singlet process dominates , and
the heavy quark pair is isolated from the rest of the interaction by a minimum momentum transfer from each parent hadron .
they predict a rather featureless @xmath9 dependence that should be _ longitudinal _ for direct production , as will be seen later .
in contrast , the higher - order csm approach of artoisenet , _ et al . _
@xcite considers multiple - gluon interactions and finds that the higher - order processes contribute coherently to the production amplitude and so build up the cross section to a point that generally agrees with current quarkonium observations .
this is a dramatic shift and has reinvigorated interest in this model .
+ in all three cases there are unique predictions for the polarization expected in direct production of quarkonium .
this makes these new tevatron results on @xmath1 polarization and the expected new lhc results on quarkonium polarization especially important .
following the success of nrqcd predictions for the j/@xmath0 and @xmath0(2s ) differential cross section results , using the freedom available in matching the long- distance parameters of the model , cdf set out to improve the measurements of the prompt polarization for these charmonium states using the larger event samples of tevatron run ii and the improved measurement accuracy of the cdf ii detector . the prompt polarization for both j/@xmath0 and @xmath0(2s ) does not increase as @xmath9 becomes significantly larger than the mass of the state , i.e. , in the perturbative regime .
the results for the j/@xmath0 are shown in fig .
[ jpsipol ] .
the measured polarization is small at all @xmath9 values , in contrast with nrqcd expectations .
the @xmath11 prediction for direct polarization has a large magnitude , but the shape matches experiment .
depolarizing effects from feed - down might reduce the discrepancy in size . more work is needed in this area .
these results are the state of the art at this moment .
new results are expected from lhc experiments soon .
+ polarization parameter @xmath12 in the @xmath6-channel helicity frame as a function of the dimuon @xmath9 .
theory curves for nrqcd and @xmath11 factorization are included for comparison .
the nrqcd band includes estimates of feed - down depolarization effects.[jpsipol],width=245 ] the charm quark mass is close to @xmath13 , and there is some speculation that disagreements with nrqcd predictions might imply that the charmonium system is too light for nrqcd assumptions to be valid .
that is certainly not the case with the bottomonium system , which is the focus of the rest of this report .
another issue whose importance has been emphasized in recent years is the choice of quantization axis along which to report polarization . in fixed target
experiments the collins - soper ( cs ) axis @xcite , the bisector of the angle between the colliding hadrons in the rest frame of the dimuon system , was the axis of choice . in hadron
colliders a natural axis was the @xmath6-channel helicity axis ( sh ) , defined as the direction of the lorentz boost that transforms the dimuon from the lab system to its rest frame . in elementary quantum mechanics we learn that a fully - polarized electron will have no spin alignment along an axis perpendicular to its quantization axis .
et al . _
@xcite emphasized the need to analyze experiments in several different lorentz frames to ensure that a small polarization was not an accident of frame choice .
more recently faccioli _
et al . _
@xcite have emphasized the importance of measuring the three independent parameters that describe the parity - conserving decay of a vector meson into a pair of spin-1/2 fermions in the vector meson rest frame : @xmath14\end{aligned}\ ] ] the new results reported here will follow the admonitions from these papers and constitute the first hadron collider results that measure all three polarization parameters in each of two lorentz frames : cs and sh .
in addition ref .
@xcite pointed out a frame - independent quantity that serves both as a systematic constraint on the measurements in two or more lorentz frames and also , should it be small , indicates that there is no significant polarization in any lorentz frame .
we shall also discuss that quantity for the new cdf results .
+ just as in the charmonium system , all three production models ( nrqcd , @xmath11 , and higher - order csm ) contain enough free parameters to match the @xmath15 cross section data well over the whole measured @xmath9 range .
the polarization measurement once again stands as the essential test of validity for the model results .
prior to this new cdf run ii measurement , there were results from cdf run i @xcite and d0 run ii @xcite , both in the sh frame .
the two experiments do not agree with each other , and neither experiment matches theory , as can be seen in fig .
[ ypol1 ] .
the experimental disagreement is not unique in polarization measurements .
these are differential measurements , and good control of acceptance and efficiency is crucial to obtaining stable results .
this experimental fact of life makes using the frame - independent test an important validation of new results .
one major difficulty with the @xmath1 measurements is handling the dimuon background , which is much larger than in the j/@xmath0 case .
the cdf and d0 background problems are very different , as can be seen in figs .
[ cdfmass ] and fig .
[ d0mass ] .
while it is clear from the preceding two pictures that the background fraction under any @xmath1 peak is smaller for cdf than for d0 , the importance of that fact really comes from the cdf observation that the combinatoric background has an angular variation that changes dramatically with dimuon mass and @xmath9 . at the tevatron the dominant dimuon background comes from heavy
flavor decays .
cdf defines a displaced dimuon sample by selecting dimuon - triggered events in which at least one muon has a miss - distance with respect to the primary vertex that exceeds 150 @xmath16 m .
this selection defines a sample dominated by combinatoric dimuons , with a tiny admixture of mis - measured @xmath1 events .
we divide the dimuon mass region into twelve intervals as shown in fig .
[ cdfmass ] : three regions containing @xmath1 signals plus background and nine pure background regions .
we make the fundamental assumption that for the combinatoric dimuons the angular distribution is independent of the apparent decay length , @xmath17 , the uncorrelated dimuons can appear to have a vertex at a distance that is uncorrelated with their angular characteristics in their rest frame
. if this hypothesis is true , then the angular characteristics of dimuons in the pure background mass slices will be the same for the prompt and displaced samples .
we have tested this hypothesis in all nine background bins and obtained excellent agreement .
we show as examples the normalized distributions in cos @xmath18 and @xmath19 for two such bins , one at low mass , one at high mass , in fig .
we make these comparisons for all three parameters in both cs and sh frames and see comparably good matches .
this agreement shows that we can extract the angular character of the combinatoric background under each of the three prompt @xmath1 signal regions by using the measured displaced sample in that same mass region .
no sideband extrapolation is needed . and
azimuth @xmath20 for dimuons from the prompt dimuon sample ( solid histogram ) and the detached dimuon sample ( points ) for low mass and high mass regions that are pure background .
the agreement in shape validates the claim that the angular characteristics of the combinatoric background are independent of the apparent vertex displacement .
[ both],width=245 ] to establish the normalization for the displaced sample , we plot the ratio of the two dimuon samples as a function of dimuon mass , shown in fig .
[ ratio ] .
the ratio is essentially constant in the nine background - only samples .
we fit the ratio to a linear function of dimuon mass and use that fit to scale the displaced sample angular distribution to predict the background prompt angular distribution in the three signal bands .
we fit the background angular distributions to the same functional form as the vector meson signal but do not limit the three coefficients for the background to the range expected for a real vector meson .
we make a simultaneous fit to the displaced angular distribution and the prompt angular distribution in each of the twelve dimuon mass ranges .
the prompt angular distribution contains a signal term ( if appropriate ) and an extrapolated background term , using the ratio as shown above .
all fits give good @xmath21 results in both cs and sh frames .
we also included a @xmath22 background term .
this term can not contribute for a real vector meson decay and it plays no role in the polarization analysis .
+ an example of the polarization fits is shown in fig .
[ polfit ] for the lowest @xmath9 bin .
the twelve dimuon background regions are each represented by a point in the two - dimensional parameter space in @xmath23 , along with the three signal fits .
error ellipses for 1@xmath24 and 2@xmath24 uncertainties are shown for each point . as can be seen from the plot , the background polarization parameter @xmath25 changes non - uniformly with dimuon mass in this @xmath9 region .
thus , having an accurate representation of background angular behavior from the data itself is essential to extracting the correct signal polarization .
this capability is unique to this analysis .
+ to establish systematic consistency for the measurement , one can use the frame - independent quantity + @xmath26 = @xmath27 defined in ref.@xcite to compare results from the cs and sh frames . because the same data are used in each frame ,
one has to evaluate matching likelihoods by using a toy monte carlo procedure . in all @xmath9
bins the results from the two frames agree well , based on the monte carlo expectations .
the complete results from this analysis , including tables of all fitted parameters in the cs and sh frames along with the frame - independent parameters are published in the supplemental material to ref .
we give illustrative examples here , starting with those for @xmath15 prompt decays , shown in fig . [ y1pol ] .
the @xmath15 polarization parameters are all small and show little variation with @xmath9 in either the cs or sh frames .
the values of the frame independent parameters are also small and @xmath9-independent , confirming that there is no lorentz frame in which one would see large polarization for this state .
+ to avoid having to interpret depolarizing effects of feed - down to @xmath15 , we can look at the @xmath8 parameters .
there is much less feed - down to this state because of smaller branching fractions from the @xmath28 states and because the higher - lying @xmath1 states have very tiny branching fractions to @xmath8 .
this is the first measurement of polarization parameters for hadroproduction of the @xmath8 state .
the sample size is smaller than that for the @xmath15 , but the results shown in fig .
[ y3pol ] give no indication of significant polarization at high @xmath9 in either the cs or sh frames .
the new cdf results for the @xmath15 agree well with the cdf run i measurement and disagree strongly with the d0 measurements , as shown in fig .
[ polcomp ] .
the situation for the polarization parameter @xmath25 , also known as @xmath12 , for the sh frame is summarized in fig .
[ polcomp ] .
there is no indication of a trend toward nrqcd behavior , illustrated by the yellow band . shown on the plot
are two curves for the @xmath11 model including feed - down from ref @xcite .
the solid line assumes spin coherence between y states , while the dashed line shows the case when the feed - down @xmath15 is depolarized .
the data seem to favor depolarization .
a direct measurement of @xmath15 polarization from @xmath29 parents is clearly quite important for understanding the situation .
not shown here but included in ref .
@xcite is a polarization band for the higher - order csm that has the same general shape as the data and is consistent with these results .
the width of the uncertainty band can be reduced by having more experimental information about feed - down contributions .
these new cdf results represent the first hadron collider determination of the complete @xmath7 and @xmath8 polarization parameter set in both collins - soper and @xmath6-channel helicity frames for promptly - produced bottomonium decays to @xmath2 .
the results show that all polarization parameters are small and have little variation with dimuon @xmath9 .
the trend of the data tends to support the @xmath11 model and the higher - order csm model but not nrqcd expectations .
we await the upcoming analyses from lhc to see the trends at higher energies and higher @xmath9 in pp collisions .
we thank the fermilab staff and the technical staffs of the participating institutions for their vital contributions .
this work was supported by the u.s .
department of energy and national science foundation ; the italian istituto nazionale di fisica nucleare ; the ministry of education , culture , sports , science and technology of japan ; the natural sciences and engineering research council of canada ; the national science council of the republic of china ; the swiss national science foundation ; the a.p .
sloan foundation ; the bundesministerium fr bildung und forschung , germany ; the korean world class university program , the national research foundation of korea ; the science and technology facilities council and the royal society , uk ; the russian foundation for basic research ; the ministerio de ciencia e innovacin , and programa consolider - ingenio 2010 , spain ; the slovak r&d agency ; the academy of finland ; and the australian research council ( arc ) .
baranov , s.p .
zotov , upsilonium polarization as a touchstone in understanding the proton dynamics in qcd " , jetp lett .
86 , 435 - 438 , 2007 .
artoisenet , p. , j.m .
campbell , j.p .
lansberg , f. maltoni and f. tramontano , @xmath31 production at fermilab tevatron and lhc energies " , phys .
101 , 152001 , 2008 .
abazov , v.m . , _
_ , measurement of the polarization of the @xmath34 and @xmath35 states in @xmath36 collisions at @xmath4 = 1.96-tev " , phys . rev
. lett . 101 , 182004 , 2008 .
aaltonen , t. , _ et al .
_ , measurements of angular distributions of muons from @xmath31 meson decays in @xmath32 collisions at @xmath37 tev " , phys .
108 , 151802 , 2012 . | production mechanisms for quarkonium states in hadronic collisions remain difficult to understand .
the decay angular distributions of j/@xmath0 or @xmath1 states into @xmath2 final states are sensitive to the matrix elements in the production process and provide a unique tool to evaluate different models .
this talk will focus on new results for the spin alignment of @xmath1 states produced in @xmath3 collisions at @xmath4 = 1.96 tev using the cdf ii detector at the fermilab tevatron .
the data sample corresponds to an integrated luminosity of 6.7 fb@xmath5 .
the angular distributions are analyzed as functions of the transverse momentum of the dimuon final state in both the collins - soper and the @xmath6-channel helicity frames using a unique data - driven background determination method .
consistency of the analysis is checked by comparing frame - invariant quantities derived from parametrizations of the angular distributions measured in each choice of reference frame .
this analysis is the first to quantify the complete three - dimensional angular distribution of @xmath7 and @xmath8 decays .
the decays are nearly isotropic in all frames , even when produced with large transverse momentum . |
the discovery of the higgs boson @xcite is a dramatic confirmation of the stan- dard model ( sm ) . despite finding last missing piece of sm we are still looking for insight into the detailed mechanism of electroweak symmetry breaking . in particular , within sm , the higgs mass receives large corrections leading to the hierarchy problem . for several resolutions of the hierarchy problem that have been put forward , like supersymmetry , large extra dimensions or composite structure the lhc found no hints so far .
given such a scenario where new physics , stemming from e. g. supersym- metry , seems to be absent , one is lead to pursue alternative approaches to the hierarchy problem . here , we follow an idea first put forward by bardeen @xcite that in classically scale invariant ( si ) theories , scale invariance
is broken by quantum corrections and quadratic divergence in fundamental scalar masses is a cutoff regularization artefact .
all dimensionfull parameters in these the- ories , including the scalar masses , come from a single renormalization scale . as has been shown by coleman and e. weinberg ( cw ) @xcite in order for this mechanism of mass generation to work within perturbation theory at least two bosonic degrees of freedom
are needed .
a seminal paper by gildener and s. weinberg @xcite ( gw ) represents a first cohesive study of the cw mechanism for multiple scalars .
classical scale invariance breaks by quantum anomaly yielding one pseudo- goldstone boson , the so - called scalon @xcite .
the intriguing idea of identifying the scalon with the higgs particle requires additional bosonic degrees of freedom due to the large top quark mass . sprung by initial explorations @xcite many works
have been put forward to realize this scenario . in order to compensate the large top contribution , extra bosons either have large couplings to the higgs @xcite or appear in large multiplets @xcite .
alternatively , the cw mechanism can be applied in a new gauge sector where the scale gets transmitted to the sm through the higgs portal @xcite . for additional work on the si extensions of sm in different contexts ,
see @xcite .
dark matter ( dm ) is another problem pressing us to extend the sm .
classically si extensions of the sm with dm have been studied in inert higgs doublet model @xcite , @xmath6 vector dark matter @xcite , multiple scalar models @xcite and non - perturbative realizations of the hidden sector @xcite .
si theories and the cw mechanism offer a possibility of a single origin of the electroweak symmetry breaking and the dark matter mass embodying the paradigm of the wimp miracle .
this attractive picture is the subject of our work .
we include si dark sector consisting of a new @xmath0 gauge group and @xmath1-charged scalar @xmath2 and majorana fermion @xmath3 .
the effective scalar potential is dominated by the contributions of the new gauge boson and leads to the breaking of the dark gauge symmetry . via the higgs portal coupling the cw mechanism
induces the breaking of electroweak symmetry . due to the residual @xmath4 symmetry
the majorana fermion is stable and represents the dm candidate .
we introduce @xmath0 gauge symmetry with doubly @xmath1-charged scalar @xmath2 and singly @xmath1-charged majorana fermion @xmath3 , both singlets under the sm gauge group .
the scalar potential and dark yukawa term are @xmath7 @xmath8 where @xmath9 is a sm higgs doublet .
the cw mechanism will be studied in the gw framework @xcite where quantum corrections are built on top of a flat direction in the tree - level potential giving mass to the scalon . assuming that both @xmath9 and @xmath2 acquire vacuum expectation values ( _ vevs _ ) @xmath10 the potential ( [ eq : pot ] ) has a flat direction when the couplings satisfy @xmath11 this is a dimensional transmutation , where one dimensionless coupling is traded for the physical scale of the model @xmath12 . at the tree level the _ vevs _
are given by @xmath13 scalar mass eigenstates are given by @xmath14 with mixing angle @xmath15 given by @xmath16 the tree - level masses of scalars are @xmath17 and @xmath1-gauge boson and fermion masses are @xmath18 where @xmath19 is the gauge coupling of the @xmath0 gauge symmetry . to find the scalon mass we need the one - loop corrected potential along the flat direction which takes the following form @xcite @xmath20 where the radial field @xmath21 is defined through @xmath22 while the fields @xmath23 satisfy the constraint @xmath24 , with their _ vevs _ being @xmath25 and @xmath26 , respectively .
the coefficients @xmath27 and @xmath28 are given as @xcite @xmath29 @xmath30 where @xmath31 is the _ vev _ of the field @xmath21 .
the mass of the scalon is given as @xmath32 the tree - level potential ( [ eq : pot ] ) in the mass eigenstate basis reads @xmath33 notice that the couplings @xmath34 , @xmath35 , @xmath36 and @xmath37 vanish at the tree level @xcite , independent of the particular value of the higgs portal coupling @xmath38 . in particular , the rigorous gw treatment for multiple scalar fields states that in the model studied here there are no higgs decays to the scalons on the tree level @xcite , contrary to @xcite .
these points can be understood from the fact that flat direction of the tree - level potential defined by the _ vev _ of the unit vector field @xmath39 is also an eigenvector of the mass matrix with zero eigenvalue .
therefore , if a term in the tree - level potential has more than two physical @xmath40 fields , or two @xmath40 fields and a dimensionfull coupling , the respective coefficient will be zero by construction .
due to the remnant @xmath4 symmetry the majorana fermion @xmath3 is a dm candidate @xcite .
our dm scenario assumes that @xmath3 is heavier than the scalar @xmath40 . in the early universe dm annihilates dominantly through the @xmath41-channel to @xmath42 pair , while contributions suppressed by a small mixing arise from @xmath43 and @xmath44 processes . at the time of the decoupling of the dm ,
scalar @xmath40 is in the thermal equilibrium with the thermal bath of the sm particles through the higgs portal interaction . while our numerical analysis covers all the mentioned annihilation processes , it is instructive to show the dominant one .
the thermally averaged @xmath45-wave annihilation cross section for @xmath46 is @xmath47 in our calculations we use the observed value for the dm relic density of the universe , @xmath48 @xcite .
the model introduces four new parameters : gauge coupling @xmath19 , yukawa coupling @xmath49 and quartic scalar couplings @xmath38 and @xmath50 .
the two constraints given by the higgs boson mass and the dark matter relic abundance leave two undetermined parameters chosen to be @xmath51 and @xmath52 .
the mass of the @xmath1-boson has to be larger than @xmath53 gev to overcome the top quark contribution in order to have positive mass for the scalon .
higgs searches at lep exclude scalon masses under 114 gev in our model .
global fit to the current lhc data @xcite gives an upper bound of @xmath54 .
we scan the parameter space of the model from the initial values @xmath55 gev , @xmath56 gev , up to @xmath57 gev , @xmath58 gev .
( left ) and dark sector gauge coupling @xmath19 ( right ) as a function of dark gauge boson mass @xmath51 and the scalon mass @xmath52 . the lower region on the plot is excluded by the lhc bound on @xmath25.,title="fig : " ] ( left ) and dark sector gauge coupling @xmath19 ( right ) as a function of dark gauge boson mass @xmath51 and the scalon mass @xmath52
. the lower region on the plot is excluded by the lhc bound on @xmath25.,title="fig : " ] ( left ) and the _ vev _ of @xmath2 , @xmath59 ( right ) as a function of dark gauge boson mass @xmath51 and the scalon mass @xmath52 .
the lower region on the plot is excluded by the lhc bound on @xmath25.,title="fig : " ] ( left ) and the _ vev _ of @xmath2 , @xmath59 ( right ) as a function of dark gauge boson mass @xmath51 and the scalon mass @xmath52
. the lower region on the plot is excluded by the lhc bound on @xmath25.,title="fig : " ] we give our results as prediction for various quantities in the @xmath60 plane . on fig .
[ fig : gm ] we give mass of the dm candidate @xmath61 and dark sector gauge coupling @xmath19 . on fig .
[ fig : sin_vphi ] we give the mixing angle @xmath25 and vev of @xmath2 , @xmath59 . holding @xmath52 fixed while increasing @xmath51 increases the _ vev _
@xmath62 naturally leading to a smaller mixing angle @xmath15 .
from ( [ eq : dmcs ] ) we see that the scale @xmath59 dominates the dm annihilation cross section @xmath63 therefore , in order to keep @xmath64 fixed to its value provided by the dm relic abundance , @xmath49 increases when @xmath59 increases .
the cw mechanism then requires that @xmath19 is increased to have a positive scalon mass .
notice the key role played by the dm constraint ; keeping @xmath52 fixed and increasing @xmath51 leads to an increase of @xmath19 .
this is in contrast to the model without dm where one naively expects @xmath65 . in other words , in the regime of large
@xmath59 the majorana fermion @xmath3 starts to play an important role in the cw mechanism .
for the minimal value @xmath66 gev , the dm constraint requires a relatively large yukawa coupling .
the cw mechanism pulls @xmath19 in the same direction , which is the origin of @xmath19 being of order one .
the perturbative restrictions provide upper bounds on masses in the hidden sector .
we find the mass of the dm to be in the range from @xmath67 gev up to a few tev .
postulating a hidden @xmath0 sector consisting of a doubly @xmath1-charged scalar and a singly charged majorana fermion we have examined a si extension of the sm with majorana fermion as a dm candidate . using the gw approach we have calculated masses of the new particles .
the majorana fermion saturates the dm relic abundance via the @xmath68 annihilation .
we predict the dm particle mass to be in the range from 500 gev to a few tev . while the lower bound
is obtained by the lhc limit on @xmath25 , the upper bound is a conservative estimate set by moderate values of the dark gauge and yukawa couplings .
the dm constraint plays an essential role here as it requires an appreciable yukawa coupling . in this sense
similar results are obtained by @xcite where dm constraint enforces a gauge coupling of order one .
a general statement may be drawn for a si extension of sm with a new gauge group ; the couplings in the hidden sector can be taken to be small provided the _ vev _ of the hidden scalar is large enough . however , identifying possible stable states of the hidden sector with dm particles , the scale set by cw mechanism requires moderate couplings in the hidden sector .
g. aad _ et al . _ [ atlas collaboration ] , _ observation of a new particle in the search for the standard model higgs boson with the atlas detector at the lhc _ ,
b * 716 * ( 2012 ) 1 [ arxiv:1207.7214 [ hep - ex ] ] . r. foot , a. kobakhidze , k. l. mcdonald and r. r. volkas , _ a solution to the hierarchy problem from an almost decoupled hidden sector within a classically scale invariant theory _ , phys .
d * 77 * ( 2008 ) 035006 [ arxiv:0709.2750 [ hep - ph ] ] .
r. foot , a. kobakhidze and r. r. volkas , _ stable mass hierarchies and dark matter from hidden sectors in the scale - invariant standard model _ , phys .
d * 82 * ( 2010 ) 035005 [ arxiv:1006.0131 [ hep - ph ] ] .
e. gabrielli , m. heikinheimo , k. kannike , a. racioppi , m. raidal and c. spethmann , _ towards completing the standard model : vacuum stability , ewsb and dark matter _ ,
d * 89 * ( 2014 ) 015017 [ arxiv:1309.6632 [ hep - ph ] ] . | we propose a classically scale invariant extension of the standard model where the electroweak symmetry breaking and the mass of the dark matter particle come from the common scale .
we introduce @xmath0 gauge symmetry and @xmath1-charged scalar @xmath2 and majorana fermion @xmath3 .
scale invariance is broken via coleman - weinberg mechanism pro- viding the vacuum expectation value of the scalar @xmath2 .
stability of the dark matter candidate @xmath3 is guaranteed by a remnant @xmath4 symmetry .
the higgs boson mass and the mass of the dark matter particle have a common origin , the vacuum expectation value of @xmath2 .
dark matter relic abundance is determined by annihilation @xmath5 .
we scan the parameter space of the model and find the mass of the dark matter particle in the range from 500 gev to a few tev .
ztf - ep-14 - 03 + * electroweak breaking and dark matter from the common scale + * * sanjin beni and branimir radovi + * + |
hydrodynamical modeling based on the interacting - winds theory ( kwok , purton , & fitzgerald 1978 ; kahn 1989 ) has proved to be quite successful in reproducing the overall shapes of many planetary nebulae ( pne ) .
but pne are also known to contain a variety of small - scale structures in the form of low - ionization patches , bullets , ansae , jets , point - symmetrical knots , filaments etc .
, whose existence was noted in the past by several authors ( e.g. , balick 1987 ; schwarz 1992 ; see also the review by mellema 1996 ) .
the attention of researchers toward this kind of `` secondary '' morphological component of pne has increased continuously since the publication of the seminal papers by balick et al .
( 1993 , 1994 , 1998 ) , who also coined the term fliers ( fast low ionization emitting regions ) to indicate the low - ionization , small - scale structures found in various pne which move supersonically relative to their ambient gas .
much work has been done since , highlighting the existence of a large variety of low - ionization microstructures in pne : some are found inside the main bodies of the nebulae ( e.g. , ngc 3242 ; balick et al .
1993 ) , others outside ( fg 1 ; lopez , roth & tapia 1993 ) , some move fast ( hb 4 ; hajian et al .
1997 ) , and others share the expansion velocity of their environments and thus are not genuine fliers ( e.g. , ngc 7662 ; balick et al .
1998 ) ; some lie along the symmetry axis of the nebulae ( ngc 7009 ; balick et al . 1998 ) , while others are apparently directed along multiple / random directions ( ic 4593 , corradi et al .
1997 ) , etc .
in particular , several authors have focused the attention on remarkable point - symmetrical features which strongly suggest the occurrence of episodic , precessing ejecta from the central stars ( e.g. , schwarz 1992 ; corradi & schwarz 1993 ; lopez , meaburn & palmer 1993 ; guerrero & manchado 1998 ) .
balick et al .
( 1998 ) discuss the difficulties of the existing dynamical models in accounting for all the properties of the sample of fliers that they observed .
while theoretical models clearly urge , further observations would help to get a more precise characterization of all those phenomena that come under the quite generic definition of `` low - ionization , small - scale structure '' in pne .
in particular , they would help in determining whether one should try to constrain all types of low - ionization microstructures in pne into a common , single theoretical framework or , as it is more likely , at least to restrict their origin and evolution to a limited number of different dynamical / radiative processes . in this paper
, we present images and radial velocity measurements for three pne from the sample of corradi et al .
( 1996b ) , who highlighted the existence of low - ionization small - scale structures in 23 pne by computing ( + ) /ratio maps in the image catalog by schwarz , corradi & melnick ( 1992 ) .
an analysis for other six pne will be presented in forthcoming papers .
images and long - slit spectra of ngc 3918 ( pn g294.6 + 04.7 ) , k 12 ( pn g253.5 + 10.7 ) , and wray 17 - 1 ( pn g258.0 - 15.7 ) were obtained on 1996 april 2627 at the 3.5-m new technology telescope ( ntt ) at eso , la silla ( chile ) , using the emmi multimode instrument . with the tek 2048@xmath4 ccd eso#36 , the spatial scale of the instrument was [email protected] pix@xmath6 both for imaging and spectroscopy .
the central wavelength and full width at half - maximum ( fwhm ) of the filter used for imaging are 658.8 nm and 3.0 nm , respectively .
further details of the observations are listed in table 1 . as with the spectroscopy
, emmi was used in the long - slit , high - resolution mode ( corradi , mampaso & perinotto 1996a ) , providing a reciprocal dispersion of 0.004 nm pix@xmath6 , and a spectral resolving power of @xmath7/@xmath8=55000 with the adopted slit width of [email protected] .
the slit length was of 6 arcmin .
the echelle order selected by using a broad filter includes the line at @xmath7=656.01 nm , at @xmath7=656.28 nm , and the doublet at @xmath7=654.81 and 658.34 nm .
the slit was positioned through the centers of the nebulae , along various position angles as detailed in table 1 .
narrow - band images of k 1 - 2 and wray 17 - 1 were also obtained at the 1.54-m danish telescope at eso la silla on 1995 january 56 . with
the tek 1000@xmath4 ccd eso#28 , the spatial scale of the instrument was [email protected] pix@xmath6 .
the central wavelength and fwhm of the filters used at the 1.54 dan were : 469.7/11.4 nm for , 486.2/7.5 nm for , 500.5/7.3 nm for , 655.9/1.3 nm for , 658.4/2.6 nm for , and 672.7/6.3 nm for .
exposure times and seeing are given in table 1 .
we also retrieved an image of ngc 3918 from the _ hst _ archive obtained on 1995 with the wfpc2 camera ( pc ccd , [email protected] pix@xmath6 ) in the f555w filter ( 525.2/122.3 nm ) .
the emission of ngc 3918 in this broad - band _ hst _ filter is dominated by the very strong lines at @xmath9495.9 and 500.7 nm .
this is confirmed by an image taken at the ntt ( not presented here ) which shows the same overall structure as the _ hst _ one , apart from the obvious coarser resolution .
however , various other nebular lines ( including he @xmath10 468.6 , and n @xmath10 575.5 ) fall within the transmission range of the filter and may be important at specific positions in the nebula .
images and spectra were reduced in a standard way using midas .
ngc 3918 is a widely studied pn
. detailed work on this object was presented by clegg et al .
( 1987 ) , who derived parameters for the central star and nebular chemical abundances , and concluded that the nebula is optically thin . from the limited morphological and kinematical information they possessed , clegg et al .
( 1987 ) adopted a biconical `` pole - on '' geometry for the nebula , and derived an expansion age of 3000 y for a distance of 1.5 kpc .
other kinematical studies limited to the central region of the nebula were presented by dopita ( 1978 ) and pea & torres - peimbert ( 1985 ) .
the morphology of ngc 3918 is complex , and we will focus the discussion only on its most relevant properties , distinguishing between the large - scale components , which determine the overall appearance of the nebula , and the low - ionization , small - scale features , which are observed in several regions of the nebula . the _ hst _ and ntt images of ngc 3918 , rotated so as to have the symmetry axis of the nebula aligned with the vertical direction , are presented in figure [ f - ima ] . at high intensity levels ( fig .
[ f - ima ] , top left ) , the _ hst _ image shows a bright inner shell of roughly elliptical shape from which two fainter protrusions extend in the vertical ( @xmath11 east west ) direction , giving to the whole structure a spindle - like appearance . in what follows
, we will call this the `` inner shell '' .
its size is 12@xmath5@xmath1220@xmath5 , measured along its minor and major axes .
it is surrounded by an elliptical `` outer shell '' of 16@xmath5 diameter . at lower intensity levels ( fig .
[ f - ima ] , top right ) , the _ hst _ image shows that the outer shell also possesses faint protrusions ; they are 1.5 times more extended than those of the inner shell . in addition , the outer shell is surrounded by a circular region of very faint emission showing a number of radial filaments .
the brightness of the outer shell , as compared to the inner one , is enhanced in ( fig .
[ f - ima ] , bottom left ) .
several low - ionization small - scale structures , located roughly along the major axis of the inner shell , are clearly visible in the image ( fig .
[ f - ima ] , bottom right ) . for the sake of clarity , features which are further discussed in the following
are labeled in fig .
[ f - ima ] with the letters @xmath13 , @xmath14 , @xmath15 , @xmath16 , @xmath17 and @xmath18 .
the most notable of these is the highly collimated feature on the lower ( eastern ) side of the nebula which appears as a thin straight lane ( @xmath15 ) extending out of the inner shell along its major axis ( see also corradi et al .
1996b ) . at its end , this structure brightens and broadens ( @xmath14 and @xmath13 ) . in the _ hst _ image ,
it is surrounded by a broad arc of emission , corresponding to the southern , faintest protrusion mentioned above . on the opposite ( western ) side of the central star ,
there is one main condensation ( @xmath16 , first identified by corradi et al .
1996b ) lying at the edge of the outermost upper protrusion of the nebula .
moreover , on the same side there is another faint , collimated bridge of emission , barely visible in our image and not detected previously , which extends far from the inner shell up to positions labeled as @xmath18 and @xmath17 .
it is also aligned with the long axis of the inner shell .
the ntt spectra of ngc 3918 are presented in figure [ f - spec ] .
the slit was located through the central star , along the major ( p.a.=@xmath19 ) and minor ( p.a.=@xmath20 ) axes of the nebula ( i.e. vertically and horizontally , respectively , as indicated by short lines on either side of the nebula in fig .
[ f - ima ] , bottom right ) . from images and spectra , we derived basic kinematical , geometrical and orientation parameters for the spindle - shaped inner shell of ngc 3918 .
first , we note that both its shape and spectrum are consistent with the prediction of the interacting - winds theory .
a very good qualitative fit to the observations is in fact given by the axisymmetrical models of mellema ( 1993 ) , which consider an initial mass distribution strongly enhanced in an `` equatorial '' plane , forming a sort of circumstellar torus with an equatorial - to - polar density contrast as large as @xmath210 ( parameters @xmath16=6 and @xmath13=0.9 in his calculations ) .
this initial mass distribution , under the action of the expanding shock driven by the fast wind , would form a shell of compressed gas with a spindle - like shape ( fig .
[ f - mellema ] , left ) , as observed for the inner shell of ngc 3918 ( fig .
[ f - ima ] , top left ) .
the shape of the long - axis , position
velocity plot is also qualitatively reproduced ( cf .
[ f - mellema ] , right , with the spectrum at p.a.=@xmath19 in fig .
[ f - spec ] ) .
according to this modeling , the protrusions of the inner shell would develop along the _ polar _ axis of the system , the observed short axis of the nebula being the projection on the sky of the equatorial density enhancement .
for this reason , in the following the protrusions and jet - like features will be referred to as `` polar '' structures .
the outer shell and its more external extensions would in turn represent the ionized unshocked remnant of the agb wind .
a more detailed comparison with the model of mellema ( 1993 ) is limited by the fact that the figures in his article show images and spectra computed at a specific evolutionary time , while on the contrary we have images and , , and spectra .
detailed hydrodynamical modeling of ngc 3918 , taking into account the evolutionary status of its central star , would be very useful in determining geometrical parameters and derive a dynamical age for the nebula . for the time being , we have derived the inclination and kinematical parameters of the spindle - shaped inner shell making use of the spatio - kinematical empirical model of solf & ulrich ( 1985 ) , which allows us to treat the kind of 3-d geometry described above .
we fit the position
velocity plots and the shape of the inner shell with a model in which the nebular expansion velocity @xmath21 increases from the equatorial plane toward the polar axis , following the equation ( solf & ulrich 1985 ) @xmath22 where @xmath23 is the latitude angle ( = 0@xmath24 in the equatorial plane , 90@xmath24 in the polar directions ) , @xmath25 and @xmath26 are the polar and equatorial velocities , and @xmath27 a shape parameter .
the other quantities involved in the model are the heliocentric systemic velocity , @xmath28 , the inclination , @xmath29 , of the nebula ( angle between the polar axis and the line of sight ) , and the product containing the inseparable effects on the apparent nebular size due to the distance , @xmath1 , and to the kinematical age , @xmath30 . as with ngc 3918
, we have measured by gaussian fitting ( with 2-pixel spatial binning ) the radial velocities in the different regions of the nebula , both at p.a.=@xmath19 and p.a.=@xmath20 .
the best fit of the shape and kinematics of the inner shell ( fig .
[ f - fit ] ) is then obtained with the following parameters : @xmath3175@xmath32 deg ; = 1.0@xmath33 y pc@xmath6 ; @xmath34=50 ; @xmath26=23 ; @xmath35 ; and v@xmath36=@xmath37@xmath384 .
the systemic velocity agrees with the value @xmath39@xmath383 measured by meatheringham , wood , & faulkner ( 1988 ) . adopting a distance @xmath1=1.5 kpc , a kinematical age of 1500 y
is computed for the inner shell , about half the value derived by clegg et al .
non - negligible deviations from the models are , however , seen in fig .
[ f - fit ] , especially in the long - axis position velocity plot .
these deviations probably reflect the simplified assumptions in the spatio - kinematical model ( a time - independent expansion velocity described by a convenient analytical expression ) .
nevertheless , the basic parameters relevant to the following discussion , namely the inclination , @xmath29 , of the nebula on the sky and the kinematical age parameter , are safely constrained in the range of values quoted above , regardless of whether the shape of the velocity field is reproduced exactly or not .
this is because @xmath29 basically depends only on the `` inclination '' of the kinematical plot , while mainly depends on the apparent size of the nebula and on the polar velocities , the former quantity being precisely measured from the images and the latter being determined with sufficient accuracy from the modeling .
comparison of the velocities with the he @xmath10 656.0 and ones shows that : i ) the expansion velocity is between 2 and 5 lower for than for h@xmath40 , while it is higher by a variable amount for the ion ( this is the so - called `` wilson effect '' ; wilson 1948 , 1950 , 1958 ; see also dopita 1978 and pea & torres - peimbert 1985 for ngc 3918 ) ; ii ) the velocity plots are much more irregular and knotty than the and ones and are likely to represent the expansion velocity of the outer , rather than the inner , shell , as suggested by the images .
the biconical model adopted by clegg et al . ( 1987 ) for ngc 3918 is evidently inconsistent with the present data and the results of our modeling .
the inclination computed for the inner shell allows us to determine the space velocities and kinematical ages of the jet - like structures _ on the hypothesis that they are located along the polar axis of the inner shell _ , as strongly suggested by the observed morphology .
their radial velocities are presented in figure [ f - veljet ] .
these are also listed in table 2 , together with distances , deprojected velocities ( adopting @xmath41 ) and kinematical age parameter , .
radial velocities along the eastern filament increase linearly ( in absolute value ) from 8 to 13 arcsec ( @xmath15 ) from the central star , and this linear trend is also followed by the outermost emission ( @xmath13 ) .
line profiles are not symmetric but show faint extended red wings . at position @xmath14 , where the jet - like structure brightens in the images ,
there are also components with substantially smaller radial velocities .
we note that the / line ratio `` flares '' at position @xmath14 , being twice as large as in the inner shell and in the rest of the jet .
thus the articulated morphology ( @xmath13+@xmath14 and the surrounding arc ) and kinematics , and the peculiarly high /enhancement define a complex region of interaction at the leading edge of the jet .
considering now the @xmath14 and @xmath15 components , and deprojecting velocities using the same inclination as for the inner shell , we find that the eastern jet expands at considerable velocities , increasing linearly from 40 at 8@xmath5 from the center to almost 100 at its tip , @xmath13 .
a remarkable result is that the kinematical age of this jet is very similar to that of the inner shell . on the other side of the nebula
, the symmetry in the morphology and kinematics shows that the very faint and extended emission on the western side of the nebula ( @xmath18 and @xmath17 ) is the counterpart of the eastern jet .
although the signal is very faint here , also on this side there are in fact hints of a linear increase of velocities from the end of the inner shell to the @xmath17 position ( fig .
[ f - veljet ] ) , and @xmath17 has a space velocity and kinematical age only slightly larger than @xmath13 .
turning now to the bright knot @xmath16 , this appears on the western side of the nebula , roughly located along the major axis of the inner shell .
however , it shows negative radial velocities indicating that it is not expanding along the polar axis of the shell .
its apparent association with the edge of the faintest polar extensions of the outer nebula ( see figs .
[ f - ima ] and [ f - veljet ] ) suggests that instead this protrusion has an orientation on the sky very different from that of the inner shell .
its origin is unclear .
the present data allowed us to determine the basic geometrical and dynamical properties of ngc 3918 , making it an excellent test case for the theories of jet formation in pne .
first , the properties of the main nebula strongly resemble those expected according to the commonly accepted scenario of the interacting - winds theories , the basic shape and kinematics being nicely reproduced by the hydrodynamical models in mellema ( 1993 ) .
a comparison with his calculations would imply that the original density distribution of the stellar envelope ejected at the tip of the agb was markedly anisotropical , most of the mass being lost along the equatorial plane in a toroidal configuration .
spatio - kinematical modeling provided us with the orientation angle on the plane of the sky , the deprojected expansion velocities and a kinematical age for the main nebula .
secondly , the location and kinematics of the extended , two - sided jet are fully consistent with the hypothesis that it is expanding along the polar axis of the original toroidal density distribution , departing approximately from the position of the outer shock driven by the fast wind into the agb remnant . using the geometrical parameters of the main nebula
, we could compute the deprojected expansion velocities for the jet , information which is often difficult to derive ( mainly because the inclination is usually not easily determined ) .
the most important properties of the jet of ngc 3918 are the following : i ) its expansion velocity increases linearly approximately from the polar velocities of the inner shell ( 4050 ) to substantially larger values at its tips ( 100 ) , and ii ) the kinematical age of the jet is very similar to that of the main nebula .
this latter result suggests that the axisymmetrical shell and the polar jet of ngc 3918 were formed in the same dynamical process ( slow fast wind interaction ? ) , without invoking additional , later ejecta producing the highly collimated outflow .
this indicates the existence of some focusing mechanisms giving rise to accelerating , collimated polar flows during the dynamical evolution of axisymmetrical shells . as remarked by mellema ( 1996 )
, it is not generally clear whether the polar jets and knots observed in several axisymmetrical pne are the result of the evolution of the main shell ( i.e. , of the slow fast wind interaction ) , or just another independent product ( e.g. , later ejecta ) of whatever causes aspherical mass loss on the agb . observations of ngc 3918 strongly support the first hypothesis . along these lines , the models by frank , balick , & livio ( 1996 ) seemed very promising since they produce polar jets extending out of the outer shock during the interacting - winds evolution of the main shell , starting with density distributions which are also fully consistent with those deduced for ngc 3918 .
the crucial point there , as compared to other calculations , is to consider a gradual transition from the slow to the fast wind , which allows the formation of a polar jet in the early phases of the nebular evolution
. objects with large equatorial - to - polar density contrasts like ngc 3918 would sustain their focused flows best and make long jets .
dwarkadas & balick ( 1998 ) , however , argued against this jet formation mechanism , which would be inhibited by the presence of thin shell nonlinear instabilities .
they also warn about the possible numerical artifacts producing converging collimated flows in 2-d simulations .
it is hoped that the present observations of ngc 3918 , however , will stimulate further effort along the above - mentioned lines .
for instance , the transient , jet - like protuberances which appear in the models by dwarkadas & balick ( 1998 ) at later stages of evolution should be further considered .
inclusion of magnetic fields offers an appealing alternative solution to the problem of jet formation ( garca - segura et al .
1999 ) . their magnetohydrodynamical simulations , in fact ,
predict the formation of polar jets in planetary nebulae as a consequence of sufficiently strongly magnetized post - agb fast winds .
a noteworthy feature of these models is that they reproduce the linear increase of the expansion velocity along the jet that is observed in ngc 3918 .
the highly collimated outflows resulting from the models of garca - segura et al .
( 1999 , see also garca - segura 1997 ) are subject to kink instabilities , which might explain the knotty appearance of the jet of ngc 3918 .
the presence of magnetic fields would also relax the need for large equatorial - to - polar density contrasts in the original agb wind in order to obtain elongated nebulae and jets
. it would be highly desirable to try to model , along these lines , the detailed shape and kinematics of the shells and jets of ngc 3918 .
k 12 is one of the pne known to have a close binary system at its center .
the orbital period is of 0.676 days , which is considered to be the outcome of shrinking of the original orbit following a common - envelope phase of the system during the agb evolution of the primary star ( bond & livio 1990 ) .
the existence of the jet - like feature of k 1 - 2 was first reported by lutz & lame ( 1989 ) and later by bond & livio ( 1990 ) .
our narrow - band images are presented in figure [ f - imak12 ] .
the main nebula has a low surface brightness and an elliptical shape in and , while it shows extended emission ( `` ears '' ) in the ew direction in the image .
superimposed on this nebulosity , a knotty , collimated structure is visible along p.a.=@xmath42@xmath24 in all the observed emission lines , but particularly prominently in the light of low - ionization species like and
. this collimated outflow consists of two strings of knots aligned radially on opposite sides of the central star . the central star ,
however , is slightly offset from the line joining most of the features , and in addition the string of knots shows some bending .
some knots are resolved in our images as slightly elongated , elliptical blobs . on the nw side ,
the brightest portion of the low - ionization structure consists of a filament ( labeled @xmath13 in fig .
[ f - imak12 ] ) which is aligned exactly along the radial direction from the central star .
this filament is marginally resolved into knots in [ n @xmath43 , and has a faint tail ( @xmath14 ) directed outwards .
a careful inspection indicates that the brightest se knots , albeit very small ( @xmath44 ) , show some evidence of ionization stratification , with the emission slightly closer to the central star than the one .
in addition to this collimated system of knots , we confirm with the new imaging presented here the discovery by corradi et al .
( 1996b ) of the existence of other low - ionization microstructures , some of them located along directions nearly perpendicular to that of the main jet - like structure .
these other knots , indicated by the six vertical arrows in fig .
[ f - imak12 ] , also have a slightly elliptical shape and form a more sparse system than the main one .
the and long - slit spectrum of k 1 - 2 passing through the main system of knots ( p.a.=@xmath42 ) is shown in fig .
[ f - spectrak12 ] .
the spectrum is the superposition of the emission from the jet - like features and some extremely faint emission from the main elliptical nebula . by binning the signal in the central [email protected] ( free from any emission from the knots ) , a splitting of the line of 50
is measured , implying an expansion velocity for the elliptical shell of 25 .
the measured heliocentric systemic velocity @xmath28=65@xmath45 is in excellent agreement with the value of 66@xmath45 of schneider et al .
the fwhm of the line in each component of the elliptical shell ( 0.1 nm ) is significantly larger than that of expected from instrumental ( 0.01 nm ) and thermal ( 0.05 nm for @xmath46=10000 k ) broadening .
this might reflect the existence of a velocity gradient through the ( thick ) shell .
the and radial velocities for the system of knots were measured by gaussian fitting . for , velocities were measured only where it was possible to estimate the relative contribution of the surrounding elliptical nebula .
[ f - spectrak12 ] , right , shows the and velocities , which agree with each other . in the figure
, we also plot the velocity field that would be expected for the main elliptical shell assuming spherical symmetry , an average shell radius of 20@xmath5 ( measured from the and images ) , and adopting an expansion velocity of 25 .
although this is a crude estimate of the kinematics in the main nebula , it is clear that all the low - ionization features except @xmath14 have velocities similar to those computed for the shell . unless this comes from a conspiracy between real distances , expansion velocities , and projection effects
, this property would indicate that most of the low - ionization features are located within the main elliptical shell and move with velocities comparable to those of the surrounding gas ( and are hence not fliers ) .
the extended tail @xmath14 of the bright nw filament @xmath13 has peculiar velocities instead , larger than all the other features .
a detailed comparison of the and images suggests that filament @xmath13 is located within the main shell , while its tail , @xmath14 , extends outwards ; this would indicate that the gas is accelerating as it leaves the elliptical nebula . under the hypotheses that i ) @xmath13 is indeed located within the main elliptical shell and shares its expansion velocity , and ii ) the central splitting of @xmath47 can be taken as a measure of the expansion velocity of the main shell also at the location of filament @xmath13 , then from the observed radial velocities we obtain an inclination on the line of sight of 40 for the collimated structure .
this number should be taken with some caution , considering the uncertainty in the assumptions from which it is derived . with this inclination , space velocities in the tail @xmath14 would increase with radius from 25 to 45 at its outermost point .
both the and lines are moderately broadened , having fwhms ( not corrected for instrumental and thermal broadening ) between 0.025 and 0.045 nm , and between 0.07 and 0.09 nm , respectively .
there is also some evidence that the line becomes slightly narrower in the tail @xmath14 of the nw filament , strengthening the idea that this gas is expanding freely after leaving the shell .
no kinematical information is yet available for the sparse knots in the almost perpendicular direction . k 12 consists of i ) a diffuse elliptical shell expanding at a velocity of 25 , ii ) a main string of collimated , low - ionization knots at p.a.=@xmath42 , and iii ) a few other knots most of which are directed in nearly perpendicular directions .
the kinematical analysis suggests that most of the knots are embedded in the shell and have a similar expansion velocity , apart from the extended tail on the nw side which is composed of material accelerated to larger velocities ( @xmath245 at its end ) as it leaves the shell . the inclination with respect to the line of sight of the collimated outflow
is estimated to be around 40 .
this makes knowledge of the orbital parameters of the central binary star ( especially the inclination of the orbital plane and the position angle of the line of nodes ) very important , since it would help in understanding whether the collimated knots correspond to an `` equatorial '' ( i.e. , along the orbital plane ) , polar , or even meridional outflow .
some precession of the collimating source would also need to be invoked to explain the bending of the main string of knots .
as for the origin of this collimated feature , soker ( 1997 ) discusses several ways in which a close binary system such as k 12 can produce jets . in his view , polar jets are generally expected from stars that have undergone a common - envelope phase with stellar companions .
although this offers an explanation for the origin of jets in pne , their ( hydro)dynamical evolution should be worked out to allow for a detailed comparison with the present observations . in any case ,
knowledge of the spatial orientation of the jet of k 12 , together with future photometric , spectroscopic and polarimetric observations of its central star , might provide the first , crucial observational proof of the occurrence of collimated polar outflows in likely post common - envelope central stars of pne . for the time being , this hypothesis is supported by the intermediate inclination computed for the jet , since the binary system of k 1 - 2 is not seen close to edge - on ( it is not an eclipising binary ) , nor it is seen pole - on ( otherwise no significant photometric orbital variations would be detected ) .
the origin of the sparse knots in the other regions of the nebula remains unclear , lacking further spectroscopical information .
the central star of wray 17 - 1 is a non - pulsating pg1159-type star .
so far there are only about 20 hydrogen - deficient stars known ( cf .
werner 1993 ) .
nine of them are surrounded by a pn , wray 17 - 1 ( also called lo 3 ) being one of those .
detailed non - lte studies of various pg1159 stars have shown that they occupy a pre - white dwarf area in the hr diagram ; thus they are rather evolved objects .
a report and short description of the jet - like features of wray 171 is given by leisy & dennefeld ( 1993 ) .
our narrow - band images are presented in figure [ f - imawra171 ] . as in the case of k 12 , at faint brightness levels
the nebula appears as a diffuse elliptical shell . around it ,
complex extended nebulosity with bright arc - shaped limbs is observed , especially in our deep image . superimposed on the diffuse nebula
, there is a pair of regions of enhanced brightness ( labeled @xmath13 and @xmath16 in the image of fig .
[ f - imawra171 ] ) located symmetrically on opposite sides of the central star along the direction defined by p.a.=@xmath48 . in an almost perpendicular direction ( p.a.=@xmath49 ) and at similar distances from the central star , there is a second , fainter pair of structures composed of two small knots ( @xmath17 and @xmath1 ) surrounded by large arcs more visible in the image .
together , the two pairs of structures form a sort of `` four - leaf - clover '' morphology .
these four regions correspond to the low - ionization , collimated features reported by corradi et al .
( 1996b ) ; they are prominent in and [ s @xmath10 ] , and show a high degree of collimation in the / ratio map in figure [ f - imawra171 ] .
each region of the brightest pair ( @xmath51 ) looks like a diffuse patch of high - ionization gas from which narrower low - ionization radial tails depart in the outward directions .
this is further illustrated in fig .
[ f - zoom ] , where we present a zoom of the and image of @xmath13 ; the emission ( and presumably also the unobserved ) forms a sort of `` shadow '' of the patch .
these low - ionization tails roughly follow the surface brightness of the main elliptical nebula , disappear where the latter does , and reappear as faint outer knots ( @xmath14 and @xmath52 in fig .
[ f - imawra171 ] ) in correspondence to the limbs of the extended nebulosity . in fig .
[ f - zoom ] , a puzzling vertical `` bridge '' connecting @xmath13 to a stellar - like object is also observed .
the bridge is real , since it persists after careful subtraction of the point spread function of the stellar - like object .
the bridge is most visible in the [ n @xmath10 ] , and /images .
comparison of the brightness of the point - like object at the apex of the bridge with that of field stars in the different narrow - band images shows that it is very probably a red star ( unfortunately it falls out of the slit used for the kinematic measurements ) .
the second pair of low - ionization features at p.a=@xmath49 have a different morphology and ionization structure : the compact knots @xmath17 and @xmath1 are located inside two large arcs of higher ionization emission ( more visible in the image ) .
the knots are observed in all emission lines , but are prominent in low - ionization species ( see also corradi et al .
the knots are marginally resolved in our ntt images , and the western one ( @xmath1 ) has very faint tails directed both inwards and outwards ( see also fig .
[ f - imawra171 ] ) .
note that the central star is slightly offset from the lines joining each pair of features .
the spectra of wray 17 - 1 , taken at p.a.= @xmath48 and p.a.=@xmath49 to include the four low - ionization regions discussed above , are displayed in figure [ f - spectrawra171 ] . by spatially binning the signal in the central 16@xmath5 of the emission at p.a.=@xmath49 , we obtain a rough estimate of the expansion velocity for the main nebulosity ( half the line splitting ) of 28 ( cf . the value of 16 for quoted in acker et al .
1992 ) .
we measured and velocities through the low - ionization features along the two slit positions .
these are also shown in figure [ f - spectrawra171 ] . except for knot @xmath14 ,
radial velocities are smaller or similar to the adopted expansion velocity of the main nebula , but in contrast to the case of k 1 - 2 , it is not possible to get reliable estimates of the projection effects because of the complex geometry of the surrounding nebula . note that the arcs surrounding @xmath17 and @xmath1 have radial velocities comparable to those of the knots ( see the spectrum in fig .
[ f - spectrawra171 ] ) , confirming the physical association seen in the images .
line profiles are generally symmetric , and there is evidence for line broadening at the position of the knots @xmath17 and @xmath1 , especially in the emission ( measured fwhm=0.12 nm ) ; this might reflect the velocity dispersion within the arcs around the knots .
the outermost knot of the low - ionization tail of @xmath13 , located on the limb of the extended nebulosity and labeled as @xmath14 in fig .
[ f - imawra171 ] , shows a very peculiar radial velocity ( @xmath53 ) , strikingly different from that of the inner portions of the tail ( @xmath54 ) , even though they are perfectly aligned in the images .
the adopted heliocentric systemic velocity for wray 171 , taken as the average of the values derived from the central splitting , and from the symmetry in velocities for each pair of low - ionization structures , is v@xmath55=51@xmath385 .
wray 17 - 1 is a complex and most peculiar nebula .
the gross structure is that of a diffuse shell , roughly circular in the inner regions but with complex extended emission .
embedded in the diffuse nebula are two remarkable pairs of symmetrical , low - ionization structures .
the two pairs are located in almost perpendicular directions ( projected on the plane of the sky ) .
each pair is similar to the other in being located at comparable distances from the central star and in having an overall ionization state that is lower than in the surrounding nebula .
but they also show remarkable morphological differences .
one pair ( @xmath56 ) is composed of two compact , low - ionization knots surrounded by an outward - facing higher - ionization cap .
the other pair ( @xmath51 ) presents broader emission patches with collimated , low - ionization outward tails , whose brightness is modulated by that of the parent nebula .
the patches are observed in all our images in different ionization stages , thus they correspond to real density enhancements .
the tails are seen only in low - ionization emission ( and ) .
to understand the nature of these tails , a key observation is the large difference in radial velocity ( 90 ) between the inner part of tail of @xmath13 and its outermost knot , @xmath14 , on the limb of the extended nebula ( the same , albeit on smaller scale , applies to @xmath16 and @xmath52 ) .
it is unlikely that this velocity difference ( with sign reversal as compared to the systemic velocity of the nebula ) corresponds to a real change in velocity and direction in the same collimated outflow to which both @xmath13 and @xmath14 would belong .
it is more likely that the low - ionization tail is just an ionization effect ; i.e. , a lower - ionization region shielded from energetic photons from the central stars by the inner , higher - ionization patch .
when we measure velocities along the tail , we would then measure not the expansion velocity of an independent , collimated outflow , but just the kinematics of the diffuse nebula , and the velocity variations would just reflect the complex 3-d geometry , orientation and kinematics of the overall structure . the physical similarities ( and occam s razor ) would suggest that both pairs of low - ionization features have the same origin .
their structural differences might therefore reflect different initial parameters ( density , size , etc . ) or a different evolutionary stage of these condensations .
are they , for instance , high - density clumps which are on the way to being photo - evaporated by photons from the central star , and observed at different evolutionary stages ( cf . mellema et al .
1998 ) ? whatever the evolution of these components , their origin is difficult to explain .
the patches and knots appear as density enhancements inside the main nebula .
the fact that the features in each pair are located on opposite symmetrical positions with respect to the central star ( but note the small offset of the latter ) excludes a random formation of these condensations within the main nebula ; rather , it indicates the occurrence of collimated , two - sided ejecta ( or impinging winds ) .
this leads to another problem , the almost perpendicular relative orientation ( projected on the sky ) of the pairs .
collimated outflows are indeed expected along the symmetry axis of aspherical nebulae ( cf .
the discussion for ngc 3918 ) , but it is much more difficult to think of a mechanism able to produce independent collimated outflows along directions with have very different orientations in space ( but see ic 4593 ; corradi et al .
1997 ) .
finally , even more mysterious is the bridge of emission connecting patch @xmath13 to a star - like object which appears , in projection , inside the nebula .
the simplest possibility is that the association is not physical , and that some extension of the emission patch fortuitously coincides , in projection , with the position of a background / foreground star ( but note that the bridge clearly breaks the symmetry of the emission patch which is otherwise nicely pointing toward the central star ; cf .
[ f - zoom ] ) .
the other possibility is that the star is physically associated with the nebula , producing the bridge to the emission patch . in this case
, the object would be a field star running through the nebula or a companion star in a loose binary system .
no distance is tabulated for wray 171 in the literature , but for an hypothetical distance of 1 kpc , the `` companion '' star would be located at about 20000 a.u . from the central star of the pn and the corresponding orbital period
would be longer than a million year : this would make it a very loose binary system indeed . in any case , if the star is associated with the nebula , how could it produce the bridge ?
is it gas stripped from the emission patch by the star in its motions through the nebula ? or are @xmath13 and the bridge gas lost from the star itself ( but how to explain then the formation of the opposite symmetric patch @xmath16 , and of @xmath17 and @xmath1 ) ? with the information available it is not possible to investigate the nature of the nebula star connection further ; this clearly deserves future study .
the three pne studied in the present article provide valuable observational constraints in the discussion of the formation of collimated outflows and low - ionization microstructures in pne .
ngc 3918 proves to be a key example of the production of an extended jet along the symmetry axis of an elongated shell . from our data , we have determined to a good level of accuracy the geometrical , orientation , and dynamical properties of the shell and jet , making ngc 3918 an excellent test case to confront theories of jet formation in pne . according to pure hydrodynamical calculations ,
the axisymmetrical geometry of the shell would have developed starting from a very aspherical ( toroidal ) mass deposition at the tip of the agb .
whether these initial conditions naturally lead to the formation of the polar jet is still matter for debate ( balick et al .
1998 ) ; the inclusion of a strong magnetic field in the post - agb wind ( garca - segura et al . 1999 ) might contribute to a solution of the problem .
k 12 shows a collimated string of knots within an elliptical shell , ejected from a close binary system which probably underwent a common - envelope phase during the agb phase of the star which is producing the pn .
estimation of the projection of these collimated structures gives us the opportunity to verify the idea that close binaries can produce `` polar '' outflows , provided that the orientation parameters of the binary systems are determined .
further photometric , spectroscopic , and/or polarimetric observations of the central source directed toward this aim are highly desirable .
finally , wray 171 is a most unusual and interesting object , and its two pairs of low - ionization structures located in almost perpendicular directions ( and with notable similarities / differences between each other ) are hardly reconcilable with current ideas of the production of collimated outflows in pne .
some of these low - ionization features ( their collimated outward tails ) seem to be produced by photoionization effects .
wray 17 - 1 is a case which shows that more than one physical process should be considered in accounting for the variety of low - ionization , small - scale structures observed in pne .
we thank garrelt mellema for providing us with fig .
[ f - mellema ] , and howard bond and don pollacco for useful discussion on the central star of k 1 - 2 .
the work of rlmc , ev , and am is supported by a grant of the spanish dges pb971435c0201 , and that of drg by a grant of the brasilian agency fapesp ( proc 98/75020 ) .
ngc 3918 & ntt & ( 2 ) & [email protected] & hst & f555w ( 3 ) & k 12 & ntt & ( 5 ) & [email protected] & dan & ( 20 ) , ( 20 ) , ( 20 ) , ( 20 ) & [email protected] wray 171 & ntt & ( 2 ) & [email protected] & dan & ( 15 ) , ( 15 ) , ( 5 ) , ( 15 ) , ( 15 ) & [email protected] & & & & & p.a .
( exposure time , min ) & ngc 3918 & ntt & @xmath19 ( 30 ) , @xmath20 ( 10 ) & [email protected] k 12 & ntt & @xmath42 ( 60 ) & [email protected] wray
171 & ntt & @xmath48 ( 60 ) , @xmath49(60 ) & [email protected] @xmath13 & 17.3 & 25 & 96 & 0.9 @xmath14 & 14.5 & 13 & 48 & 1.5 @xmath15 & 10.3 & 13 & 51 & 1.0 inner shell & 9.3 & & 50 & 1.0 @xmath16 & 13.4 & 10 & & @xmath18 & 24.5 & @xmath5821 & 80 & 1.5 @xmath17 & 27.8 & @xmath5828 & 106 & 1.3 | we analyze optical images and high - resolution , long - slit spectra of three planetary nebulae which possess collimated , low - ionization features .
ngc 3918 is composed of an inner , spindle - shaped shell mildly inclined with respect to the plane of the sky .
departing from the polar regions of this shell , we find a two - sided jet expanding with velocities which increase linearly with distance from 50 to 100 .
the jet is probably coeval with the inner shell ( age @xmath0@xmath1 y , where @xmath1 is the distance in kpc ) , suggesting that its formation should be ascribed to the same dynamical processes which also shaped the main nebula , and not to a more recent mass loss episode .
we discuss the formation of the aspherical shell and jet in the light of current hydrodynamical and magnetohydrodynamical theories .
k 1 - 2 is a planetary nebula with a close binary nucleus which shows a collimated string of knots embedded in a diffuse , elliptical shell .
the knots expand with a velocity similar to that of the elliptical nebula ( @xmath225 ) , except for an extended tail located out of the main nebula , which linearly accelerates up to @xmath3 .
we estimate an inclination on the line of the sight of @xmath240 for the string of knots ; once the orientation of the orbit is also determined , this information will allow us to test the prediction of current theories of the occurrence of polar jets from close binary systems .
wray 17 - 1 has a complex morphology , showing two pairs of low - ionization structures located in almost perpendicular directions from the central star , and embedded in a large , diffuse nebula .
the two pairs show notable similarities and differences , and their origin is very puzzling .
# 1.#2 # 1#2 |
decay of @xmath0 mesons plays an important role in particle physics since it is a testing ground for the standard model and a window for possible new physics .
we can obtain good information on the cabibbo - kobayashi - maskawa ( ckm ) matrix elements and cp violation from @xmath0 decays . since the contribution of qcd in @xmath0 decays changes the whole structure of the theory , the study of @xmath0 decays is an intertwined field of particle physics . among these decays ,
nonleptonic @xmath0 decays have been the subject of intense interest .
especially the treatment of nonperturbative effects from the strong interaction is a serious theoretical problem in nonleptonic decays .
precise experimental observation of nonleptonic @xmath0 decays makes it urgent to give firm theoretical prediction including the effects of cp violation .
there has been a lot of theoretical progress in nonleptonic @xmath0 decays and we suggest how to consider nonleptonic @xmath0 decays from the viewpoint of the soft - collinear effective theory ( scet ) . the effective hamiltonian for nonleptonic decays from the standard model has been derived and the wilson coefficients of the operators for @xmath0 decays have been calculated to next - to - leading order , and next - to - next - to - leading order for some operators @xcite . in order to evaluate the hadronic matrix elements of four - quark operators , naive factorization @xcite was assumed , in which the matrix elements were reduced to products of current matrix elements .
but there was no justification for this assumption except the argument of color transparency @xcite . besides that , decay amplitudes depend on an arbitrary renormalization scale @xmath8 in naive factorization since the wilson coefficients depend on @xmath8 , while the matrix elements of operators do not .
ali et al . @xcite
have improved the problem of the scale dependence by including radiative corrections of the operators before taking hadronic matrix elements .
then the @xmath8 dependence of the wilson coefficients is cancelled by that of the radiative corrections of the operators .
however , the decay amplitudes depend on calculation schemes since the off - shell renormalization scheme is used @xcite .
politzer and wise @xcite suggested to take the heavy quark mass limit in computing corrections to the decay rate ratio @xmath9 .
they have considered radiative corrections for nonfactorizable contributions and found that they are finite and the decay amplitudes are factorized .
beneke et al .
@xcite have extended this idea to general two - body decays including two light final - state mesons , in which the decay amplitudes can be expanded in a power series of @xmath10 .
they show that nonfactorizable contributions including spectator interactions are factorized as a convolution of the hard scattering amplitudes with the meson wave functions , and the corrections are suppressed by powers of @xmath10 .
this is an important step toward theoretical understanding of nonleptonic @xmath0 decays .
first the amplitudes can be obtained from first principles and a systematic @xmath10 expansion is possible .
second , since the on - shell renormalization is used , there is no scheme dependence . and they improved previous approaches by including momentum - dependent parts , which had not been included previously .
however , when higher - twist light - cone wave functions are included , there appear infrared divergences in the amplitudes .
these are treated as theoretical uncertainties , but from the theoretical viewpoint it is a problem to be solved in this approach .
bauer et al .
@xcite have proposed an effective field theory in which massless quarks move with large energy .
this effective theory , called the soft - collinear effective theory , is appropriate for light quarks with large energy .
it has been applied to hard scattering processes and @xmath0 decays @xcite .
it is also an appropriate effective theory for nonleptonic @xmath0 decays to two light mesons . in this paper
, we apply scet to nonleptonic @xmath0 decays into light mesons .
we construct all the relevant operators in the effective theory at leading order in a gauge - invariant way by integrating out all the off - shell modes .
the wilson coefficients are calculated by matching the full theory onto scet .
we show that the four - quark operators in scet are factorized to all orders in @xmath1 and the argument of color transparency is explictly shown .
and we consider nonfactorizable spectator contribution in scet , and find that they are also factorized to all orders in @xmath1 .
the basic idea of the factorization properties in @xmath0 decays into two light mesons was sketched in ref .
@xcite . in this paper , we extend the argument and discuss intricate characteristics of scet in nonleptonic decays , the details of the procedure of matching , and present all the wilson coefficients and technical details in the calculation .
we also present the analysis of @xmath11 decays as an application , which is consistent with the calculation in the heavy quark mass limit at lowest order in @xmath1 .
the organization of the paper is as follows : in section [ sec2 ] , we consider the effective theories @xmath2 and @xmath5 , which we employ in the analysis of nonleptonic @xmath0 decays and explain the field contents and discuss how to match these effective theories . in section [ sec3 ] , we construct four - quark operators in @xmath2 , which are gauge invariant by integrating out off - shell modes .
this is achieved by attaching gluons to fermion legs and by integrating out off - shell intermediate states .
we discuss the factorization properties of the operators . in section [ sec4 ] , we match the full theory and @xmath2 and obtain the wilson coefficients of the four - quark operators . in section [ sec5 ] ,
we consider nonfactorizable spectator interaction , in which subleading operators are enhanced to give the leading - order result .
the structure of the subleading operators and the mechanism for the enhancement are determined by the power counting and the matching between effective theories . in section [ sec6 ]
, we consider the contributions to the heavy - to - light form factor .
the time - ordered products of subleading operators from the heavy - to - light current and the interaction of an ultrasoft ( usoft ) quark and a collinear quark contribute to the form factor .
in section [ sec7 ] we combine all the results to apply scet to nonleptonic decays @xmath12 . and finally a conclusion is presented . in appendix
a , we show how the auxiliary field method can be used to derive gauge - invariant four - quark operators in @xmath5 . in appendix b
, we employ the auxiliary field method to derive gauge - invariant subleading operators in @xmath2 .
the formalism of scet and the procedure of the two - step matching are discussed comprehensively in refs .
@xcite and in ref .
we will not discuss them in detail , but we review them briefly here .
the momentum of an energetic quark moving in the light - cone direction @xmath15 can be decomposed as @xmath16 where @xmath15 , @xmath17 are two light - like vectors satisfying @xmath18 , @xmath19 .
there are three scales @xmath20 , @xmath21 , and @xmath22 , where @xmath23 is the hard scale .
the momentum squared @xmath24 is typically of order @xmath25 , where @xmath26 is a typical hadronic scale @xmath27 0.5 gev .
we introduce a small expansion parameter @xmath28 to facilitate the power counting .
we construct scet in two steps , namely @xmath29 for @xmath30 , and @xmath5 for @xmath31 .
formally we can construct @xmath5 directly from the full theory , but it is conceptually easy to construct @xmath29 and @xmath5 successively . in scet , the collinear quarks @xmath32 and gluons @xmath33 have momenta @xmath34 .
there are the soft fields @xmath35 , and @xmath36 , with momenta @xmath37 , and the usoft fields @xmath38 , @xmath39 with momenta @xmath40 .
the collinear quarks interact with collinear gluons or usoft gluons , but not with soft gluons since they make collinear particles or heavy quarks off the mass shell . in order to derive the power counting method in @xmath2 , we move all the dependence on @xmath41 to the interaction vertices , and determine the @xmath41 dependence of all the fields @xcite
. then we can construct all the operators in powers of @xmath41 systematically by integrating out the off - shell modes of order @xmath23 .
the guiding principles of constructing operators in scet are gauge invariance @xcite , reparameterization invariance @xcite and the power counting @xcite .
the wilson coefficients of the operators can be obtained through the matching between the full theory and @xmath2 by requiring that the matrix elements of operators in both theories be the same at any order of the perturbation theory .
there is no mixing of operators with different powers of @xmath41 through radiative corrections as long as @xmath42 since the matching is performed perturbatively . in @xmath5 for @xmath43 , we integrate out all the off - shell modes of order @xmath44 .
then the collinear fields in @xmath5 have momenta @xmath45 , and the soft degrees of freedom have momenta @xmath46 , where @xmath47 is a new small expansion parameter in @xmath5 . here
the collinear modes have momenta @xmath48 , which is appropriate to describe hadrons of size @xmath49 .
the degrees of freedom of order @xmath50 in @xmath2 are all integrated out , and the usoft degrees of freedom in @xmath2 of order @xmath51 remain in @xmath5 with momentum @xmath52 , and we call them soft degrees of freedom in @xmath5 .
the matching of @xmath2 onto @xmath5 is performed at @xmath53 .
a subtle technical point in matching is to require gauge invariance . in order to see how to keep the gauge invariance ,
let us consider a simple example of a heavy - to - light current . in the full theory the current @xmath54
is matched to the operator in @xmath2 as @xmath55 where @xmath56 is the wilson coefficient , and @xmath57 is the label momentum operator .
the factor @xmath58 is the wilson line which is given by @xmath59.\ ] ] this wilson line is obtained by attaching collinear gluons to the heavy quark and integrate out the off - shell modes of the intermediate states of order @xmath23 . in @xmath5 , the emission of soft gluons @xmath60 makes the collinear fields off the mass shell and we have to integrate out the off - shell modes .
this is achieved by attaching usoft gluons to external fermion lines in @xmath2 and by integrating out the off - shell modes of order @xmath61 .
it corresponds to factorizing the usoft - collinear interactions with the field redefinition @xcite @xmath62 the collinear effective lagrangian can be written in terms of @xmath63 and @xmath64 , that is , the collinear - usoft interactions are factorized with the field redefinitions given in eq .
( [ redef1 ] )
. then we match @xmath2 onto @xmath5 at a scale @xmath65 .
for example , the heavy - to - light current is matched as @xmath66 where @xmath67 . in @xmath5 ,
the fields with the superscript @xmath68 are fundamental objects .
we rename the usoft field as the soft field in @xmath5 .
we write @xmath69 , and lower the off - shellness of the collinear fields . since the leading collinear lagrangian is the same in @xmath2 and @xmath5 , we can simply replace @xmath70 , and we obtain the final form of the operator @xmath71 where the superscript @xmath72 , which denotes @xmath5 , will be dropped from now on . from the two - step matching , it is clear why the wilson coefficients do not depend on the soft momentum .
the advantage of the two - step matching is manifest when we consider time - ordered products in @xmath2 since these can induce jet functions involving the fluctuations of order @xmath73 . in @xmath2
we can compute the jet functions with a well - defined set of feynman rules independent of the computation of the wilson coefficients at @xmath74 . and
the scaling of operators in @xmath5 is constrained by the power counting in @xmath2 and especially , @xmath2 puts a constraint of the number of factors of @xmath75 , which can be induced from the fluctuations @xmath76 .
therefore we can know how many powers of @xmath76 appear for a given process at a given order in @xmath1 , and this power counting is not spoiled by the loop effects since there is no dependence on the soft momentum in the wilson coefficients .
this feature will be seen explicitly when we evaluate the decay amplitudes of nonleptonic decays which involve time - ordered products . in nonleptonic
decays into two light mesons , there are collinear quarks moving in opposite directions .
we choose these two directions as @xmath15 and @xmath17 . for an energetic quark moving in the @xmath17 direction ,
the momentum is decomposed as @xmath77 we denote a collinear quark by @xmath32 ( @xmath78 ) which moves in the @xmath15 ( @xmath17 ) direction .
these fields satisfy the relations @xmath79 the collinear gauge field in the @xmath15 ( @xmath17 ) direction is written as @xmath33 ( @xmath80 ) .
the effective lagrangian for @xmath78 and @xmath80 , which can be obtained from the colllinear lagrangian for @xmath32 and @xmath33 by replacing @xmath81 , @xmath82 respectively .
this situation is simlar to two jets in the opposite direction @xcite , but we have a heavy quark interacting with collinear gluons in both directions , which makes the analysis more interesting and complicated .
the effective hamiltonian for @xmath0 decays in the full theory is given as @xmath83 where the local @xmath84 operators are given by @xmath85 here @xmath86 and @xmath87 denotes down - type quarks , @xmath88 is the chromomagnetic field strength tensor , and @xmath89 are the color @xmath90 generators .
the process of obtaining the gauge - invariant operators in scet requires two - step matching @xcite because the scet involves two different scales @xmath91 and @xmath92 .
first we match the full theory onto @xmath2 at @xmath93 , and we match successively onto @xmath5 at @xmath94 .
a concrete example of the two - step matching was illustrated in ref .
@xcite . in order to construct the operators in scet
, we first have to specify which quark or antiquark goes to a certain direction to form a light meson .
we set @xmath17 as the direction of a quark - antiquark pair to form a light meson , and @xmath15 as the direction of the remaining quark which combines with a spectator quark in a @xmath0 meson to form another meson .
therefore the construction of the operators is process - dependent in the sense that we first specify the direction of each outgoing quark , and the number of operators in scet is doubled because we have two possibilities to assign two quark fields in both directions .
a generic four - quark operator for nonleptonic @xmath0 decays in scet has the form @xmath95 , or @xmath96 , where @xmath97 and @xmath98 are dirac matrices , and @xmath99 is the heavy quark field in the heavy quark effective theory .
these operators are derived from the operator @xmath100 in the full theory where @xmath101 ( @xmath102 ) are light quarks .
the operator @xmath103 is obtained by replacing @xmath104 by @xmath32 , and @xmath105 and @xmath106 by @xmath78 . for the operator @xmath107 ,
we replace @xmath105 by @xmath32 and @xmath104 and @xmath106 by the @xmath78 fields .
the second operator produces a light meson , in which one quark comes from the heavy - to - light current and another antiquark from the light - to - light current to form a meson in the @xmath17 direction .
a remaining quark goes to the @xmath15 direction .
we can transform this operator to the form @xmath108 by fierz transformation . in order to simplify the organization of the computation
, all the operators will be written in the form @xmath109 .
if a fierz transformation is necessary to obtain this form , we apply fierz transformation in the full theory and perform matching accordingly .
therefore we only have to consider operators of the form @xmath110 , but the two types of the operators with or without fierz transformation are regarded as distinct and their wilson coefficients are different . though we know the form of the operators , the operator @xmath111 itself is not gauge invariant under the collinear , soft and usoft gauge transformations . in order to obtain gauge - invariant operators in @xmath5 , we employ two - step matching .
we consider the emissions of collinear gluons @xmath33 and @xmath80 from external fermions in the full theory and integrate out the off - shell intermediate states of order @xmath112 to obtain the collinear gauge - invariant operators in @xmath2 .
then we go down to @xmath5 , in which we consider the emissions of soft gluons @xmath36 which have momenta of order @xmath113 .
the emission of soft gluons can cause intermediate states off the mass shell by the amount @xmath114 , which should be integrated out to obtain the final form of the operators in @xmath5 . to simplify the calculation , we will directly match the full - qcd operators to the operators in @xmath5 by considering the emissions of the collinear and the soft gluons from each fermion in the full qcd and integrate out all the intermediate states of order @xmath112 and @xmath115 to obtain the gauge - invariant operators in @xmath5 , as was done in ref .
the result is equivalent to the two - step matching described above .
we will use the two - step matching explicitly in treating the time - ordered products .
we can construct gauge - invariant operators by attaching collinear or soft gluons to each fermion in the full qcd and integrate out off - shell modes .
note that the soft gluons here have momenta of order @xmath116 , defined in @xmath5 . because the ordering of gauge fields is important due to the nonabelian nature of the gauge fields , we consider corrections to order @xmath117 .
typical feynman diagrams at order @xmath117 are shown in fig .
[ fig1 ] , and the remaining diagrams in which gluons are attached to other fermions are omitted .
but it is straightforward to attach two gluons to all the fermion lines making intermediate states off - shell . if collinear and soft gluons are attached to the heavy quark [ fig .
[ fig1 ] ( a ) , ( b ) ] , the intermediate heavy quark and gluon states are off the mass shell and we integrate them out .
for the collinear quark @xmath32 , the interaction with @xmath80or @xmath36 makes the collinear quark off the mass shell . and the interaction with @xmath33 or @xmath36 makes the @xmath78 field off the mass shell .
we add all the possible configurations in which the intermediate states become off the mass shell and collect the terms at leading order in @xmath26 .
let us introduce the factors @xmath118 where @xmath119 ( @xmath120 ) is the momentum of the collinear gluon @xmath33 ( @xmath80 ) , and @xmath121 is the soft momentum of the soft gluon @xmath36 .
the wilson lines corresponding to each type of gluons are obtained by exponentiating the above factors as @xmath122 , \ \
\overline{w}= \sum_{\mathrm{perm } } \exp \bigl
[ -g \frac{1}{n \cdot \mathcal{q } } n\cdot a_{\overline{n } } \bigr ] , \nonumber \\ \overline{s}&= & \sum_{\mathrm{perm } } \exp \bigl [ -g \frac{1}{\overline{n } \cdot \mathcal{r } } \overline{n}\cdot a_s \bigr ] , \
\ s= \sum_{\mathrm{perm } } \exp \bigl [ -g \frac{1}{n \cdot \mathcal{r}}n \cdot a_s \bigr ] .
\end{aligned}\ ] ] here @xmath123 ( @xmath124 ) is the label momentum operator for collinear fields in the @xmath15 ( @xmath17 ) direction , and the operator @xmath125 is the operator extracting the soft momentum from soft fields .
when we sum over all these diagrams , the color singlet operators of the form @xmath126 and the nonsinglet operators of the form @xmath127 are affected by the gauge fields differently and the final form of the four - quark operators can be written as @xmath128 where @xmath129 , @xmath130 ( @xmath131 ) are the color factors .
one interesting case arises when we attach @xmath33 and @xmath80 to a heavy quark . at leading order in @xmath26 , the amplitude in fig .
[ fig1 ] ( a ) , with @xmath33 and @xmath80 carrying the incoming momentum @xmath119 and @xmath120 respectively , is given by @xmath132 b,\end{aligned}\ ] ] and the amplitude for fig .
[ fig1 ] ( b ) with @xmath33 and @xmath80 is written as @xmath133 at first sight , these amplitudes contain complicated denominators which can not be expressed in terms of @xmath134 , @xmath0 , @xmath135 or @xmath136 .
however , if we add @xmath137 and @xmath138 , we obtain @xmath139 where there appear only @xmath134 and @xmath0 .
therefore the role of the triple gluon vertex is critical not only in determining the order of the wilson lines but also in making the final expression simple .
we will derive gauge - invariant operators using the auxiliary fields in appendix a , and this cancellation is important in constructing the lagrangian for the auxiliary fields . for singlet operators ,
when we sum over all the feynman diagrams , we have @xmath140_{\alpha \beta } , \
l^s_{\gamma\delta } = \delta_{\gamma\delta } , \label{dfsing}\ ] ] to order @xmath117 . here
we show only the products of two different gauge fields , since they indicate the ordering of the wilson lines . from eq .
( [ dfsing ] ) , @xmath141 suggests the wilson lines to be @xmath142 . for color nonsiglet operators
, we have @xmath143_{\gamma\beta } , \nonumber
\\ l^n_{\alpha\delta } & = & \bigl[g(-a+b - c+d ) \nonumber \\ & & + g^2 ( ac+db - cb - dc - ab - ad)\bigr]_{\alpha\delta } , \label{dfnon}\end{aligned}\ ] ] where @xmath144 suggests the wilson line in the order @xmath145 , and @xmath146 suggests the form @xmath147 .
therefore the gauge - invariant singlet and nonsinglet operators in @xmath5 are given by @xmath148\cdot \bigl[(\overline{\chi}\overline{w})_{\beta } \gamma_2 ( \overline{w}^{\dagger } \chi)_{\beta}\bigr ] , \nonumber \\ o_n & = & \bigl [ ( \overline{\xi } ws^{\dagger } \overline{s})_{\beta } \gamma_1 ( \overline{s}^{\dagger } h ) _ { \alpha } \bigr]\cdot \bigl [ ( \overline{\chi } \overline{w})_{\alpha } \gamma_2 ( \overline{w}^{\dagger } \chi)_{\beta } \bigr ] .
\label{gio2}\end{aligned}\ ] ] the operator @xmath149 can be written as @xmath150 , but this is identical to @xmath149 since @xmath151 .
all the four - quark operators for @xmath0 decays are of the form @xmath152 or @xmath149 with different dirac structure . and the form of the operators @xmath152 and @xmath149 in eq .
( [ gio2 ] ) manifestly shows the factorization of four - quark operators at leading order in scet and to all orders in @xmath1 . in the operators @xmath152 and @xmath149 , the interactions of @xmath33 and @xmath36 occur only in the heavy - to - light current sector , while the interactions of @xmath80 occur only in the light - to - light current sector .
if a collinear or a soft gluon is emitted from one sector and is absorbed by the other sector , the interaction vanishes , or if it does not vanish , the momentum transfer is of order @xmath153 , which is already integrated out to produce the wilson coefficients or jet functions .
that is , the gluon exchange between the two current sectors is not allowed , and the gluon exchange is allowed only in each sector .
due to this property , the form of the operators is preserved even though there are any possible exchange of gluons to all orders .
the wilson coefficients may be different at different orders of the perturbation theory .
this is an explicit proof of color transparency in scet , and we can safely calculate the matrix elements of the operators in terms of a product of the matrix elements of the two currents .
note that the terminology `` factorization '' is used in two ways .
first , it means that the matrix elements of the four - quark operators reduce to products of the matrix elements of the currents , and the matrix element can be written as @xmath154 this was first assumed in naive factorization .
it is possible only when there is no gluon exchange which connects the two currents and it is explicitly shown in scet to all orders in @xmath1 .
another meaning of factorization appears in high - energy processes in which a physical amplitude can be separated into a short - distance part and a long - distance part . for example , exclusive hadronic form factors at momentum transfer @xmath155 factor into nonperturbative light - cone wave functions @xmath156 for mesons , convoluted with a hard scattering kernel @xmath157 as @xcite @xmath158 we discuss both types of factorization in this paper .
so far , we have considered the factorization of matrix elements .
the second meaning of factorization will be discussed when we consider spectator contributions . by matching the four - quark operators in the full theory onto scet
, the effective hamiltonian for @xmath0 decays in scet can be written as @xmath159 , \label{scetham}\end{aligned}\ ] ] where @xmath160_{v - a } \bigl [ ( \overline{\chi}^d \overline{w})_{\beta } ( \overline{w}^{\dagger}\chi^u)_{\beta } \bigr]_{v - a } , \nonumber \\ o_{2r } & = & \bigl [ ( \overline{\xi}^u w s^{\dagger } \overline{s})_{\beta } ( \overline{s}^{\dagger } h)_{\alpha } \bigr]_{v - a } \bigl [ ( \overline{\chi}^d \overline{w } ) _ { \alpha } ( \overline{w}^{\dagger}\chi^u)_{\beta } \bigr]_{v - a } , \nonumber \\ o_{3r } & = & \bigl [ ( \overline{\xi}^d w)_{\alpha } ( s^{\dagger } h)_{\alpha } \bigr]_{v - a } \sum_q \bigl [ ( \overline{\chi}^q \overline{w})_{\beta } ( \overline{w}^{\dagger}\chi^q)_{\beta } \bigr]_{v - a } , \nonumber \\ o_{4r } & = & \bigl [ ( \overline{\xi}^d w s^{\dagger } \overline{s})_{\beta } ( \overline{s}^{\dagger } h)_{\alpha } \bigr]_{v - a } \sum_q \bigl [ ( \overline{\chi}^q \overline{w } ) _ { \alpha } ( \overline{w}^{\dagger}\chi^q)_{\beta } \bigr]_{v - a } , \nonumber \\ o_{5r } & = & \bigl [ ( \overline{\xi}^d w)_{\alpha } ( s^{\dagger } h)_{\alpha } \bigr]_{v - a } \sum_q \bigl [
( \overline{\chi}^q \overline{w})_{\beta } ( \overline{w}^{\dagger } \chi^q)_{\beta } \bigr]_{v+a } , \nonumber \\ o_{6r } & = & \bigl [ ( \overline{\xi}^d w s^{\dagger } \overline{s})_{\beta } ( \overline{s}^{\dagger } h)_{\alpha } \bigr)]_{v - a } \sum_q \bigl [ ( \overline{\chi}^q \overline{w } ) _ { \alpha } ( \overline{w}^{\dagger}\chi^q)_{\beta } \bigr]_{v+a } , \nonumber \\ o_{1c } & = & \bigl [ ( \overline{\xi}^d w s^{\dagger } \overline{s})_{\beta } ( \overline{s}^{\dagger } h)_{\alpha } \bigr]_{v - a } \bigl [ ( \overline{\chi}^u \overline{w } ) _ { \alpha } ( \overline{w}^{\dagger}\chi^u)_{\beta } \bigr]_{v - a } , \nonumber \\ o_{2c } & = & \bigl [ ( \overline{\xi}^d w)_{\alpha } ( s^{\dagger } h)_{\alpha } \bigr]_{v - a } \bigl [ ( \overline{\chi}^u \overline{w})_{\beta } ( \overline{w}^{\dagger}\chi^u)_{\beta } \bigr]_{v - a } , \nonumber \\ o_{3c } & = & \sum_q \bigl [ ( \overline{\xi}^q w s^{\dagger } \overline{s})_{\beta } ( \overline{s}^{\dagger } h)_{\alpha } \bigr]_{v - a } \bigl [ ( \overline{\chi}^d \overline{w } ) _ { \alpha } ( \overline{w}^{\dagger}\chi^q)_{\beta } \bigr]_{v - a } , \nonumber \\ o_{4c } & = & \sum_q \bigl [ ( \overline{\xi}^q w)_{\alpha } ( s^{\dagger } h)_{\alpha } \bigr]_{v - a } \bigl [
( \overline{\chi}^d \overline{w})_{\beta } ( \overline{w}^{\dagger } \chi^q)_{\beta } \bigr]_{v - a}. \label{effop}\end{aligned}\ ] ] here the summation over @xmath161 goes over to light massless quarks .
since we specify the direction of each quark , we have the original operators @xmath162 , obtained from eq .
( [ fullop ] ) , in which the light - to - light current forms a meson in the @xmath17 direction .
when a light quark in the heavy - to - light current moves in the @xmath17 direction and forms a meson with an antiquark in the light - to - light current , we use the fierz transformation first in the full theory and match the operators .
note that there are no operators @xmath163 and @xmath164 in scet , which are obtained by fierzing @xmath165 and @xmath166 .
neglecting color structure and the wilson lines , @xmath163 and @xmath164 have the form @xmath167 , which is identically zero at leading order in scet because @xmath168 since @xmath169 . in the literature @xcite , the effects of the operators @xmath163 and @xmath164 are sometimes known as chirally - enhanced contributions .
even though the effect is formally suppressed , the numerical values may not be negligible .
when the equation of motion for the currents is applied , there is an enhancement factor @xmath170 , where @xmath171 is a meson mass and @xmath172 is the current quark mass .
but in scet there is simply no such operator as @xmath163 and @xmath164 at leading order .
this contribution is formally suppressed in powers of @xmath26 in scet , but the coefficient of these operators can be large . in phenomenological applications , we have to know how to treat the chirally - enhanced contributions , but we will not consider them here .
the wilson coefficients of the four - quark operators in scet can be determined by matching the full theory onto scet .
we require that the matrix elements of an operator in the full theory be equal to the matrix elements of the corresponding operator in scet .
we use the @xmath173 scheme with naive dimensional regularization scheme and anticommuting @xmath174 .
all the external particles are on the mass shell . the radiative corrections for the four - quark operators in the full theory are shown in fig .
[ fig2 ] . as a specific example , the amplitudes for fig .
[ fig2 ] ( a ) to ( d ) for the operator @xmath175 are given as @xmath176 , \nonumber \\
im_b^{(1 ) } & = & \frac{\alpha_s}{4\pi } ( t_a)_{jk } ( t_a)_{il } ( \overline{d}_j u_k)_{v - a } ( \overline{u}_i b_l)_{v - a } \nonumber \\ & \times & \bigl [ -\frac{4}{\epsilon_{\mathrm{uv } } } + \frac{1}{\epsilon^2 } -\frac{2}{\epsilon } \bigl ( \ln \frac{x_2 m_b}{\mu } -1 \bigr ) -5 + 4\ln \frac{m_b}{\mu } + 2\ln^2 \frac{x_2 m_b}{\mu } \nonumber \\ & -&\frac{2}{1-x_2 } \ln x_2 + 2\mathrm{li}_2 ( 1-x_2)+\frac{\pi^2}{12 } \bigr ] , \nonumber \\
im_c^{(1 ) } & = & \frac{\alpha_s}{4\pi } ( t_a)_{jk } ( t_a)_{il } ( \overline{d}_j u_k)_{v - a } ( \overline{u}_i b_l)_{v - a } \nonumber \\ & \times & \bigl [ -\frac{4}{\epsilon_{\mathrm{uv } } } + \frac{2}{\epsilon^2 } + \frac{2}{\epsilon } \bigl(2-\ln \frac{-xx_1 m_b^2}{\mu^2 } \bigr ) -1 + \ln^2 \bigl ( \frac{-xx_1 m_b^2}{\mu^2 } \bigr ) -\frac{\pi^2}{6 } \bigr ] , \nonumber \\
im_d^{(1 ) } & = & \frac{\alpha_s}{4\pi } ( t_a)_{jk } ( t_a)_{il } ( \overline{d}_j u_k)_{v - a } ( \overline{u}_i b_l)_{v - a } \nonumber \\ & \times & \bigl [ \frac{1}{\epsilon_{\mathrm{uv } } } -\frac{2}{\epsilon^2 } -\frac{2}{\epsilon } \bigl(2 -\ln \frac{-xx_2 m_b^2}{\mu^2 } \bigr ) \nonumber \\ & -&8 + 3 \ln \frac{-xx_2m_b^2}{\mu^2 } -\ln^2 \frac{-xx_2 m_b^2}{\mu^2 } + \frac{\pi^2}{6 } \bigr ] , \label{mat1 } \end{aligned}\ ] ] where @xmath177 is the dilogarithmic function . and
@xmath178 , @xmath179 are the energy fractions given by @xmath180 , @xmath181 for @xmath182 .
the prescription for @xmath183 is given by @xmath184 .
if we add all these `` nonfactorizable '' contributions , the infrared divergence cancels and the only infrared divergence comes from the vertex corrections of the currents [ fig . [ fig2 ] ( e ) and ( f ) ] . since the vertex correction of the light - to - light current is the same both in the full theory and in scet , it cancels in matching . and fig .
[ fig2 ] ( f ) is given by @xmath185 .
\end{aligned}\ ] ] for @xmath186 , the corresponding amplitudes are given as @xmath187 , \nonumber \\
im_b^{(5 ) } & = & \frac{\alpha_s}{4\pi } ( t_a)_{jk } ( t_a)_{il } ( \overline{d}_i b_l ) _ { v - a } ( \overline{q}_j q_k)_{v+a } \nonumber \\ & \times & \bigl[-\frac{1}{\epsilon_{\mathrm{uv } } } + \frac{1}{\epsilon^2 } -\frac{2}{\epsilon } \bigl ( \ln \frac{x_2 m_b}{\mu } -1 \bigr ) + 1 - 2\ln \frac{m_b}{\mu } + 2\ln^2 \frac{x_2 m_b}{\mu } \nonumber \\ & & -\frac{2 - 3x_2}{1-x_2 } \ln x_2 + 2\mathrm{li}_2 ( 1-x_2 ) + \frac{\pi^2}{12 } \bigr ] , \nonumber
\\ im_c^{(5 ) } & = & \frac{\alpha_s}{4\pi } ( t_a)_{jk } ( t_a)_{il } ( \overline{d}_i b_l ) _ { v - a } ( \overline{q}_j q_k)_{v+a } \nonumber \\ & \times & \bigl[-\frac{1}{\epsilon_{\mathrm{uv } } } + \frac{2}{\epsilon^2 } + \frac{2}{\epsilon } \bigl ( 2-\ln \frac{-xx_1 m_b^2}{\mu^2 } \bigr ) + 5 -3 \ln \frac{-xx_1 m_b^2}{\mu^2 } \nonumber \\ & & + \ln^2 \frac{-xx_1 m_b^2}{\mu^2 } -\frac{\pi^2}{6 } \bigr ] , \nonumber \\
im_d^{(5 ) } & = & \frac{\alpha_s}{4\pi } ( t_a)_{jk } ( t_a)_{il } ( \overline{d}_i b_l ) _ { v - a } ( \overline{q}_j q_k)_{v+a } \nonumber \\ & \times & \bigl[\frac{4}{\epsilon_{\mathrm{uv } } } -\frac{2}{\epsilon^2 } -\frac{2}{\epsilon } \bigl ( 2-\ln \frac{-xx_2 m_b^2}{\mu^2 } \bigr ) -2 -\ln^2 \frac{-xx_2 m_b^2}{\mu^2 } + \frac{\pi^2}{6}\bigr ] , \label{mat5}\end{aligned}\ ] ] and @xmath188 is the same as @xmath189 except the dirac structure
. the infrared divergence of the nonfactorizabie contributions in eq .
( [ mat5 ] ) cancels , and the infrared divergence from the vertex corrections is cancelled in matching . using the color identity @xmath190 the operators in eq .
( [ mat1 ] ) becomes @xmath191 , and the operators in eq .
( [ mat5 ] ) becomes @xmath192 . in the radiative corrections of @xmath193 ,
the operator becomes @xmath194 due to the color factors in fig .
[ fig2](a ) and ( d ) , while it becomes @xmath195 from fig .
[ fig2](b ) and ( c ) .
therefore the radiative corrections for @xmath193 , or for nonsinglet operators in general , have infrared divergences in all the diagrams . the corresponding radiative corrections for the four - quark operators in scet are shown in fig . [ fig3 ] . for singlet operators , the radiative corrections exist only in the heavy - to - light current sector [ fig .
[ fig3 ] ( a ) , ( b ) ] , in which the infrared divergence appearing in the calculation is exactly the same as the infrared divergence in the full theory .
the diagram ( c ) is included , but since the result is the same in the full theory , it is cancelled in matching . for nonsinglet operators ,
we need all the diagrams in fig .
[ fig3 ] with additional collinear gluon interactions in the light - to - light current sector [ fig .
[ fig3 ] ( e ) and ( f ) ] and soft gluon interactions [ fig .
[ fig3 ] ( b ) and ( d ) ] .
when we calculate these diagrams , the infrared divergence is exactly the same as that in the full theory .
therefore we can safely match the full theory onto scet since the infrared divergence cancels , and the wilson coefficients can be calculated . in matching the theory at one loop ,
we calculate perturbative matrix elements in the full and effective theories .
all the long - distance physics is reproduced in the effective theory , and the difference between the two calculations determines the short - distance wilson coefficients .
we treat both ultraviolet and infrared divergences using dimensional regularization , with the final collinear quark on the mass shell . in this case
all the feynman diagrams in scet are proportional to @xmath196 .
the ultraviolet divergences are cancelled by the counterterms in scet , and all the infrared divergences cancel in the difference between the full theory and scet . therefore the wilson coefficients of various operators in scet can be obtained by calculating radiative corrections in the full theory . in the full theory
we also have to consider the feynman diagrams with fermion loops and the effect of the chromomagnetic operator .
these contributions are included in fig .
[ fig2 ] ( g ) and ( h ) .
the wilson coefficients are , in general , functions of the operators @xmath197 , @xmath198 and @xmath199 .
the matrix elements of the four - quark operators can be written in terms of a convolution as @xmath200 due to the momentum conservation @xmath178 , @xmath201 and @xmath202 satisfiy @xmath203 .
for nonleptonic decays into two light mesons at leading order in scet , we can set @xmath204 , @xmath205 , and @xmath206 , and the matrix element can be written as @xmath207 where @xmath208 with @xmath209 .
however , we list all the wilson coefficients for general @xmath178 , @xmath201 and @xmath202 , which will be useful for other decay modes , or nonleptonic @xmath0 decays at subleading order in scet . by adding the wave function renormalization of the heavy quark , the wilson coefficients at next - to - leading order are given as @xmath210 c_1 + \frac{a_3}{2 } c_2 , \
c_{2r } = \frac{a_1}{2 } c_1 + \bigl [ 1-\frac{a_3}{2n } + c_f a_4 \bigr ] c_2 , \nonumber \\ c_{3r}^p & = & \bigl[1-\frac{a_1}{2n } + c_f a_2 \bigr ] c_3 + \frac{a_3}{2 } c_4 , \ c_{4r}^p = \frac{a_1}{2 } c_3 + \bigl [ 1-\frac{a_3}{2n } + c_f a_4 \bigr ] c_4 , \nonumber \\ c_{5r}^p & = & \bigl [ 1-\frac{a_5}{2n } + c_f a_2 \bigr ] c_5 + \frac{a_6}{2 } c_6 , \
c_{6r}^p = \frac{a_5}{2 } c_5 + \bigl[1-\frac{a_6}{2n } + c_f a_7 \bigr ] c_6 , \nonumber \\ c_{1c}&= & \bigl [ 1-\frac{a_3}{2n } + c_f a_4 \bigr ] c_1 + \frac{a_1}{2 } c_2 , \
c_{2c } = \frac{a_3}{2 } c_1 + \bigl[1-\frac{a_1}{2n } + c_f a_2 \bigr ] c_2\nonumber \\
c_{3c}^p&=&\bigl [ 1-\frac{a_3}{2n } + c_f a_4 \bigr ] c_3 + \frac{a_1}{2 } c_4-\frac{1}{2n } c_l^p , \nonumber \\ c_{4c}^p & = & \frac{a_3}{2 } c_3 + \bigl[1-\frac{a_1}{2n } + c_f a_2 \bigr ] c_4 + \frac{1}{2 } c_l^p , \label{wilco}\end{aligned}\ ] ] where @xmath211 s are the wilson coefficients from the full theory . and the coefficients @xmath212 to order @xmath1 , evaluated at @xmath213 are given as @xmath214 \nonumber
\\ a_2&= & \frac{\alpha_s}{4\pi } \bigl[-5 -2 \ln^2 x + \frac{2-x}{1-x } \ln x -2 \mathrm{li}_2 ( 1-x ) -\frac{\pi^2}{12 } \bigr ] \nonumber \\ a_3 & = & \frac{\alpha_s}{4\pi } \bigl [ -9 + \frac{2-x}{1-x } \ln x -2 \mathrm{li}_2 ( 1-x)+2\ln^2 x_2 + 2\ln^2 ( -x_1 ) -\ln^2 \frac{-x_1}{x}\nonumber \\ & & -\frac{2}{1-x_2 } \ln x_2 + 2 \mathrm{li}_2(1-x_2 ) -\frac{\pi^2}{6}\bigr ] , \nonumber \\ a_4 & = & \frac{\alpha_s}{4\pi } \bigl[-14 -2\ln^2 x_1 -\ln^2 ( -xx_2 ) + 3\ln ( -xx_2 ) + \frac{2 - 3x_1}{1-x_1 } \ln x_1 \nonumber \\ & & -2\mathrm{li}_2 ( 1-x_1 ) + \frac{\pi^2}{12 } \bigr ] , \nonumber \\ a_5 & = & \frac{\alpha_s}{4\pi } \bigl[6 -3 \ln ( - x ) + \ln \frac{x^2}{x_1 x_2 } \ln \frac{x_1}{x_2 } -\frac{1 - 3x_1}{1-x_1 } \ln x_1 -\frac{2 - 3x_2}{1-x_2 } \ln x_2 \nonumber \\ & & -2 \mathrm{li}_2 ( 1-x_1 ) + 2 \mathrm{li}_2 ( 1-x_2 ) \bigr ] , \nonumber \\ a_6 & = & \frac{\alpha_s}{4\pi } \bigl[3 + 2\ln^2 x_2 + 2\ln^2 ( -x_1 ) -\ln^2 \frac{-x_1}{x } -\frac{1 - 2x}{1-x } \ln x -3\ln ( -x_1)\nonumber \\ & & -\frac{2 - 3x_2}{1-x_2 } \ln x_2 -2 \mathrm{li}_2 ( 1-x ) + 2 \mathrm{li}_2 ( 1-x_2 ) -\frac{\pi^2}{6 } \bigr ] , \nonumber \\ a_7 & = & \frac{\alpha_s}{4\pi } \bigl [ -2 -2\ln^2 x_1 -\ln^2 ( -xx_2 ) + \frac{2}{1-x_1 } \ln x_1 \nonumber \\ & & -2\mathrm{li}_2 ( 1-x_1 ) + \frac{\pi^2}{12 } \bigr ] . \end{aligned}\ ] ] and the contribution @xmath215 from fermion loops and the chromomagnetic operator [ fig . [ fig2 ] ( g ) and ( h ) ] at @xmath213 are given by @xmath216 , \label{clp}\end{aligned}\ ] ] where @xmath217 ( @xmath218 ) , and @xmath219 is given by @xmath220 we can add @xmath221 in @xmath222 and @xmath223 , and @xmath224 in @xmath225 and @xmath226 with @xmath227 in eq .
( [ clp ] ) .
but in physical processes in which the final - state consists of color singlet mesons , the contribution of @xmath215 cancels , hence omitted in eq .
( [ wilco ] ) .
in nonleptonic @xmath0 decays , we also have to consider the spectator quark which can interact with the @xmath228 quark or other quarks forming light mesons . in this section ,
we concentrate on the spectator quark interacting with the quark and antiquark pair ( @xmath78 fields ) produced by the weak current , which we call nonfactorizable spectator contributions .
the spectator quark can interact with the heavy - to - light current , which will be treated in the next section . here
we apply the two - step matching explicitly .
the time - ordered products are constructed in @xmath2 , and they are evolved down to @xmath5 .
nonfactorizable spectator contributions arise from the interactions of the gluons @xmath33 in the light - to - light current with a spectator quark , which becomes a collinear quark as a result .
however the operators @xmath152 and @xmath149 at leading order in scet do not involve the interaction of @xmath33 .
therefore we take into account subleading operators which involve @xmath33 in the light - to - light current . and
the lagrangian describing the interaction of collinear and usoft quarks begins with @xmath229 compared to the leading collinear lagrangian .
but the propagator of the exchanged gluon is of order @xmath230 .
therefore the nonfactorizable spectator contribution is of the same order as the leading contributions from the four - quark operators . in order to evaluate decay amplitudes at leading order in scet , we need to include the nonfactorizable spectator contributions , as shown in fig .
[ fig4 ] . here
we consider a collinear gluon @xmath33 from the light - to - light current interacting with a spectator quark to produce a collinear quark @xmath32 .
the lagrangian for the collinear and usoft quarks at order @xmath41 is given by @xcite @xmath231 w^{\dagger } \xi + \mathrm{h.c.}.\ ] ] in @xmath2 , there are two independent subleading operators suppressed by @xmath41 , in which collinear gluons @xmath33 come from the current @xmath232 .
they are given by @xmath233\bigl)_{\alpha } \gamma_{2i } ( \overline{w}^{\dagger } \chi)_{\beta } \biggr\ } \nonumber \\ o^{(1b)}_i & = & \bigl ( ( \overline{\xi}w)_{\beta } \gamma_{1i } h_{\alpha } \bigr ) \biggl\ { ( \overline{\chi } \overline{w})_{\alpha } \gamma_{2i } \bigl ( \bigl [ w^{\dagger } i\overrightarrow{\fmslash{d}}_{n\perp } w\bigr ] \frac{\fmslash{n}}{2 } \frac{1}{n\cdot \mathcal{q } } \overline{w}^{\dagger } \chi \bigr)_{\beta } \biggr\ } , \label{o1ab}\end{aligned}\ ] ] where the index @xmath234 runs over all the possible forms of the operators as shown in eq .
( [ effop ] ) .
the wilson coefficients of these operators are 1 at lowest order in @xmath1 .
the form of the operators @xmath235 can be obtained in a straightforward manner , but the ordering of the wilson lines is nontrivial . it will be explained using the auxiliary field method in appendix b. in @xmath2 , the nonfactorizable spectator contribution is given by the matrix elements of the time - ordered product @xmath236.\ ] ] in order to go down to @xmath5 , we decouple the collinear - usoft interaction using the field redefinitions @xcite @xmath237 the collinear effective lagrangian can be written in terms of @xmath63 , @xmath238 and @xmath64 , @xmath239 , that is , the collinear - usoft interactions are factorized with the field redefinitions given in eq .
( [ redef ] ) . matching at @xmath240 , the small expansion parameter changes from @xmath241 to @xmath242 with a definite power counting procedure for any operators .
recall that we rename the usoft field of order @xmath243 as the soft field in @xmath5 . in that sense
the usoft fields become soft and the wilson line @xmath244 becomes the wilson line with soft gluons in @xmath5 but `` soft '' means the momentum of order @xmath245 . without introducing addtional wilson lines
, we denote this wilson line as @xmath246 and in the matching we replace @xmath247 without any ambiguity . and
the operators are matched onto the operators in @xmath5 .
the nonfactorizable spectator contribution comes from the matrix elements of six - quark operators . in calculating the matrix elements of @xmath248 , we first factorize the usoft - collinear interactions using eq .
( [ redef ] ) with @xmath247 , @xmath249 , and project out color indices in such a way that the quark bilinears ( @xmath250 , @xmath251 ) and ( @xmath252 , @xmath253 ) forming mesons are color singlets .
since the final expression involving the wilson lines are nontrivial , we show explicitly how the color projection is performed for @xmath254 . at order @xmath1
, we will extract a collinear gluon from the factor ( @xmath255 ) in @xmath256 and @xmath254 to contract them . neglecting the dirac structure and keeping color - dependent parts only , the time - ordered product of @xmath254 with @xmath257 has the form @xmath258 \bigl [ \bigl(\overline{\chi } \overline{w } \overline{s}^{\dagger } s t_a s^{\dagger } \bigr)_{\alpha } \cdot ( \overline{s } \overline{w}^{\dagger } \chi ) _ { \beta } \bigr ] \bigl [ ( \overline{q}_{s } s)_{\gamma } ( t_a)_{\gamma \delta } ( w^{\dagger}\xi ) _ { \delta } \bigr ] \nonumber \\ & = & \bigl [ ( \overline{\xi } w)_{\mu } ( s^{\dagger})_{\mu \beta } \cdot h_{\alpha } \bigr ] \bigl [ ( \overline{\chi } \overline{w } ) _ { \nu } \bigl ( \overline{s}^{\dagger } st_a s^{\dagger } \bigr)_{\nu\alpha } \cdot ( \overline{s})_{\beta \rho } ( \overline{w}^{\dagger } \chi ) _ { \rho } \bigr ] \nonumber \\ & & \times \bigl [ ( \overline{q}_{s } s)_{\gamma } ( t_a)_{\gamma \delta } ( w^{\dagger}\xi ) _ { \delta } \bigr ] \nonumber \\ & = & \bigl [ ( \overline{\xi } w)_{\mu } \cdot h_{\alpha } \bigr ] \bigl [ ( \overline{\chi } \overline{w } ) _ { \nu } \cdot ( \overline{w}^{\dagger } \chi ) _ { \rho } \bigr ] \bigl [ ( \overline{q}_{s } s)_{\gamma } \cdot ( w^{\dagger}\xi ) _ { \delta } \bigr ] \nonumber \\ & & \times \bigl ( \overline{s}^{\dagger } st_a s^{\dagger } \bigr)_{\nu\alpha}(s^{\dagger}\overline{s})_{\mu \rho } ( t_a)_{\gamma \delta } \nonumber \\ & \longrightarrow & \bigl [ ( \overline{\xi } w ) _ { \beta } \cdot h_{\alpha } \bigr ] \bigl [ ( \overline{\chi}\overline{w } ) _ { \nu } \cdot ( \overline{w}^{\dagger } \chi)_{\nu } \bigr ] \bigl [ ( \overline{q}_{s } s)_{\gamma } \cdot ( w^{\dagger } \xi)_{\beta } \bigr ] \nonumber \\ & & \times \frac{1}{n^2 } ( s^{\dagger } \overline{s})_{\mu \rho } ( \overline{s}^{\dagger } st_a s^{\dagger})_{\rho \alpha } ( t_a)_{\gamma\mu } \nonumber \\ & = & \frac{c_f}{n^2 } \bigl [ ( \overline{\xi } w ) _ { \beta } \cdot ( s^{\dagger}h ) _ { \alpha } \bigr ] \bigl [ ( \overline{\chi}\overline{w } ) _ { \nu } \cdot ( \overline{w}^{\dagger } \chi)_{\nu } \bigr ] \bigl [ ( \overline{q}_{s } s)_{\alpha } \cdot ( w^{\dagger } \xi)_{\beta } \bigr ] , \label{cpro}\end{aligned}\ ] ] where the dots denote the dirac structure and the color projection is performed after the arrow in eq . ( [ cpro ] ) . the ime - ordered product with @xmath259 can be projected in the same way .
the time - ordered products of the operators @xmath260 with the color singlet structure vanish due to the color structure .
the matrix element of the time - ordered product @xmath261 at order @xmath1 is given by @xmath262 \bigl[(\overline{\chi } \overline{w } ) _ { \gamma } \frac{1}{n\cdot \mathcal{q}^{\dagger } } \frac{\fmslash{n}}{2 } \gamma_{\perp}^{\mu } \gamma_{2i } ( \overline{w}^{\dagger } \chi)_{\gamma } \bigr ] \nonumber \\ & & \times \bigl [ ( \overline{q}_s s)_{\alpha } ( \overline{n}\cdot x ) \frac{1}{n\cdot \mathcal{r}^{\dagger } } \gamma_{\mu}^{\perp}\frac{1}{\overline{n}\cdot \mathcal{p } } ( w^{\dagger } \xi)_{\beta } ( 0 ) \bigr ] \rangle \nonumber \\ & + & \langle \bigl[(\overline{\xi } w)_{\beta } \gamma_{1i } ( s^{\dagger } h)_{\alpha } \bigr ] \bigl[(\overline{\chi } \overline{w } ) _ { \gamma } \gamma_{2i } \gamma_{\perp}^{\mu } \frac{\fmslash{n}}{2 } \frac{1}{n\cdot \mathcal{q } } ( \overline{w}^{\dagger } \chi)_{\gamma } \bigr ] \nonumber \\ & & \times \bigl [ ( \overline{q}_s s)_{\alpha}(\overline{n}\cdot x ) \frac{1}{n\cdot \mathcal{r}^{\dagger}}\gamma_{\mu}^{\perp } \frac{1}{\overline{n}\cdot \mathcal{p } } ( w^{\dagger } \xi)_{\beta } ( 0)\bigr ] \rangle \biggr\}. \label{spec}\end{aligned}\ ] ] let us consider in detail the matrix elements in eq .
( [ spec ] ) for different dirac structure @xmath263 . for simplicity
we omit all the wilson lines and the momentum operators in the following calculation . for @xmath264
, we can evaluate the first term in the curly bracket in eq .
( [ spec ] ) as @xmath265 in the first line we can replace @xmath266 by @xmath267 and we use the fierz transformation to arrive at the last relation . the first term in the last relation in eq .
( [ matcal ] ) can be further simplified as @xmath268 \xi_{\beta } \cdot \overline{\chi } \fmslash{n } ( 1-\gamma_5 ) \chi \nonumber \\ & & = \frac{1}{2 } \overline{\xi } \gamma^{\mu } ( 1-\gamma_5 ) \xi \cdot \overline{q } \gamma_{\mu } ( 1-\gamma_5 ) h \cdot \overline{\chi } \fmslash{n } ( 1-\gamma_5 ) \chi \nonumber \\ & & = \frac{1}{4 } \overline{\xi } \fmslash{\overline{n } } ( 1-\gamma_5 ) \xi \cdot \overline{q } \fmslash{n } ( 1-\gamma_5 ) h \cdot \overline{\chi } \fmslash{n } ( 1-\gamma_5 ) \chi .
\label{firterm}\end{aligned}\ ] ] the part proportional to @xmath269 vanishes when we apply the fierz transformation . in the third line we use the fierz transformation for the product of the first two currents .
similarly , the second term in eq .
( [ matcal ] ) is simplified as @xmath270 where the fierz transformations are applied successively . when we add eqs .
( [ firterm ] ) and ( [ secterm ] ) , they cancel . the second term in eq .
( [ spec ] ) is given by eq .
( [ secterm ] ) with an opposite sign . as a result , for @xmath271 ,
only the second term in eq .
( [ spec ] ) contributes and the matrix element is given as @xmath272 note that the matrix element of the six - quark operators in eq .
( [ me1 ] ) is written in a factorized form .
it is because a collinear or a soft gluon in one sector can not interact with the particles in the other sector .
that type of interaction makes the intermediate states off the mass shell by @xmath112 or @xmath115 .
these off - shell modes are already taken care of in the wilson coefficients or in the jet functions .
therefore the six - quark operators describing the spectator interactions are factorized and the matrix elements of the six - quark operators can be computed by the products of the matrix elements of each current .
for @xmath273 , only the first term in eq .
( [ spec ] ) contributes and the matrix element is given by @xmath274 and for @xmath275 , the matrix element is given as @xmath276 as can be clearly seen in the final expressions in eqs .
( [ me1])([me3 ] ) , the gluons @xmath33 , @xmath80 , and @xmath36 can be exchanged only inside each meson .
this is true at higher orders of @xmath1 .
though the wilson coefficients can be different , the structure of the operator is the same to all orders in @xmath1 .
therefore the nonfactorizable spectator contribution is factorized to all orders in @xmath1 .
let us calculate the matrix element @xmath248 explicitly for @xmath277 .
it is given by @xmath278 where we integrate out @xmath279 and @xmath280 explicitly .
the matrix element involving the @xmath0 meson can be calculated as @xmath281 \nonumber \\ & & = -\frac{if_b m_b}{4 } \int dr_+ e^{-ir_+ \overline{n } \cdot x/2 } \mathrm{tr}\ \bigl [ \frac{1+\fmslash{v}}{2 } \fmslash{\overline{n } } \gamma_5 \fmslash{n } ( 1-\gamma_5 ) \bigr ] \phi_b^+ ( r_+ ) \nonumber \\ & & = -if_b m_b \int dr_+ e^{-ir_+ \overline{n } \cdot x/2 } \phi_b^+ ( r_+),\end{aligned}\ ] ] where the leading - twist @xmath0 meson light - cone wave function is defined through the projection of the @xmath0 meson as @xcite @xmath282.\ ] ] and the light - cone wave function for the light mesons is defined as @xmath283 \phi_{m2 } ( u).\end{aligned}\ ] ] for @xmath284 , the matrix element which is given by @xmath285 for @xmath286 , the matrix element is zero if we use the leading - twist @xmath0 meson wave function because @xmath287 \nonumber \\ & = & -\frac{if_b m_b}{8 } \phi_b^+ \mathrm{tr } \
\fmslash{v } \fmslash{\overline{n } } \fmslash{n } \gamma_{\perp}^{\mu } = 0.\end{aligned}\ ] ] if we use higher - twist wave function for the @xmath0 meson , there may be nonzero contributions , but this is expected to be suppressed .
in section [ sec5 ] , we have considered the nonfactorizable spectator contributions arising from the subleading operators in which we include only the subleading part from the light - to - light current . however , there is also a subleading operator coming from the heavy - to - light current , but this contributes to the heavy - to - light form factor .
it is considered first in ref .
@xcite , and we discuss in detail here in the context of nonleptonic @xmath0 decays .
the operators in @xmath2 , which contribute to the form factor , are given by @xmath288 here we list only singlet operators .
the nonsinglet operators give the same matrix elements as the singlet operators with the color suppression factor @xmath289 .
we will consider only singlet operators from now on .
the interaction of collinear and usoft quarks is given by @xcite @xmath290 where @xmath291 , \ ig \fmslash{m } = [ i\overline{n } \cdot d^n , i\fmslash{d}^{us } + \frac{\fmslash{\overline{n}}}{2 } gn\cdot a_n].\ ] ] at leading order in scet , the relevant time - ordered products are given as @xmath292 , \
t_{1i}^f = \int d^4 x t[j^{(1a)}_i ( 0 ) i\mathcal{l}_{\xi q}^{(1 ) } ( x ) ] , \nonumber \\
t_{2i}^f & = & \int d^4 x t[j^{(1b)}_i ( 0 ) i\mathcal{l}_{\xi q}^{(1 ) } ( x ) ] , \ t_{3i}^f = \int d^4 x t[j^{(0)}_i ( 0 ) i\mathcal{l}_{\xi q}^{(2b ) } ( x ) ] , \nonumber \\
t_{4i}^{nf } & = & \int d^4 x t[j^{(0)}_i ( 0 ) i\mathcal{l}_{\xi q}^{(2a ) } ( x ) ] , \nonumber \\
t_{5i}^{nf } & = & \int d^4 x d^4 y t[j^{(0)}_i ( 0 ) i\mathcal{l}_{\xi \xi}^{(1 ) } ( x ) i\mathcal{l}_{\xi q}^{(1 ) } ( y ) ] , \nonumber \\
t_{6i}^{nf } & = & \int d^4 x d^4 y t[j^{(0)}_i ( 0 ) i\mathcal{l}_{cg}^{(1 ) } ( x ) i\mathcal{l}_{\xi q}^{(1 ) } ( y ) ] , \label{formtime}\end{aligned}\ ] ] where @xmath293 is the leading collinear lagrangian and @xmath294 is the subleading gluon lagrangian , which can be found in refs .
the feynman diagrams at lowest order in @xmath1 for the time - ordered products are shown in fig .
[ fig5 ] schematically .
compared to the case of the nonfactorizable spectator interactions , the leading - order heavy - to - light current @xmath295 contains @xmath33 , therefore it contributes to the heavy - to - light form factor starting from the leading order . as suggested in ref .
@xcite , we absorb the nonfactorizable parts @xmath296 ( @xmath297 ) into the definition of the soft nonperturbative effects for the form factors at this order . among the factorizable contributions @xmath298 , only @xmath299
is nonzero at order @xmath1 . for @xmath300 , the matrix element of @xmath299
is given by @xmath301|b \rangle \nonumber \\ & = & \frac{g^2}{4\pi } \frac{1}{m_b \overline{n } \cdot p_1 } \int d\overline{n } \cdot x \int \frac{dn\cdot k}{n\cdot k } e^{in\cdot k \overline{n } \cdot x/2 } \nonumber \\ & \times & \frac{c_f}{4n } \langle \overline{\xi } w \fmslash{\overline{n } } ( 1-\gamma_5 ) w^{\dagger } \xi \cdot \overline{\chi } \overline{w } \fmslash{n } ( 1-\gamma_5 ) \overline{w}^{\dagger } \chi \cdot \overline{q}_s s ( \overline{n } \cdot x ) \fmslash{n } ( 1-\gamma_5 ) s^{\dagger } h \rangle \nonumber \\ & = & \frac{\alpha_s}{4\pi } \frac{4\pi^2 c_f}{n } if_{m_1 } f_{m2 } f_b \frac{2e}{m_b } m_b \int du \frac{\phi_{m_1 } ( u)}{u } \int \frac{dr_+}{r_+ } \phi_b^+ ( r_+ ) , \label{t1mat}\end{aligned}\ ] ] where we use fierz transformations successively . for @xmath302 , we obtain the same result as in eq .
( [ t1mat ] ) . for @xmath303
, it vanishes .
the form factors for @xmath0 decays into light pseudoscalar mesons are defined as @xmath304 \nonumber \\ & & + f_0 ( q^2 ) \frac{m_b^2-m_p^2}{q^2 } q^{\mu},\end{aligned}\ ] ] where @xmath305 . in scet , the form factor at order @xmath1 is given by @xmath306 where @xmath307 is a nonperturbative function introduced in ref .
@xcite in large - energy effective theory @xcite .
a similar nonperturbative function is introduced in refs .
@xcite , based on scet .
the procedure in obtaining independent nonperturabtive functions in the form factor is different in the large - energy effective theory and in scet , but the number of independent nonperturbative functions is the same . and @xmath308 at @xmath213 is given by @xmath309 the expression for @xmath310 coincides with the result in ref .
@xcite with @xmath311 at order @xmath1 , and @xmath308 is calculated in refs .
as an application , we consider the decay amplitudes for @xmath312 , which can be written as @xmath313 where the operators @xmath314 are given by @xmath315 + a_2^p \bigl [ ( \overline{\xi}^d h)_{v - a } ( \overline{\chi}^u \chi^u ) _ { v - a}\bigr ] \nonumber \\ & & + a_3^p \bigl [ ( \overline{\xi}^d h)_{v - a } ( \overline{\chi}^q \chi^q ) _ { v - a } \bigr ] + a_4^p \bigl [ ( \overline{\xi}^q h)_{v - a } ( \overline{\chi}^d \chi^q ) _ { v - a } \bigr ] \nonumber \\ & & + a_5^p \bigl[(\overline{\xi}^d h)_{v - a } ( \overline{\chi}^q \chi^q ) _ { v+a } \bigr].\end{aligned}\ ] ] the notation @xmath316 $ ] means that @xmath317 are the sum of the amplitudes initiated by the operator @xmath318 .
they include the contribution from the operator itself , and the spectator contributions .
that is , @xmath316 $ ] can be written as @xmath319 = t_i^p + n_i^p + f_i^p,\ ] ] where @xmath320 is the contribution from the four - quark operators , @xmath321 is the nonfactorizable spectator contribution , and @xmath322 is the spectator contribution for the heavy - to - light form factor with @xmath323 .
before we present the decay amplitudes for @xmath324 at order @xmath1 explicitly , we show the amplitudes @xmath317 in general to all orders in @xmath1 , which are derived from scet . the form of @xmath320 from eq .
( [ fqop ] ) can be written as @xmath325 where the upper ( lower ) sign corresponds to @xmath326 ( @xmath327 ) .
the effective wilson coefficients @xmath328 are evaluated at @xmath329 and evolved down to @xmath330 to be matched onto @xmath5 .
then they evolve down to the scale @xmath8 , where the matrix elements are evaluated .
the matrix element of the heavy - to - light current is given by @xcite @xmath331 the nonperturbative parameter @xmath332 is matched at @xmath333 and it evolves down to @xmath8 , and the wave functions are evaluated at @xmath8 .
note that we do not include the radiative corrections in @xmath332 , since they are taken into account either in the wilson coefficients or in the spectator contributions .
the spectator contribution @xmath321 involves the time - ordered products , which can be written as @xmath334 & = & \delta \bigl ( \frac{n\cdot x}{2 } \bigr ) \delta^2 ( x_{\perp } ) \int d\overline{\eta } d\eta \int dr_+ e^{ir_+ \overline{n } \cdot x/2 } \nonumber \\ & \times & j^n_i ( \eta , \overline{\eta } , r_+ ) \mathcal{o}_i ( \eta,\overline{\eta},r_+),\end{aligned}\ ] ] where @xmath335 are the jet functions , which are obtained by matching @xmath2 onto @xmath5 .
the index @xmath234 denotes all the possible operators for the decay . and
the operator @xmath336 is given by @xmath337 therefore @xmath321 can be written as @xmath338 \nonumber \\ & = & \int du dv dr_+ c_{\mathrm{eff},i}^{n } ( \mu_0 , \mu ) j^n_i ( u , v , r_+ , \mu_0 , \mu ) n_i f_b f_{m1 } f_{m_2 } \nonumber \\ & & \times \phi_{m1 } ( u , \mu ) \phi_{m2 } ( v,\mu ) \phi_b^+ ( r_+,\mu ) , \label{gen2 } \end{aligned}\ ] ] where @xmath339 are the effective wilson coefficients , and the @xmath8 dependence is shown explicitly . here
@xmath340 are the normalization constants when we evaluate the matrix elements of @xmath336 . since @xmath341
, the variables @xmath342 and @xmath343 satisfy the relations @xmath344 the factorizable spectator contribution to the form factor can be written similarly as @xmath345 where @xmath346 are the effective wilson coefficients , and @xmath347 are the jet functions obtained from matching @xmath2 onto @xmath5 .
the spectator interactions @xmath321 in eq .
( [ gen2 ] ) and @xmath322 in eq .
( [ gen3 ] ) are factorized into the short - distance and the long - distance parts to all orders in @xmath1 , and the convolution integrals are finite .
now let us calculate the decay amplitudes for @xmath312 at order @xmath1 , based on the general expressions on @xmath320 , @xmath321 and @xmath322 .
the contributions @xmath320 are obtained by the convolutions of the following effective wilson coefficients @xmath348 with the hadronic matrix elements of the four - quark operators . for the decay
amplitudes at leading order , we can put @xmath204 and let @xmath349 and @xmath350 , and the amplitudes @xmath320 s , evaluated at @xmath213 are given as @xmath351 , \nonumber \\
t_2^p & = & im_b^2 \zeta f_{\pi } \bigl [ \bigl(c_2 + \frac{c_1}{n}\bigr ) ( 1+k ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_1 f \bigr],\nonumber \\ t_3^p & = & im_b^2 \zeta f_{\pi } \bigl [ \bigl(c_3 + \frac{c_4}{n}\bigr ) ( 1+k ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_4 f \bigr ] , \nonumber \\ t_4^p & = & im_b^2 \zeta f_{\pi } \biggl\ { \bigl[\bigl(c_4 + \frac{c_3}{n}\bigr ) ( 1+k ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } \bigl [ c_3 f + c_1 \bigl ( \frac{2}{3 } -g(s_p ) \bigr ) \nonumber \\ & + & c_3\bigl ( \frac{4}{3 } -g(0 ) -g(1 ) \bigr ) + c_4 \bigl ( \frac{2n_f}{3 } -3g(0 ) -g(s_c ) -g(1 ) \bigr ) \nonumber \\ & + & c_6 \bigl ( -3g(0 ) -g(s_c ) -g(1 ) \bigr ) + g_{\pi,8 } ( c_5+c_8 ) \bigr ] \biggr\ } , \nonumber \\ t_5^p & = & -im_b^2 \zeta f_{\pi } \bigl [ \bigl(c_5 + \frac{c_6}{n}\bigr ) ( 1+k ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_6 ( -f-12 ) \bigr],\end{aligned}\ ] ] where @xmath352 , @xmath353 , @xmath354 . and @xmath355 is given as @xmath356.\end{aligned}\ ] ] the nonfactorizable spectator contributions @xmath321 are given by eq .
( [ specc ] ) as @xmath357 where the effective wilson coefficients are given by @xmath358 the spectator contributions to the heavy - to - light form factor @xmath322 are given as @xmath359 where the effective wilson coefficients are given by @xmath360 the final expression can be simplied when we use the definition of @xmath310 given in eq .
( [ fplus ] ) .
for example , @xmath361 is written as @xmath362 , \nonumber \\ & & + i \frac{c_f}{n }
\pi \alpha_s f_{\pi}^2 f_b m_b \bigl(c_1 + \frac{c_2}{n}\bigr ) \int du\frac{\phi_{\pi } ( u)}{u } \int dr_+ \frac{\phi_b^+ ( r_+)}{r_+ } \nonumber \\ & = & im_b^2 \zeta f_{\pi } \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_2 f + im_b^2 f_{\pi } \bigl ( c_1 + \frac{c_2}{n } \bigr ) \bigl [ \zeta ( 1+k ) \nonumber \\ & & + \pi \alpha_s \frac{c_f}{n } \frac{f_{\pi } f_b}{m_b } \int du\frac{\phi_{\pi } ( u)}{u } \int dr_+ \frac{\phi_b^+ ( r_+)}{r_+ } \bigr ] \nonumber \\ & & \approx im_b^2 f_+ f_{\pi } \bigl[\bigl(c_1 + \frac{c_2}{n}\bigr ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_2 f \bigr],\end{aligned}\ ] ] where the definition of @xmath310 in eq .
( [ fplus ] ) is used in the last line with @xmath363 .
and we replace @xmath332 by @xmath310 in the term proportional to @xmath355 .
this induces terms of @xmath364 , which is neglected .
this relation also holds for the combinations @xmath365 for all @xmath234 . in summary , the decay amplitudes @xmath317
are given by @xmath366 , \nonumber \\
a_2^p & = & im_b^2 f_+ f_{\pi } \bigl [ \bigl(c_2 + \frac{c_1}{n}\bigr ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_1 f^{\prime } \bigr],\nonumber \\ a_3^p & = & im_b^2 f_+ f_{\pi } \bigl [ \bigl(c_3 + \frac{c_4}{n}\bigr ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_4 f^{\prime } \bigr ] , \nonumber \\
a_4^p & = & im_b^2 f_+ f_{\pi } \biggl\ { \bigl[\bigl(c_4 + \frac{c_3}{n}\bigr ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } \bigl [ c_3 f^{\prime } + c_1 \bigl ( \frac{2}{3 } -g(s_p ) \bigr ) \nonumber \\ & + & c_3\bigl ( \frac{4}{3 } -g(0 ) -g(1 ) \bigr ) + c_4 \bigl ( \frac{2n_f}{3 }
-3g(0 ) -g(s_c ) -g(1 ) \bigr ) \nonumber \\ & + & c_6 \bigl ( -3g(0 ) -g(s_c ) -g(1 ) \bigr ) + g_{\pi,8 } ( c_5+c_8 ) \bigr ] \biggr\ } , \nonumber \\ a_5^p & = & -im_b^2 f_+ f_{\pi } \bigl [ \bigl(c_5 + \frac{c_6}{n}\bigr ) + \frac{\alpha_s}{4\pi } \frac{c_f}{n } c_6 ( -f^{\prime}-12 ) \bigr ] , \label{final}\end{aligned}\ ] ] where @xmath367 the decay amplitudes in eq .
( [ final ] ) at order @xmath1 are consistent with the result obtained by beneke et al .
@xcite at order @xmath1 .
however , our result goes further in the sense that we consider the operators to all orders in @xmath1 , as shown in eqs .
( [ gen1 ] ) , ( [ gen2 ] ) and ( [ gen3 ] ) . at higer orders ,
our result will be different from the result in ref .
@xcite since there are two scales @xmath368 and @xmath369 involved and the effects of these two scales can not be reproduced in the heavy quark limit .
we have considered the four - quark operators relevant to nonleptonic @xmath0 decays into two light mesons in scet at leading order in @xmath26 and to next - to - leading order in @xmath1 .
the construction of the four - quark operators in scet is process - dependent since we first have to specify the directions of the outgoing quarks and construct the operators accordingly . in matching onto @xmath5
, we integrate out off - shell modes by attaching collinear and soft gluons to fermion lines .
the result is given as gauge - invariant four - quark operators .
the form of the gauge - invariant operators is obtained to all orders in @xmath1 by using the auxiliary field method .
the wilson coefficients of these operators can be computed by matching the amplitudes between the full theory and @xmath2 since the infrared divergence of the full theory is reproduced in @xmath2 . and
we obtain jet functions through the matching between @xmath2 and @xmath5 .
when the effects of collinear and soft gluons are included , we can obtain gauge - invariant operators , and the explicit form of these operators guarantees the color transparency at leading order in @xmath26 but to all orders in @xmath1 .
now the idea of the naive factorization in which the matrix elements of four - quark operators are reduced to products of the matrix elements of two currents has a theoretical basis .
furthermore when we include spectator interactions which contribute to the nonfactorizable contribution and to the heavy - to - light form factor , the amplitudes , which include four - quark and six - quark operators , are factorized to all orders in @xmath1 .
that is , the amplitudes can be written as a convolution of short - distance effects represented by the effective wilson coefficients and long - distance effects represented by the light - cone wave functions of mesons . and the convolution integrals appearing in the hard spectator interactions and in the hard contribution to the form factors are finite .
therefore we have proved that the decay amplitudes for nonleptonic @xmath0 decays into two light mesons at leading order in scet and to all orders in @xmath1 are factorized .
note that we have not included renormalization group running of the effective wilson coefficients . in order to include the renormalization group running ,
the scaling of the wilson coefficients can be achieved by calculating the anomalous dimensions of the operators , say , from eqs .
( [ mat1 ] ) and ( [ mat5 ] ) .
but the running of the wilson coefficients may not be appreciable at this order for the scale change from @xmath213 to @xmath370 . as an application of scet
, we have calculated the decay amplitudes for @xmath371 at leading order in scet . the results are consistent at order @xmath1 with the result in the heavy quark mass limit , which is shown explicitly here .
this is not a coincidence because the leading - order decay amplitudes in the heavy quark mass limit employing leading - twist meson wave functions correspond to the leading - order decay amplitudes in scet .
however scet extends the analysis to all orders in @xmath1 , and proves that the decay amplitudes are factorized .
the two types of the factorization properties , which correspond to the color transparency and the separation of long- and short - distance effects are satisfied to all orders in @xmath1 in nonleptonic @xmath0 decays .
we can go beyond the leading - order calculation and consider subleading corrections in order to check the validity of the approach using scet .
for example , we can ask questions on how chirally - enhaced contributions can be treated in scet , or how to include higher - twist wave functions of mesons , and how we can organize higher - order corrections in scet .
however we stress that this is a first step toward understanding nonleptonic @xmath0 decays , and those questions are under investigation .
this work was supported by korea research foundation grant ( krf-2002 - 041-c00052 ) .
in section [ sec2 ] we derived gauge - invariant four - quark operators through the matching calculation to order @xmath117 , in which all the off - shell modes were integrated out .
we can derive the gauge - invariant four - quark operators of the form @xmath372 using the auxiliary field method .
we construct a lagrangian of the on - shell fields with the off - shell fields which are auxiliary fields .
then we integrate out the off - shell auxiliary fields to obtain gauge - invariant operators entirely in terms of the on - shell fields . in integrating out the off - shell modes ,
the wilson lines @xmath58 , @xmath373 , @xmath246 and @xmath374 appear and the final form is gauge invariant .
the auxiliary field method offers a consistent way to derive gauge - invariant operators to all orders .
the auxiliary field method has been used in scet for the gauge - invariant heavy - to - light current operators @xcite , and for the processes in which there are collinear particles in the @xmath15 and @xmath17 directions @xcite .
we extend the idea of the auxiliary field method to derive the gauge - invariant four - quark operators in nonleptonic decays .
since there are collinear fields both in the @xmath15 and the @xmath17 directions , the auxiliary field method for collinear quarks @xmath32 and @xmath78 is the same as the method presented in ref .
@xcite . here
we consider how to treat the heavy quark interacting with two types of collinear gluons @xmath33 and @xmath80 . for clarity
, we first construct the lagrangian of the on - shell fields with the off - shell fields of order @xmath375 , and integrate out the off - shell modes to obtain the gauge - invariant four - quark operators . through this procedure
, we obtain the lagrangian and the four - quark operators in @xmath2 .
once we get familiar with the construction of the lagrangian with the off - shell modes of order @xmath375 , we include the off - shell modes of order @xmath376 , and integrate out all the off - shell modes .
we can obtain the lagrangian and the four - quark operators in @xmath5 in a gauge - invariant way .
let us categorize all the fields in @xmath2 in powers of @xmath26 .
note that the small expansion parameter @xmath41 in @xmath2 is of order @xmath377 and @xmath378 in @xmath5 is of order @xmath379 . in order to avoid confusion
, we express the scaling of all the momenta in powers of @xmath26 .
the on - shell fields are the collinear fields @xmath32 , @xmath33 , which scale as @xmath380 , the collinear fields @xmath78 , @xmath80 , which scale as @xmath381 , the soft fields @xmath35 , @xmath36 , which scale as @xmath382 , and the usoft fields @xmath38 , @xmath39 , @xmath99 , which scale as @xmath383 . in @xmath2
, we include the following off - shell fields .
when the on - shell heavy quark @xmath99 interacts with a collinear gluon @xmath33 ( @xmath80 ) , the final heavy quark has momentum of order @xmath384 [ @xmath385 .
we label this off - shell heavy quark as @xmath386 ( @xmath387 ) . if the on - shell heavy quark @xmath99 interacts with @xmath33 and @xmath80 , the final off - shell heavy quark has momentum of order @xmath388 , and @xmath389 .
we label this off - shell field as @xmath390 .
a gluon interacting with @xmath33 and @xmath80 through the triple gluon vertex also has a momentum @xmath391 , and we label this gluon as @xmath392 .
we list all the on - shell fields and the off - shell auxiliary fields in table [ intmod ] .
the collinear quarks can also be off - the mass shell of order @xmath375 , which we denote as @xmath393 and @xmath394 .
the treatment of the off - shell collinear quarks is presented in ref .
@xcite , and we will not repeat it here . since the soft and usoft quarks and gluons are not relevant in our argument , we do not present them in table [ intmod ] .
.[intmod ] list of fields to construct the auxiliary lagrangian in @xmath2 .
the on - shell fields are collinear , and heavy quark fields , and the wilson lines obtained by the corresponding gluons are listed in the last column .
the off - shell fields , which appear as intermediate states , are classified by their momentum scaling behavior . [ cols="^,^,^,^",options="header " , ] now we construct the lagrangian including the off - shell modes of order @xmath376 . through this procedure
, we obtain the lagrangian and the four - quark operators in @xmath5 . in @xmath5
, we have to integrate out all the off - shell modes in table [ modscet ] by constructing the lagrangian with additional auxiliary fields with additional soft momentum , with the index @xmath395 .
the on - shell and the off - shell fields in @xmath5 are listed in table [ modscet ] .
note that the momentum scaling of each fields has changed in order to accomodate @xmath5 .
since @xmath396 , we do not put the index @xmath395 to @xmath390 and @xmath392 .
for the same reason , we express the off - shell heavy quarks as @xmath386 and @xmath387 .
we include the off - shell modes due to the soft momentum .
we label these fields with the index @xmath395 to denote the off - shellness @xmath73 .
the auxiliary lagrangian including the off - shell modes of order @xmath376 is given by @xmath397 \psi_n + \overline{\psi}_{\overline{n } } n\cdot\bigl [ \mathcal{q } + g ( a_{\overline{n}}^x + a_{\overline{n } } ) \bigr ] \psi_{\overline{n } } \nonumber \\ & & + \overline{\psi}_h \overline{n } \cdot \bigl[\mathcal{p } + g ( a_q + a_n ) \bigr ] \psi_h .
\label{hqq}\end{aligned}\ ] ] note that there is no soft gluon @xmath36 involved in eq .
( [ hqq ] ) for the heavy quark .
it appears in the lagrangian for the collinear quarks .
the explicit lagrangian for the collinear quarks involving the off - shell fields of order @xmath376 can be found in eqs .
( a4 ) and ( a5 ) in ref .
@xcite .
the difference between the lagrangians given in eqs .
( [ hqaux ] ) and ( [ hqq ] ) is that all the off - shell fields are labeled by @xmath395 , and @xmath398 , @xmath399 .
therefore the heavy quark field , after the off - shell fields are integrated out , is written as @xmath400 and the collinear quarks are written as @xcite @xmath401 where @xmath402 , and @xmath403 are the fourier transforms of the wilson lines @xmath404 \biggr\ } , \nonumber \\
s_x ( y ) & = & \mathrm{p } \exp \biggl\ { ig \int_{-\infty}^y ds \bigl [ n \cdot a_{nx } ( sn ) + n \cdot a_s ( sn ) \bigr ] \biggr\ } , \label{gluong2}\end{aligned}\ ] ] and @xmath405 and @xmath406 are obtained by replacing @xmath15 by @xmath17 in eq .
( [ gluong2 ] ) .
the wilson lines satisfy @xmath407 the last two relations in eq .
( [ trans ] ) were obtained in ref .
@xcite and the first relation is new . by integrating out the off - shell fields , the singlet four - quark operator in @xmath5 is given by @xmath408 using eq .
( [ wilson ] ) , and similarly the nonsinglet four - quark operator is given as @xmath409 this result is the same as the explicit calculation obtained in section 2 , and it is a proof to all orders in @xmath1 using the auxiliary field method .
we can also derive the subleading operators @xmath235 in eq .
( [ o1ab ] ) and @xmath410 and @xmath411 in eq .
( [ fsubop ] ) using the auxiliary field method .
here we have to consider off - shell modes from the collinear gluons both in the @xmath15 and @xmath17 directions . for simplicity
, we consider the subleading operators in @xmath2 disregarding the off - shell modes by soft gluons .
the leading - order result was derived in ref .
@xcite , and here we present a new result which yields gauge - invariant operators at subleading order .
in addition to the solutions in eq .
( [ eom ] ) at leading order , when we include the subleading terms in the covariant derivative in eq .
( [ cov ] ) , we obtain the equations of motion at subleading order , which are given as @xmath412 & = & 0 , \nonumber \\ { } [ \overline{w}_q n\cdot \mathcal{q } \overline{w}_q^{\dagger } , [ w_q \overline{n } \cdot \mathcal{p } w_q^{\dagger } , i\mathcal{d}_{q\perp}^{\nu } ] ] & = & 0 , \label{subeom}\end{aligned}\ ] ] where @xmath413 . here
we make an ansatz @xmath414 and they satisfy eq .
( [ subeom ] ) .
first the intermediate form of the subleading four - quark operators relevant to the nonfactorizable spectator contribution and the heavy - to - light form factor can be obtained by neglecting the off - shell modes in the @xmath17 direction since we are interested in the operators proportional to @xmath415 at leading order in @xmath416 . as a definite example , let us concentrate on the operator @xmath259 and obtain the gauge - invariant form by integrating out the off - shell modes .
the derivation for the operator @xmath254 can be done in a similar way . in @xmath2 , without the off - shell modes in the @xmath17 direction , the operator @xmath259 is given by @xmath417 \frac{\fmslash{n}}{2 } ( \chi + \chi_q ) , \label{sample}\end{aligned}\ ] ] where the color indices are suppressed .
when we include off - shell modes , each field is written as @xmath418 then the operator in eq .
( [ sample ] ) is written as @xmath419 \cdot \bigl[(\overline{\chi } w_q^{\dagger } ) \gamma_{2i } \overline{w}_q \frac{1}{n\cdot \mathcal{q } } \overline{w}_q^{\dagger } \bigl ( [ i\fmslash{\mathcal{d}}_{n\perp q } w_q ] \frac{\fmslash{n}}{2 } \chi \bigr ) \bigr].\ ] ] if we use the ansatz in eqs .
( [ lan ] ) and ( [ san ] ) , the singlet operator becomes @xmath420 \nonumber \\ & & \times \bigl[(\overline{\chi } w_q^{\dagger})_{\beta } \gamma_2 \bigl ( \overline{w}_q \frac{1}{n\cdot \mathcal{q } } \overline{w}_q^{\dagger } [ i\fmslash{\mathcal{d}}_{n\perp q } w_q ] \frac{\fmslash{n}}{2 } \chi \bigr)_{\beta } \bigr ] \nonumber \\ & = & \bigl [ ( \overline{\xi } w)_{\alpha } \gamma_1 h_{\alpha } \bigr ] \nonumber \\ & & \times \bigl[(\overline{\chi } w_q^{\dagger}\overline{w}_q)_{\beta } \gamma_2 \frac{1}{n\cdot \mathcal{q } } \bigl([i\fmslash{d}_{n\perp } \overline{w}_q^{\dagger } w_q ] \frac{\fmslash{n}}{2 } \chi \bigr)_{\beta } \bigr ] \nonumber \\ & = & \bigl[(\overline{\xi } w)_{\alpha } \gamma_1 h_{\alpha } \bigr ] \cdot \bigl [ ( \overline{\chi } \overline{w})_{\beta } \gamma_2 \frac{1}{n\cdot \mathcal{q}}\bigl ( [ w^{\dagger } i\fmslash{d}_{n\perp } w ] \frac{\fmslash{n}}{2 } \overline{w}^{\dagger } \chi \bigr)_{\beta } \bigr ] , \end{aligned}\ ] ] and the nonsinglet operator becomes @xmath421 \cdot \bigl[(\overline{\chi } w_q^{\dagger})_{\alpha } \gamma_2 \bigl ( \overline{w}_q \frac{1}{n\cdot \mathcal{q } } \overline{w}_q^{\dagger } [ i\fmslash{\mathcal{d}}_{n\perp q } w_q ] \frac{\fmslash{n}}{2 } \chi \bigr)_{\beta } \bigr ] \nonumber \\ & = & \bigl [ \overline{\xi}_{\beta } \gamma_1 ( \overline{w } h)_{\alpha } \bigr ] \cdot \bigl[\overline{\chi}_{\alpha } \gamma_2 \frac{1}{n\cdot \mathcal{q } } \bigl ( \overline{w}_q^{\dagger } [ i\fmslash{\mathcal{d}}_{n\perp q } w_q ] \frac{\fmslash{n}}{2 } \chi \bigr)_{\beta } \bigr ] \nonumber \\ & = & \bigl [ \overline{\xi}_{\beta } \gamma_1 ( \overline{w } h)_{\alpha } \bigr ] \cdot \bigl[\overline{\chi}_{\alpha } \gamma_2 \frac{1}{n\cdot \mathcal{q } } \bigl ( [ i\fmslash{d}_{n\perp } \overline{w}_q^{\dagger } w_q ] \frac{\fmslash{n}}{2 } \chi \bigr)_{\beta } \bigr ] \nonumber \\ & = & \bigl [ ( \overline{\xi } w\bigr)_{\beta } \gamma_1 h_{\alpha}\bigr ] \cdot \bigl [ \bigl(\overline{\chi } \overline{w } \bigr)_{\alpha } \gamma_2 \bigl([w^{\dagger } i\fmslash{d}_{n\perp } w ] \frac{1}{n \cdot \mathcal{q } } \frac{\fmslash{n}}{2 } \overline{w}^{\dagger } \chi\bigr)_{\beta}\bigr],\end{aligned}\ ] ] these are the subleading operators which are gauge invariant .
other operators can be shown to have the forms presented in the paper in a similar way .
99 g. buchalla , a. j. buras and m. e. lautenbacher , rev .
* 68 * ( 1996 ) 1125 ; a. j. buras and r. fleischer , in : a. j. buras , m. lindner ( ed . ) , heavy flavours ii , world scientific , singapore , 1998 , p. 65 .
m. bauer and b. stech , phys .
b * 152 * ( 1985 ) 380 ; m. bauer , b. stech and m. wirbel , z. phys .
c * 34 * ( 1987 ) 103 .
j. d. bjorken , nucl .
( proc . suppl . )
b * 11 * ( 1989 ) 325 .
a. ali and c. greub , phys .
d57 ( 1998 ) 2996 ; a. ali , g. kramer and c. d. lu , phys .
d * 58 * ( 1998 ) 094009 ; a. ali , g. kramer and c. d. lu , phys .
d * 59 * ( 1999 ) 014005 .
a. j. buras and l. silvestrini , nucl . phys .
b 548 ( 1999 ) 293 .
h. d. politzer and m. b. wise , phys .
b * 257 * 399 ( 1991 ) .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda .
83 ( 1999 ) 1914 ; nucl .
b * 606 * ( 2001 ) 245 . c. w. bauer , s. fleming and m. luke , phys .
d * 63 * ( 2001 ) 014006 . c. w. bauer , s. fleming , d. pirjol and i. w. stewart , phys .
d * 63 * ( 2001 ) 114020 . c. w. bauer and i. w. stewart , phys .
b * 516 * ( 2001 ) 134 . c. w. bauer , d. pirjol and i. w. stewart , phys .
. lett . * 87 * ( 2001 ) 201806 . c. w. bauer , s. fleming , d. pirjol , i. z. rothstein and i. w. stewart , phys .
d * 66 * ( 2002 ) 014017 .
j. chay and c. kim , phys .
d * 65 * , 114016 ( 2002 ) .
a. v. manohar , t. mehen , d. pirjol and i. w. stewart , phys .
b * 539 * ( 2002 ) 59 . c. w. bauer , d. pirjol and i. w. stewart , phys .
d * 65 * ( 2002 ) 054022 .
j. chay and c. kim , hep - ph/0301055 . c. w. bauer , d. pirjol , and i. w. stewart , hep - ph/0303156 . c. w. bauer , d. pirjol and i. w. stewart , phys .
d * 66 * ( 2002 ) 054005 .
r. j. hill and m. neubert , nucl .
b * 657 * , 229 ( 2003 ) . c. w. bauer , d. pirjol and i. w. stewart , phys .
d * 67 * , 071502 ( 2003 ) .
d. pirjol and i. w. stewart , phys .
d * 67 * , 094005 ( 2003 ) .
g. p. lepage and s. j. brodsky , phys .
d * 22 * ( 1980 ) 2157 .
m. beneke , a. p. chapovsky , m. diehl and t. feldmann , nucl .
b * 643 * ( 2002 ) 431 .
j. charles , a. le yaouanc , l. oliver , o. pene , and j. c. raynal , phys .
d * 60 * ( 1999 ) 014001 .
m. j. dugan and b. grinstein , phys .
b * 255 * ( 1991 ) 583 . a. g. grozin and m. neubert , phys .
d * 55 * ( 1997 ) 272 .
m. beneke , and t. feldmann , nucl .
b * 592 * ( 2001 ) 3 . | we consider nonleptonic @xmath0 decays into two light mesons at leading order in soft - collinear effective theory , and show that the decay amplitudes are factorized to all orders in @xmath1 .
the operators for nonleptonic @xmath0 decays in the full theory are first matched to the operators in @xmath2 , which is the effective theory appropriate for @xmath3 with @xmath4 gev .
we evolve the operators and the relevant time - ordered products in @xmath2 to @xmath5 , which is appropriate for @xmath6 . using the gauge - invariant operators in @xmath5 ,
we compute nonleptonic @xmath0 decays in scet , including the nonfactorizable spectator contributions and spectator contributions to the heavy - to - light form factor . as an application
, we present the decay amplitudes for @xmath7 in soft - collinear effective theory . , |
understanding the nature of mechanical stability , how it arises , and how it can be controlled is important to fields ranging from civil engineering to biology .
james clerk maxwell was among the first to address this problem with mathematical rigor in his 1864 paper @xcite in which he established the conditions for the mechanical stability of a frame of points connected by lines or struts of fixed length .
his approach , which can easily be generalized to treat particles interacting via central - force potentials or bonds between particles interacting via an angular potential , has since been used to study network glasses @xcite , randomly packed spheres near jamming @xcite , networks of semi - flexible polymers @xcite , regular @xcite and random @xcite periodic lattices , and a host of engineering problems @xcite . in this paper , we study the mode structure of a particular lattice - the kagome lattice shown in fig .
[ fig : kagorand ] - in which nearest neighbor ( @xmath1 ) sites are connected by harmonic central - force springs of spring constant @xmath19 and next - nearest - neighbor ( @xmath4 ) sites are randomly connected by similar springs but with a different spring constant @xmath20 . in the absence of @xmath4 springs , finite realizations of this lattice
are just on the verge of mechanical stability in the maxwell sense .
the addition of @xmath4 springs renders the system stable .
when these springs are added uniformly with a spring constant @xmath5 that is allowed to approach zero continuously @xcite , the result is a kind of critical point at @xmath21 with a rich mode structure and diverging length and vanishing frequency scales as @xmath22 that have direct counterparts in the properties of randomly packed spheres near jamming @xcite . here we use the coherent potential approximation or cpa @xcite ( otherwise know as the effective medium approximation ) to explore in detail the static and dynamical mechanical properties and mode structure of the kagome lattice as a function of the probability @xmath12 that @xmath4 sites are connected by springs .
we find that the effective medium @xmath4 spring constant @xmath23 scales as @xmath10 for small @xmath12 , and it gives rise to length and frequency scales @xmath24 and @xmath25 . beyond a characteristic frequency @xmath26 ,
the imaginary part of @xmath27 rises quickly and phonons scatter strongly .
the ioffe - regel limit beyond which certain plane wave states become ill - defined is reached at a frequency slightly larger than @xmath28 .
maxwell @xcite argued that each constraint in a system of @xmath2 points in @xmath29-dimensions reduces the number , @xmath30 , of zero frequency modes of free points by one .
thus , the number of zero frequency modes in a system with @xmath31 constraints is @xmath32 .
of these , @xmath33 are those of rigid translations and rotations , leaving @xmath34 internal zero frequency modes , which are usually refereed to as floppy modes @xcite or mechanisms in the engineering literature @xcite , when @xmath35 .
mechanical stability requires that there be no floppy modes , or equivalently that @xmath36 .
though maxwell considered general frames of points , his analysis can immediately be applied to the determination of mechanical stability of extended crystals and glasses - as was first done in the context of network glasses , which are unstable in the absence of bending forces , by phillips and thorpe @xcite - in which points are replaced by particles or atoms and struts are replaced by inter - particle potentials .
if each particle in an extended system is on average connected by central - force potential creating bonds to @xmath8 other particles , then , because each bond is shared by two particles , the system experiences @xmath37 constraints if bond lengths are restricted to be at their equilibrium length and there are no redundant bonds @xcite , and it has @xmath38 floppy modes .
thus , in the limit of large @xmath2 , a central - force system is mechanically stable if @xmath39 . in an infinitely extended system
, @xmath8 is independent of @xmath2 . in a finite system cut from an infinite one , particles at the boundary
have fewer neighbors than those in the interior , and @xmath8 ( which is an average quantity ) is of order @xmath40 smaller than it would be in the ideal infinite system .
systems with @xmath41 are called isostatic .
a finite piece of @xmath2 particles cut from an infinite isostatic has a surface with of order @xmath42 cut bonds , and thus that many fewer constraints and that many more floppy modes .
these modes generally remain floppy even if they have finite amplitudes if there are free boundary conditions . because of their geometry with straight segments extending throughout their interior
, the square , the kagome , and related lattice of @xmath2 sites with periodic rather than free boundary conditions have @xmath42 infinitesimal floppy modes - modes that have zero energy for infinitesimal displacements but finite energy for large displacements @xcite .
the simplest example of an infinitesimal floppy modes is that of a mass connected to two identical co - linear springs that are rigidly attached to walls on either side of the mass .
if the springs are not under tension , infinitesimal displacements of the particle perpendicular to the spring cost no energy but finite displacements do . thus the harmonic kagome lattice has @xmath43 floppy modes .
the recently most explored random isostatic system is perhaps the system of randomly close - packed spheres at the jamming transition @xcite in the system of frictionless soft spheres with one - sided repulsion .
this transition occurs when the volume faction @xmath44 of spheres just exceeds the critical value @xmath45 at which they form just enough contacts to support a compressional load .
thus at @xmath45 , the bulk modulus @xmath46 jumps discontinuously to a nonzero value , the number of neighbors per particle grows as @xmath47 , and the shear modulus grows continuously from zero as @xmath48 .
associated with this transition are a diverging length scale @xmath49 and vanishing frequency scale @xmath14 whose behavior follows from quite general cutting " arguments @xcite .
isostaticity is not limited to random systems .
there are a number of periodic isostatic lattices with @xmath50 @xmath1 bonds per particle .
these include the hypercubic lattice in @xmath29 dimensions and the kagome lattice in two - dimensions and its generalization to higher dimensions with @xmath1 bonds occupied by springs of spring constant @xmath51 , all with @xmath52 .
they can be moved off isostaticity by introducing springs of spring constant @xmath5 on @xmath4 bonds either homogeneously or randomly . in homogeneous systems ,
the isostatic limit is continuously approached by allowing @xmath22 . in the square - lattice version of this model @xcite ,
the bulk modulus is nonzero at @xmath53 , but the shear modulus vanishes as @xmath5 in this limit . in the kagome version ,
both the shear and the bulk moduli are nonzero at @xmath53 . in both models , as in the jamming problem , there is a divergent length scale , @xmath54 and a vanishing frequency scale @xmath55 .
the coherent potential , or equivalently the effective medium , approximation ( cpa ) @xcite is a powerful tool for the calculation of properties of random systems , from the electronic structure of alloys @xcite to the elasticity of composite materials @xcite .
it has been used with great success in the study of rigidity percolation problem @xcite and , more recently , in the dynamics of off - lattice systems near a rigidity threshold @xcite .
it also provides a quantitative theory of the static and dynamic phonon response in the nearly isostatic square lattice in which next - nearest - neighbor springs ( @xmath4 ) are randomly added to a lattice in which all nearest - neighbor - site connected by equivalent springs @xcite . in this lattice ,
the effective - medium @xmath4 spring constant @xmath56 has a zero - frequency limit , @xmath57 , that scales as @xmath58 leading to lengths and frequencies that scale as @xmath15 and @xmath59 as they do in the jamming problem .
the finite - frequency spring constant is a scaling function of @xmath17 at small @xmath12 : @xmath60 .
as @xmath12 is increased , there is a crossover from nonaffine response at small @xmath12 to affine response with @xmath61 at larger @xmath12 , much as there is in networks of semi - flexible polymers as the average polymer length is increased @xcite in this paper , we use the cpa to explore in detail the static and dynamic properties of the kagome lattice as it is moved away from isostaticity by the random addition of @xmath4 springs .
our approach is identical to that employed in our treatment of the square lattice @xcite , but calculations are considerably more complicated because each unit cell has three sites rather than a single one .
the results of our calculations follow what we expect from our experience with the homogeneous square and kagome lattices and with the random square lattice . the bulk and shear moduli are nonzero and proportional to the @xmath1 spring constant @xmath51 when @xmath62 . the static @xmath4 effective medium spring constant @xmath57 scales as @xmath10 leading to @xmath49 and @xmath63 in agreement with cutting arguments . as is the case in the square lattice
, @xmath16 is a basically a scaling function of @xmath17 at small @xmath12 , but with small yet important deviations at small @xmath64 that describe rayleigh scattering , i.e. , a mean - free - path that scales as @xmath65 in @xmath66 .
the outline of this paper is as follows . in sec .
[ sec : kagoelas ] we review the elasticity of the homogeneous nearly isostatic kagome lattice . in sec .
[ sec : cpa ] we discuss the cpa on random nearly isotropic kagome lattice with the @xmath4 bonds randomly occupied with probability @xmath12 . in sec .
[ sec : results ] we discuss the results of the cpa calculation , including the crossover of @xmath27 from @xmath10 to @xmath12 behavior as @xmath12 increase or @xmath20 decreases , and the rapid increase of scattering at the characteristic frequency @xmath67 .
bonds denoted by purple dashed lines .
the unit cell triangle is marked with filled triangles .
particle @xmath68 , @xmath69 , @xmath70 in each unit cell are marked in the bottom right unit cell . ]
in this section we briefly review the elastic energy in central - force network models , in which the elastic energy @xmath71 can be written as a sum of the energy of each central - force bond @xmath72 where @xmath73 is the length of the bond and @xmath74 is the potential energy of the bond as a function of the length .
we consider a displacement field on the network that maps particle @xmath75 which is at position @xmath76 to a new position @xmath77 , thus the length of bond @xmath78 between particle @xmath75 and @xmath79 is changed into @xmath80 we refer to the original space in which particle @xmath75 is at @xmath76 as the _ reference space _ , and the space after applying the displacement field as the _ target space_. we consider harmonic potentials @xmath81 where @xmath82 is the rest length of the bond , and @xmath83 is the spring constant . the length @xmath73 can be expanded for small displacement @xmath84 as @xmath85 where @xmath86 , @xmath87 , and @xmath88 is the unit vector pointing along the bond in the reference space .
thus we have @xmath89 where @xmath90 is the magnitude of the force on the bond in the reference space .
in general we consider the case in which the reference state is in mechanical equilibrium , which means that the total force on each particle vanishes @xmath91 where the sum @xmath92 is over all occupied bonds connected to @xmath75 .
however , to capture the properties of random networks , which often carry residual stress , the length of each bond is not necessarily at its rest length , i.e. , @xmath93 in general .
the change of the elastic energy from the reference space to the target space of the whole system is then a quadratic form of the displacement field @xmath94 which can also been written as @xmath95 where @xmath96 is the component of @xmath97 parallel to @xmath98 and @xmath99 is the component perpendicular to @xmath98 . by doing a gradient expansion on the displacement field , @xmath100 where @xmath101 is the displacement field at position @xmath102
, we recover the elastic energy of the continuum theory @xmath103 with @xmath104 where the summation @xmath105 is over bonds connecting to one particle , and we are using a simple lattice with one particle per unit cell in this illustration .
the volume of a unit cell is denoted by @xmath106 .
the kagome lattice is a lattice with three particles per unit cell , and we shall use the following displacement vector to denote the deformation of the lattice @xmath107 where @xmath75 labels the unit cell and @xmath108 label the particles in the unit cell as in fig . [
fig : kagome ] .
as we have discussed , the elastic energy can be expanded in small displacement field @xmath109 , and to leading order we have the quadratic form @xmath110 which can be built from the analysis as in eq .
( [ eq : deltvuugene ] ) , but generalized to the case of the kagome lattice which has three sites per unit cell .
the matrix @xmath111 is called the dynamical matrix of the lattice .
this elastic energy can be written in momentum space as @xmath112 where the fourier transform into momentum space is defined as @xmath113 where @xmath2 is the number of unit cells .
the dynamical matrix for the homogeneous kagome lattice with all @xmath1 bonds occupied with springs of spring constant @xmath19 and all @xmath4 bonds with springs of spring constant @xmath20 is a @xmath114 matrix given by @xmath115 where the @xmath116 vectors and their derivation are given in app .
[ app : dm ]
. there are six translational degrees of freedom per unit cell in the kagome lattice giving rise to six phonon branches . of these , three are optical branches with frequencies of order @xmath117 , two are acoustic branches with sound velocities of order @xmath117 , and one is the anomalous branch .
the later three branches , which determines the low - energy elastic theory of the kagome lattice , have modes in the space spanned by the following three vectors @xmath118 which correspond to two translations of the whole unit cell in @xmath119 and @xmath120 directions and the rotation of the unit - cell triangle around its center .
the low - energy theory is governed by the @xmath121 reduced dynamical matrix obtained by integrating out the three high - energy optical branches , as shown in app .
[ app : rdm ] . for small momentum , @xmath122 , the reduced dynamical matrix is simply diagonalized by longitudinal and transverse acoustic phonons ( which are linear combinations of @xmath123 and @xmath124 ) with speeds of sound @xmath125 and @xmath126 and the rotational mode with a characteristic frequency @xmath127 at @xmath128 .
the bulk modulus @xmath46 and the shear modulus @xmath129 are related , respectively , to the longitudinal and transverse sound velocities through @xmath130 where @xmath131 is the mass density , which because there are three atoms per unit cell , is equal to @xmath70 in our units .
thus , @xmath132 and @xmath133 .
there is only weak mixing between the rotational modes and the acoustic phonons , and the system is isotropic
. for large momentum @xmath134 , strong mixing between the transverse acoustic modes and the rotational modes occurs , and the strong anisotropy of the isostatic state is retrieved .
the mixing is maximal along @xmath135 and symmetry equivalent directions , which we refer to as the isostatic directions , and the resulting two modes are shown in fig .
[ fig : kagome - modes ] .
the anomalous branch , with frequency of order @xmath20 , is the lower branch of the two . in the limit of @xmath136 ,
the lattice becomes isostatic , the isotropic region is squeezed to the origin , and the anomalous modes reduce to the isostatic floppy modes with zero frequency along @xmath135 ( @xmath137 line in fig .
[ fig : kagome - modes ] ) and symmetry equivalent directions as depicted in fig .
[ fig : kagome ] .
the name `` anomalous modes '' follows the nomenclature of ref .
@xcite , referring to the modes developed from the floppy modes as the system is moved away from the isostatic point . for a more detailed discussion of the low energy theory of the elasticity of the kagome lattice ,
see app .
[ app : rdm ] .
and the solid lines are for @xmath138 .
the floppy and anomalous branches are in red .
( b ) shows anomalous and shear modes along @xmath137 and indicates characteristic frequencies and wavenumbers .
frequencies @xmath139 , @xmath140 and @xmath141 are defined in the text .
( from ref .
@xcite ) ] of particular interest is the frequency of the anomalous modes in the vicinity of @xmath135 .
the squared frequency of these modes can be written as @xmath142 where @xmath143 . the function @xmath144 is plotted in fig .
[ fig : isosmode ] .
it reaches a maximum value of @xmath145 at a @xmath66 saddle point at @xmath146 and a local minimum value of @xmath147 at the zone edge @xmath148 . for small @xmath5 , @xmath149 , @xmath150 and @xmath151 .
all of the characteristic frequencies @xmath152 are proportional to @xmath153 for small @xmath5 .
@xmath144 is well approximated between @xmath146 and @xmath154 by @xmath155 , \ ] ] as is evident from fig .
[ fig : isosmode ] .
this relation will prove useful in our evaluation of integrals in our cpa analysis in sec .
[ sec : cpa ] lengths scaling as @xmath156 can be extracted from the phonon dispersion relations in various ways .
one length is the hybridization length @xmath157 obtained from the hybridization wavenumber @xmath158 separating the domain of predominantly transverse phonon behavior at low @xmath159 from the domain of predominantly rotation behavior at high @xmath159 .
other lengths can be obtained by comparing the @xmath160 term in @xmath161 to @xmath162 and @xmath163 : @xmath164 and @xmath165 .
an interesting property of @xmath166 is that the hybridization frequency @xmath167 obtained by setting @xmath168 in the transverse phonon frequency is identical to @xmath169 : @xmath170 .
one experimentally relevant quantity is the fourier transform of the finite temperature static phonon correlation function @xmath171 : @xmath172 where @xmath173 and @xmath174 label the basis defined in eq .
( [ eq : basissix ] ) of the @xmath175dimensional space of @xmath109 , @xmath176 labels the phonon band , and @xmath177 is the @xmath178-dimensional eigenvector associated with mode @xmath179 .
this correlation function is in the zero - frequency limit and thus independent of dissipation of the system .
the quantities @xmath180 are merely the eigenvalues of the dynamical matrix with the zero - frequency value of the spring constant ( the effective medium spring constant can depend on frequency as we discuss below in the cpa ) . the diverging length scale @xmath181 can be extracted this way from the static phonon correlation function in experiments .
, for @xmath182 .
the eigenvalue of the full @xmath114 dynamical matrix , @xmath183 , normalized by @xmath20 , is denoted by the red dots , and the eigenvalue of the @xmath184 reduced dynamical matrix ( [ eq : dmsplf ] ) is denoted by the red line .
the blue line represent the approximation ( [ eq : omegaa ] ) we used in the asymptotic calculation in the @xmath185 in cpa . ]
the cpa is a widely used method in the study of disordered systems @xcite . in it ,
a random system is mapped into an effective medium with no disorder that is described by a green s function with a suitable self - energy that can capture the effect of the disorder average of the randomness . to achieve this ,
one imposes a self - consistency constraint that the effective medium green s function perturbed by the presence of single impurity in the effective medium reduces to the effective medium green s function when averaged over the probability distribution of the impurity .
more specifically , the @xmath186-matrix of this perturbation vanishes upon averaging over configurations that contain and do not contain the impurity . for the case of the nearly isostatic kagome lattice
, the effective medium has all @xmath4 bonds occupied with an effective - medium spring of spring constant @xmath187 , and the effective medium green s function is identical to that of a homogeneous system with @xmath188 .
the cpa procedure consists of replacing one arbitrary @xmath4 bond with a new bond of spring constant @xmath189 , which takes on the value @xmath20 with probability @xmath12 ( bond occupied ) and the value @xmath190 with probability @xmath191 ( bond unoccupied ) .
this procedure leads to a modified dynamical matrix @xmath192 where @xmath193 where @xmath68 represents the arbitrary @xmath4 bond we have chosen to replace into @xmath189 .
this form of @xmath194 follows directly from the calculations leading to eq .
( [ eq : dmcons ] ) .
it depends on the wavenumbers @xmath195 and @xmath196 because the perturbed system is not translationally invariant .
the phonon green s function for the effective medium is @xmath197 in the perturbed system with one bond replaced , the green s function becomes @xmath198 and is no longer translationally invariant .
this green s function can be expanded for small @xmath194 @xmath199 where we have dropped the frequency @xmath200 dependence which is the same for every @xmath201 and @xmath194 .
this series can be written as @xmath202 where @xmath203 is the @xmath186-matrix expressed in the wavenumber basis . in the cpa , the effective medium spring constant @xmath27
is determined by requiring that the average value of @xmath204 be equal to @xmath205 or equivalently that the disorder average of the @xmath186-matrix vanish : @xmath206 the evaluation of the @xmath186-matrix is simplified by the following identity , @xmath207 where @xmath208 with @xmath209 the area of the unit cell in real space and @xmath210 is the area of the first brillouin zone in reciprocal space .
the integral is over the first brillouin zone .
the green s function @xmath211 is the phonon green s function in the effective medium so it depends on @xmath27 .
using these relations in eq .
( [ eq : t - matrix1 ] ) gives @xmath212 thus , the self - consistency equation ( [ eq : scet ] ) requires that @xmath213 from which one can solve for the effective medium @xmath4 spring constant @xmath27 for any given @xmath12 and @xmath200 .
the form of this solution at small @xmath27 depends on the behavior of the function @xmath214 at small @xmath27 , which is in turn determined by the form of the anomalous mode along the @xmath135 and other isostatic directions .
details of the calculation of @xmath214 are presented in app .
[ app : fint ] . in the following discussion
unless otherwise stated , we use reduced units with @xmath215 and lattice constant @xmath216 , and thus unitless spring constants , and elastic moduli , and frequencies : @xmath217 , @xmath218 , and @xmath219 .
we first consider the case of @xmath220 , which characterizes the static response of the system . for small @xmath12
, we expect that the effective medium spring constant @xmath27 also to be small and that we can , therefore , ignore the @xmath221 term in the cpa self - consistency equation ( [ eq : cpasce ] ) . using the asymptotic small @xmath9 limit @xmath222 , where @xmath223 , derived in app .
[ app : fint ] [ eq .
( [ eq : fzerofreq ] ) ] , we obtain the equation for @xmath27 at small @xmath12 @xmath224 which has the solution @xmath225 this solution has two limits : @xmath226 where @xmath227 . in the first case , @xmath228 , and the solution for @xmath27
is obtained by ignoring the first term in eq .
( [ eq : scesp ] ) ; in the second case , the opposite is true , and @xmath27 is obtained by ignoring the second term in this equation . in the second case , every @xmath4 bond distorts in the same way under stress , and response is affine .
in the first case @xmath229 , indicating that the response is nonaffine with local rearrangements in response to stress . within the cpa ,
this result emerges because of the divergent elastic response encoded in @xmath201 ( and @xmath230 ) as @xmath231 ( see app .
[ app : fint ] ) .
the nonaffine regime arises when @xmath4 springs are strong enough for the second term in eq .
( [ eq : scesp ] ) to dominate the first . as @xmath20 approaches zero at fixed @xmath12 ,
distortions produced by the extra bond decrease and the nonaffine regime becomes vanishingly small .
numerical solutions of the cpa self - consistency equation ( [ eq : cpasce ] ) with the full @xmath114 dynamical matrix is plotted in fig .
[ fig : zerofreqcpa ] , along with a comparison to the analytical solution ( [ eq : scesolusp ] ) and the two asymptotic forms in eq .
( [ eq : affinona ] ) .
this crossover of the effective medium spring constant @xmath27 between @xmath10 and @xmath12 is different from the nonaffine - affine crossover in the case of random nearly isostatic square lattice @xcite which describes the response of the square lattice to shear stress , because the shear modulus ( more precisely , @xmath232 ) is proportional to the effective medium spring constant @xmath27 for the @xmath4 bonds in the square lattice . in the case of kagome lattice in this paper ,
the shear modulus is finite and proportional to @xmath19 , whereas @xmath27 mainly determines the rigidity of the lattice with respect to the floppy mode , which are essentially rotations of the triangles as shown in fig .
[ fig : kagome ] .
length and frequency scales can be extracted in the static limit much as they were extracted in the homogeneous case discussed in sec .
[ sec : kagoelas ] .
the finite temperature static phonon correlation function @xmath233 is the inverse of the dynamical matrix evaluated at @xmath234 , whose eigenvalues and eigenvectors are identical to those of the homogeneous case with @xmath5 replaced by @xmath235 .
the eigenvalues allows us to identify frequencies by taking the square roots of the appropriate eigenvalues of @xmath111 : @xmath236 unlike the situation homogeneous lattices , these frequencies are not equal to any physical dynamical - mode frequency of the system .
they do , however , provide information about the static properties of the phonon correlation function @xmath233 that could in principle be measured at finite temperature via scattering or particle tracking experiment .
they also lead to diverging lengths just as they do in the homogeneous case : @xmath237 as a function of @xmath12 at @xmath220 of the cpa self - consistency equation ( [ eq : cpasce ] ) with the full @xmath114 dynamical matrix .
@xmath4 bond spring constant @xmath238 are shown in red dots , purple squares , and blue diamonds respectively .
the analytical solution ( [ eq : scesolusp ] ) of these three cases at small @xmath12 are shown in red , purple , and blue lines .
also shown are the nonaffine ( @xmath239 ) and affine ( @xmath240 at @xmath241 ) limits in green dashed line and red dash - dotted line . at large @xmath12 the numerical solution , especially the one for @xmath242 deviate significantly from the nonaffine limit form because eq .
( [ eq : scesp ] ) is an approximation at small @xmath12 by ignoring the highest order term in eq .
( [ eq : cpasce ] ) . ] for finite frequency @xmath200 , the effective medium spring constant is complex , @xmath243 , where the imaginary part @xmath244 , which describes damping of phonons in this random network , is odd in @xmath200 and positive for @xmath245 . from the analysis for the static limit @xmath220
, we see that the interesting case is the nonaffine regime with @xmath246 , in which the self - consistency equation ( [ eq : scesp ] ) simplifies to @xmath247 in the static limit , @xmath248 is singular in the @xmath249 limit . as we show in app .
[ app : fint ] , at finite frequency , @xmath250/\fcnnnm$ ] , which leads to @xmath251 as depicted in fig .
[ fig : finifreqcpa ] .
taking @xmath220 , this solution reduces to the zero - frequency solution [ eq .
( [ eq : affinona ] ) ] in the nonaffine limit . equation ( [ eq : cpaasymsolu ] ) develops an imaginary part when @xmath252 , which , as we discussed above must be negative for @xmath245 .
it is straightforward to see that this solution satisfies the scaling form @xmath253 , as does the cpa effective @xmath4 spring constant in the square lattice ref .
this solution shows a rapid increase of damping beyond a characteristic frequency @xmath254 marking another characteristic frequency that scales also as @xmath12 . numerical solution of the cpa self - consistency equation ( [ eq : cpasce ] ) using the full @xmath114 dynamical matrix is also shown in fig .
[ fig : finifreqcpa ] .
we see that the asymptotic form ( [ eq : cpaasymsolu ] ) captures the solution fairly well .
this special behavior of the imaginary part of the effective medium spring constant @xmath27 is related to the phonon spectrum of the kagome lattice . as we have discussed in sec
[ sec : kagoelas ] , at low frequencies , there is only very weak mixing between the rotational branch which is strongly affected by the @xmath4 bonds and the acoustic phonon branches which are only very weakly scattered by the @xmath4 bonds , and thus the damping to the acoustic phonons is very weak . on the other hand , for higher frequencies the transverse phonon strongly mixes with the rotational modes , making the damping rapidly increase with scatterings from the @xmath4 bonds beyond @xmath28 , although the exact value of @xmath28 may be an artifact of the cpa method . the weak scattering below @xmath255 is not captured by the asymptotic form ( [ eq : cpaasymsolu ] ) for small @xmath27 because the function @xmath23 in eq .
( [ eq : cpaasymsolu ] ) was obtained using the dominant small @xmath27 limit of the integral @xmath256 .
there are , however , contributions to this integral that do not diverge and that contribute a subdominant imaginary part to @xmath27 , even when @xmath257 , that is of order @xmath258 at small @xmath200 corresponding to rayleigh scattering . more discussion
is included in app .
[ app : fint ] . and @xmath259 . the numerical solution to eq .
( [ eq : cpasce ] ) with the full @xmath114 dynamical matrix is shown as the data points .
blue circles and red squares represent real and ( negative of ) imaginary parts of @xmath27 at @xmath260 , and green diamonds and brown triangles represent the real and ( negative of ) imaginary parts of @xmath27 at @xmath259 .
the asymptotic form ( [ eq : cpaasymsolu ] ) is shown as the blue ( real ) and red ( negative of imaginary ) lines . in this plot frequency
is rescaled by @xmath12 , and the effective medium spring constant @xmath27 is rescaled by its value at zero frequency which is real . ] in the homogeneous case , the eigenvalues of the dynamical matrix lead naturally to the identification of characteristic frequencies @xmath169 , and @xmath261 that vanish as @xmath153 in the limit of @xmath22 . in the random case
, we have to deal with both the frequency - dependence @xmath262 and the fact that it is a complex number , and we must ask whether these frequencies have any real meaning . as discussed earlier in sec .
[ sec : hkl ] , we can extract frequencies from the static dynamical matrix in exactly the same way that we did for the homogeneous case , and they satisfy @xmath263 thus all of these frequencies are greater than the frequency @xmath264 . as a result ,
the signatures in the phonon dispersion relation including the hybridization and the saddle point are washed out by the strong scattering , as is shown in fig .
[ fig : scattering ] . the phonon density of states ( dos ) can be calculated from the retarded green s function through @xmath265 where the trace is over both momentum @xmath195 and the phonon modes . using this we get the phonon dos of the effective medium solved from the cpa , as plotted in fig .
[ fig : dos ] . as a comparison
, we also show the phonon dos of a periodic kagome lattice with the @xmath4 spring constant equal to @xmath266 , which is real valued . for small frequencies ,
at which the imaginary part of the cpa solution @xmath23 is very small , the two dos are very close , and can be fitted nicely by the debye - like total dos of the transverse and the longitudinal phonons @xmath267 where @xmath268 and @xmath269 are respectively the longitudinal and transverse speed of sound ( we have taken @xmath215 as stated earlier ) . at the critical frequency @xmath28 ,
the imaginary part of @xmath23 rapidly increase , inducing a rapid increase of the phonon dos . on the other hand , the periodic lattice exhibit a jump in dos at @xmath270 corresponding to the minimum of the phonon dispersion relations at the edge of the 1bz . at @xmath271
the dos of the periodic lattice has a logarithmic singularity , corresponding to the saddle point of the phonon dispersion at @xmath272 on the isostatic directions @xcite . for the cpa effective medium ,
this singularity is totally washed out due to the strong damping beyond @xmath28 , which is similar to the case of the square lattice @xcite . from the cpa solution at finite frequency , we identified a frequency scale @xmath28 beyond which phonon scattering rapidly increase . in this subsection
we examine the scattering of phonons in more detail .
the scattering of the transverse phonons is characterized by the imaginary part of the phonon response function projected to the transverse direction @xmath273 .
the phonon response function is defined as @xmath274 where @xmath275 and @xmath276 label time , @xmath173 and @xmath174 label the basis defined in eq .
( [ eq : basissix ] ) of the @xmath175dimensional space of @xmath109 .
this response function is related to the phonon green s function through @xmath277 .
the imaginary part of the transverse component of this response function @xmath273 characterizes the scattering of the transverse phonon by disorder .
@xmath273 is calculated for small momentum and frequency using the asymptotic cpa solution ( [ eq : cpaasymsolu ] ) , and shown in fig .
[ fig : scattering ] . also shown in the figure is the frequency at which the phonon green s function of the anomalous branch has a complex pole , which is solved from the equation @xmath278 , which characterizes the dynamic dispersion relation .
we use the form of @xmath279 as defined in app .
[ app : rdm ] for this calculation .
it is clearly shown in the figure that below @xmath28 , the response function has dirac - delta peaks at the frequencies determined by the transverse phonon dispersion relation @xmath280 .
above @xmath28 the imaginary part @xmath281 takes off , and the phonon peaks progressively broaden , showing that the transverse phonon is no longer a good eigenstate of the system .
furthermore , the hybridization of the transverse phonon and the rotational modes into the anomalous mode and the van hove singularities in the density of states are washed out by the strong scattering . as a result , @xmath282 , @xmath283 , @xmath284 no longer play meaningful role in the dynamic response function . as a function of @xmath200 for various values of @xmath159 ( we took @xmath135 to follow the isostatic @xmath137 direction ) .
the green line in the bottom plane marks @xmath28 , the blue and red lines marks the solution @xmath285 and @xmath281 of the equation @xmath278 , which solves for the pole of the green s function for the anomalous branch .
the derivation of @xmath279 is shown in app .
[ app : rdm ] . ] the strength of the scattering can be characterized by the ioffe - regel ( ir ) limit , which states that the plane wave states are no longer well defined if the mean free path @xmath286 is comparable to or less than the phonon wavelength @xmath287 .
an equivalent condition is that the relaxation time becomes comparable to the period of the wave , i.e. , @xmath288 .
the solution for the positions of the complex poles of the green s function of the anomalous branch @xmath278 shows that the imaginary part @xmath281 becomes comparable to the real part @xmath285 not far beyond @xmath28 and that @xmath289 .
the associated ir length scale can be derived from @xmath28 and @xmath290 to be of order @xmath291 .
this ir length scale differs from that in jammed solids , @xmath292 as derived in refs .
@xcite ( called @xmath293 in ref .
this discrepancy can be attributed to the different scaling of the shear modulus @xmath129 . in the kagome lattice ,
the shear modulus @xmath129 is proportional to @xmath19 and thus scales as @xmath294 , whereas in jammed solids @xmath295 .
thus , the transverse speeds of sound scales as @xmath296 and @xmath297 in these two cases respectively . in both cases ,
the frequency beyond which plane wave states are strongly scattered is @xmath298 .
therefore , the scattering length scale , @xmath299 are respectively @xmath300 and @xmath301 in the kagome lattice and jammed solids . up to now , three examples of random nearly isostatic systems have been studied , including the random nearly isostatic square lattice , random nearly isostatic kagome lattice discussed in this paper , and jammed solids near point j. in all cases , the characteristic frequency for the onset of the anomalous mode plateau @xmath298 and the isostatic length scale @xmath302 . on the other hand ,
the scaling of elastic moduli in the three cases are different because of different _ network architecture _ : in the square lattice @xmath303 and @xmath304 , in kagome lattice @xmath305 , and in jammed solids @xmath295 and @xmath304 ( a jump from zero to finite value at point j ) . as a result , the scattering length scales are also different .
the square lattice is anisotropic , and we studied the scattering of the @xmath306 vibrations along @xmath159 direction and found that the scattering length corresponded to the point @xmath307 in the first brillouin zone , @xmath308 , or @xmath309 . in kagome lattice @xmath310 . in jammed solids , the ir length scale @xmath311 . in ref .
@xcite , wyart studied transport properties of amorphous solids modeled by an isostropic random network near its percolative rigidity threshold . in this system ,
both @xmath46 and @xmath129 vanish as @xmath312 , and the crossover frequency between plane - wave and strongly scattered states is @xmath313 .
both the longitudinal and transverse sound velocities scale as @xmath297 , and the ir length @xmath314 scales as @xmath315 for both modes .
the cpa self - consistency equation for low frequency in ref .
@xcite , eq .
( 7 ) , can be rewritten in the form @xmath316 by ignoring the @xmath317 term which correspond to rayleigh scattering in the second subequation .
the solution to this equation @xmath318 has a form very similar to the that of the effective medium @xmath4 spring constant @xmath27 in the kagome lattice as shown in eq .
( [ eq : cpaasymsolu ] ) . although @xmath319 in the amorphous solid and @xmath320 in the kagome lattice scale differently with @xmath312 , the frequency dependence of @xmath321 and @xmath322 are almost identical : they are both of the form @xmath323 where @xmath324 and @xmath78 are constants and @xmath325 .
the @xmath317 term ignored in the above analysis leads to rayleigh scattering and corresponds to the subdominant terms in @xmath185 discussed in sec .
[ sec : cpaff ] .
the difference between @xmath317 in ref .
@xcite and @xmath183 for the kagome lattice is the spatial dimension .
to summarize , we examined the random nearly isostatic kagome lattice via the cpa , we obtained effective - medium @xmath4 spring constant @xmath27 that scales with the occupancy probability @xmath326 of the @xmath4 bonds as @xmath10 at small @xmath12 .
below the characteristic frequency @xmath26 , there is only weak damping of acoustic phonons arising from rayleigh scattering , whereas above @xmath28 scattering increases rapidly and the system shows proximity to the ir limit .
we compare the kagome lattice to other nearly isostatic systems including the square lattice , jammed solids near point j , and a model random isotropic network @xcite .
the characteristic frequency scale @xmath298 , marking both the onset of the plateau of the anomalous modes and the strong scattering of plain wave states , is found to be a universal property of all of these systems .
the elastic modulus @xmath327 and thus the transport length scale depends on the _ network architecture _ and are not universal . _
acknowledgments_this work was supported in part by nsf - dmr-0804900 .
to construct the dynamical matrix of the kagome lattice , we use the form of the elastic energy given in eq .
( [ eq : deltvuugene ] ) .
because we consider the reference state of all bonds at their rest length , we have @xmath328 , thus there is only projection of @xmath329 onto the direction along the bond .
we first consider the case of simple lattice with one particle in each unit cell and rewrite eq .
( [ eq : deltvuugene ] ) as @xmath330 where @xmath331 labels the two particles connected by the bond @xmath78 .
thus the dynamical matrix @xmath111 , as defined in eq .
( [ eq : defidm ] ) , is given by @xmath332 it is convenient to express the dynamical matrix in momentum space via the fourier transform defined in eq .
( [ eq : ftdefi ] ) @xmath333 where the @xmath334 above the summation of @xmath335 denote a summation over particles connected to @xmath336 , and @xmath337 represent the bonds connected to an arbitrary particle ( note the difference from @xmath78 in the previous equation , which represent all bonds in the system ) .
one can define the dynamical matrix for translational invariant system as @xmath338 where the summation @xmath339 is over bonds connected to an arbitrary particle , and the vector @xmath340 is a convenient way to express the dynamical matrix . for the kagome lattice , which has three particles per unit cell ,
one need to modify the above construction of the dynamical matrix , and in the basis of @xmath341 with particles @xmath342 labeled as in fig .
[ fig : kagome ] , the dynamical matrix can be expressed as @xmath343 with the @xmath116 vectors for @xmath1 bonds for each unit cell ( each bond is counted once ) @xmath344 and the @xmath116 vectors for @xmath4 bonds for each unit cell @xmath345
to calculate the asymptotic form of @xmath256 we first simplify the problem by reduce the dynamical matrix into the space of its three low energy modes by integrating out its three high energy modes @xcite .
the resulting low - energy dynamical matrix is conveniently represented in the basis of longitudinal and transverse phonons and the rotational mode ( mode @xmath346 ) @xmath347 in which the dynamical matrix takes the form @xmath348 in leading order of small @xmath20 and quadratic order in @xmath349 ( the cross term of order @xmath350 is considered higher order and has been dropped ) .
eigenmodes of the dynamical matrix are identified by diagonalizing @xmath351 .
strong mixing between the transverse mode and the rotational mode occurs along @xmath135 ( i.e. , @xmath352 ) and symmetry equivalent isostatic directions .
the resulting two eigenvalues ( by diagonalizing the lower right @xmath353 block ) are @xmath354 obtained from the quadratic order of the renormalized @xmath121 matrix @xmath351 .
the lower eigenvalue @xmath355 correspond to the _ anomalous mode _ , which is close to the transverse mode @xmath356 ( which is simply @xmath123 for @xmath135 direction ) for @xmath357 . for @xmath358
this anomalous mode corresponds to the linear combination of @xmath359 , which is actually the floppy mode of the kagome lattice in the @xmath231 limit , in which @xmath360 .
the function @xmath256 , as given in eq .
( [ eq : fdef ] ) , can be analyzed using the simplified dynamical matrix ( [ eq : dmsplf ] ) , which is the leading order form in small @xmath27 and @xmath349 .
thus we can obtain an asymptotic analytical calculation of the integral @xmath361 by projecting from the @xmath178-dimensional basis in eq .
( [ eq : basissix ] ) onto the @xmath70-dimensional basis in eq .
( [ eq : dmsplfbs ] ) built from the three low energy modes of the system @xmath362 where @xmath363 is the orthogonal transformation from the basis of @xmath364 to the basis @xmath365 in eq .
( [ eq : dmsplfbs ] ) with the longitudinal , transverse , and the rotational mode . in these new basis ,
the dynamical matrix is modified by integrating out the high energy modes and keeping to leading order in small @xmath27 and @xmath349 , which lead to the simple form of eq .
( [ eq : dmsplf ] ) @xcite , and thus the green s function can be analyzed correspondingly .
note that we use the green s function @xmath366 calculated from the renormalized dynamical matrix ( [ eq : dmsplf ] ) , so it is different from the bare value @xmath367 .
the transformed the @xmath368 vector in the basis of @xmath365 takes the form @xmath369 the leading order term of this integral in small @xmath27 is from @xmath370 , the anomalous mode , which has a small frequency of order @xmath371 over the whole range of momentum from @xmath372 to the edge of the brillouin zone along the isostatic directions , and thus correspond to diverging contributions to the @xmath185 integral in small @xmath27 . for an approximation of the @xmath185 integral at small @xmath27 , we use the dynamical matrix of the form ( [ eq : dmsplf ] ) , which kept to leading order in @xmath27 and quadratic order in @xmath349 . at small momentum
, the dynamical matrix ( [ eq : dmsplf ] ) is diagonalized by the basis @xmath365 , and the green s function @xmath211 takes the form of a diagonal matrix @xmath373 which is isotropic and valid for small momentum @xmath374 .
thus , the small momentum region contribute to @xmath185 the following terms @xmath375 at large momentum , the dynamical matrix can be diagonalized to leading order in @xmath27 in the basis @xmath376 , in which @xmath377 takes the form @xmath378 and the green s function @xmath211 takes the form of a diagonal matrix @xmath379 -\frac{3q_x^2}{16}}\big),\end{aligned}\ ] ] which is for the direction of @xmath135 , and we have used the approximated form [ eq : omegaa ] of @xmath380 , that represent the dispersion relation of the anomalous mode at large frequency , as depicted in fig . [
fig : isosmode ] .
for this calculation we use the small @xmath27 values @xmath381 and @xmath382 . for the other two directions
one should change the third term above from @xmath383 into the perpendicular direction of the two isostatic directions accordingly .
thus we need to divide the first brillouin zone into 3 parts : @xmath384 , @xmath385 , and @xmath386 , and integrate each of them out separately and then calculate the sum . here
we just do the @xmath384 part as an example , which uses the form of the green s function in eq .
( [ eq : gflm ] ) .
this part of the integral is @xmath387 and the integral for the other two directions can be calculated similarly .
the leading order contribution to @xmath256 in small @xmath27 is from the third term in eq .
( [ eq : flarge ] ) , which represent the isostatic mode .
we first consider the @xmath220 case , for which the leading order term of @xmath256 is @xmath388 where we took the limit of @xmath231 and make use of the identity @xmath389 .
adding up the contribution from @xmath390 and @xmath391 part we have @xmath392 other terms in eq .
( [ eq : fsmall ] ) and ( [ eq : flarge ] ) contribute higher order terms in small @xmath27 , and is discussed in sec .
[ sec : highorde ] . in the case of @xmath245 ,
the leading order term can be calculated in a similar way @xmath393 the @xmath394 integral can either be evaluated using the @xmath395 function trick by assuming an infinitesimal imaginary part of @xmath200 ( @xmath396 ) , or by extending the integral limit of @xmath394 to @xmath397 ( because the integrand decays fast when @xmath394 is large ) and using contour integral .
we also assumed that @xmath398 to make the simplification that @xmath399 . adding up the contribution from @xmath390 and @xmath391 part we have @xmath400 other terms in eq .
( [ eq : fsmall ] ) and ( [ eq : flarge ] ) contribute higher order terms in small @xmath27 , and is discussed in sec .
[ sec : highorde ] . to get the correction to the asymptotic solution of @xmath23 as in ( [ eq : fffasym ] ) , in particular the small imaginary part rather than zero at small frequency , we calculate the imaginary part of @xmath185 at small frequencies and solve for the correction to @xmath23 perturbatively in the cpa equation . because we consider small frequencies @xmath401 , the contribution is from the two acoustic modes , which are isotropic , and thus can be calculated as @xmath402 and @xmath403 thus we have the correction to eq .
( [ eq : fffasym ] ) that is valid for small @xmath200 as @xmath404 we then solve the leading order cpa equation in small @xmath27 nonaffine regime perturbatively using this corrected form of @xmath185 at small @xmath200 , and get @xmath405 where @xmath406 is the zeroth order solution ( [ eq : cpaasymsolu ] ) .
this correction is very small and can not be observed in our numerical solutions within precision . | the kagome lattice has coordination number @xmath0 , and it is mechanically isostatic when nearest neighbor ( @xmath1 ) sites are connected by central force springs .
a lattice of @xmath2 sites has @xmath3 zero - frequency floppy modes that convert to finite - frequency anomalous modes when next - nearest - neighbor ( @xmath4 ) springs are added .
we use the coherent potential approximation ( cpa ) to study the mode structure and mechanical properties of the kagome lattice in which @xmath4 springs with spring constant @xmath5 are added with probability @xmath6 , where @xmath7 and @xmath8 is the average coordination number .
the effective medium static @xmath4 spring constant @xmath9 scales as @xmath10 for @xmath11 and as @xmath12 for @xmath13 , yielding a frequency scale @xmath14 and a length scale @xmath15 . to a very good approximation at at small nonzero frequency , @xmath16 is a scaling function of @xmath17 .
the ioffe - regel limit beyond which plane - wave states becomes ill - define is reached at a frequency of order @xmath18 . |
the lower levels of the atmosphere are usually strongly influenced by the earth s surface .
known as the atmospheric boundary layer , this is the part of the atmosphere where the surface influences the temperature , moisture , and velocity of the air above through the turbulent transfer of mass .
the stability of the atmospheric boundary layer depends on the profiles of the density and the temperature as a function of the height above the ground . during normal summer days
the land mass warms up and the temperature is higher at lower elevations . if it were not for the decrease in density of the air as a function of the height , such a situation of heating from below would have been always highly unstable .
in fact , the boundary layer is considered stable as long as the temperature decreases at the dry adiabatic lapse rate ( @xmath0 per kilometer ) throughout most of the boundary layer .
with such a rate of cooling one balances out the decrease in density . with a higher degree of cooling one refers to the atmospheric boundary layer as unstably stratified , whereas with a lower degree of cooling the situation is stably stratified .
stably stratified boundary layer occurs typically during clear , calm nights . in extreme cases turbulence tends to cease , and radiational cooling from the surface results in a temperature that increases with height above the surface .
the tendency of the atmosphere to be turbulent does not depend only on the rate of cooling but also on the mean shear in the vertical direction .
the commonly used parameter to describe the tendency of the atmosphere to be turbulent is the
gradient " richardson number ( richardson , 1920 ) , defined as @xmath1 ^ 2 } \,,\ ] ] where @xmath2 is the stream - wise direction , @xmath3 is the height above the ground , @xmath4 is the mean potential temperature profile , ( which differs from the mean temperature profile @xmath5 by accounting for the adiabatic cooling of the air during its expansion : @xmath6 ) , @xmath7 is the buoyancy parameter in which @xmath8 is the adiabatic thermal expansion coefficient ( for an ideal gas @xmath9 ) , and @xmath10 is the gravitational acceleration . the mean shear @xmath11 is defined in terms of the mean velocity @xmath12 , which in the simplest case of flat geometry depends only on the vertical coordinate @xmath3 .
the parameter @xmath13 represents the ratio of the generation or suppression of turbulence by buoyant production of energy to the mechanical generation of energy by wind shear .
this paper is an extended presentation of an invited talk given on international conference
turbulent mixing and beyond " devoted , in particular , to the problems of fluid dynamics , turbulence , geophysics and statistics , that are long - standing challenging tasks .
here we consider the description of stably stratified turbulent boundary layers ( tbl ) , taking as an example the case of stable thermal stratification .
since the 50 s of twentieth century , traditional models of stratified tbl generalize models of unstratified tbl , based on the budget equations for the kinetic energy and mechanical momentum ; see reviews of umlauf and burchard ( 2005 ) , weng and taylor ( 2003 ) .
the main difficulty is that the budget equations are not closed ; they involve turbulent fluxes of mechanical moments @xmath14 ( known as the reynolds stress " tensor ) and a thermal flux @xmath15 ( for the case of thermal stratification ) : @xmath16 where @xmath17 and @xmath18 stand for the turbulent fluctuating velocity and the potential temperature with zero mean .
the nature of the averaging procedure behind the symbol @xmath19 will be specified below .
earlier estimates of the fluxes ( [ def3 ] ) are based on the concept of the down - gradient turbulent transport , in which , similarly to the case of molecular transport , a flux is taken proportional to the gradient of transported property times a corresponding ( turbulent ) transport coefficient : [ dt ] @xmath20 here the turbulent - eddy viscosity @xmath21 and turbulent thermal conductivity @xmath22 are estimated by dimensional reasoning via the vertical turbulent velocity @xmath23 and a scale @xmath24 ( which in the simplest case is determined by the elevation @xmath3 ) .
the dimensionless coefficients @xmath25 and @xmath26 are assumed to be of the order of unity .
this approach meets serious difficulties ( zeman , 1981 ) , in particular , it predicts full suppression of turbulence when the stratification exceeds a critical level , for which @xmath27 . on the other hand , in observations of the atmospheric turbulent boundary layer turbulence exists for much larger values than @xmath28 : experimentally above @xmath29 and even more ( see galperin et al .
( 2007 ) and references therein ) . in models for weather predictions
this problem is fixed " by introducing fit functions @xmath30 and @xmath31 instead of the constant @xmath32 and @xmath26 in the model parametrization ( [ dt ] ) .
this technical solution " is not based on any physical derivation and just masks the shortcomings of the model . to really solve the problem one has to understand its physical origin , even though from a purely formal viewpoint it is indeed possible that a dimensionless coefficient like @xmath26 can be any function of @xmath13 . to expose the physical reason for the failure of the down - gradient approach , recall that in a stratified flow , in the presence of gravity , the turbulent kinetic energy is _ not _ an integral of motion .
only the total mechanical energy , the sum of the kinetic and the potential energy , is conserved in the inviscid limit . as it was shown already by richardson ,
the difficult point is that an important contribution to the potential energy comes not just from the mean density profile , but from the density fluctuations . clearly ,
any reasonable model of the turbulent boundary layer must obey the conservation laws .
the physical requirement of conserving the total mechanical energy calls for an explicit consideration not only of the mean profiles , but also of _ all _ the relevant second - order , one - point , simultaneous correlation functions of _ all _ the fluctuating fields together with some closure procedure for the appearing third order moments .
first of all , in order to account for the important effect of stratification on the anisotropy , we must write explicit equations for the entire reynolds stress tensor , @xmath33 .
next , in the case of the temperature stratified turbulent boundary layers we follow tradition [ see , e.g. zeeman ( 1981 ) , hunt et al .
( 1988 ) , schumann and gerz ( 1995 ) , hanazaki and hunt ( 2004 ) , keller and van atta ( 2000 ) , stretch et al .
( 2001 ) , elperin et al .
( 2002 ) , cheng et al .
( 2002 ) luyten et al .
( 2002 ) , and rehmann and hwang ( 2005 ) ] and account for the turbulent potential energy which is proportional to the variance of the potential temperature deviation , @xmath34 . and last but not least , we have to consider explicitly equations for the vertical fluxes , @xmath35 and @xmath36 , which include the down gradient terms proportional to the velocity and temperature gradients , and counter - gradient terms , proportional to @xmath37 ( in the equation for @xmath35 ) and to @xmath38 ( in the equation for @xmath36 ) .
unfortunately , the resulting second order closure seems to be inconsistent with the variety of boundary - layer data , and many authors took the liberty to introduce additional fitting parameters and sometimes fitting functions to achieve a better agreement with the data ( see reviews of umlauf and burchard ( 2005 ) , weng and taylor ( 2003 ) , zeeman , ( 1981 ) , melor and yamada ( 1974 ) , and references therein ) . moreover , in the second order closures the problem of critical richardson number seems to persists ( cheng et al .
, 2002 ; canuto , 2002 ) .
notice that in spite of obvious inconsistency of the first - order schemes , most of the practically used turbulent models are based on the concept of the down - gradient transport .
one of the reasons is that in the second - order schemes instead of two down - gradient equations ( [ dt ] ) one needs to take into account eight nonlinear coupled additional equations i.e. four equations for the reynolds stresses , three equations for the heat fluxes and equation for the temperature variance . as the result , the second - order schemes have seemed to be rather cumbersome for comprehensive analytical treatment and have allowed to find only some relationships between correlation functions ( see , e.g. , cheng et al . , 2002 ) .
unfortunately , the numerical solutions to the complete set of the second - order schemes equations which involve too many fitting parameters are much less informative in clarification of physical picture of the phenomenon than desired analytical ones .
in this paper we suggest a relatively simple second - order closure model of turbulent boundary layer with stable temperature stratification that , from one hand , accounts for main relevant physics in the stratified tbl and , from the other hand , is simple enough to allow complete analytical treatment including the problem of critical ri@xmath39 . to reach this goal
we approximate the third order correlations via the first- and second - order ones , accounting only for the most physically important terms .
we will try to expose the approximations in a clear and logical way , providing the physical justification as we go along .
resulting second - order model consist of nine coupled equations for the mean velocity and temperature gradients , four components of the reynolds stresses , two components of the temperature fluxes and the temperature variance .
thanks to the achieved simplicity of the model we found an approximate analytical solution of these equations , expressing all nine correlations as functions of only one governing parameter , @xmath40 , where @xmath41 is the outer scale of turbulence ( depending on the elevation @xmath3 and also known as the dissipation scale " ) and @xmath42 is the obukhov length .
we would like also to stress , that in our approach @xmath40 is an external parameter of the problem . for small elevations @xmath43
, it is well accepted that @xmath41 is proportional to @xmath3 , while the @xmath41 dependence is still under debate for @xmath3 comparable or exceeding @xmath42 . for @xmath44 the assignment and discussion of the actual dependence of the outer scale of turbulence , @xmath41 , which is manifested in the nature is out of the scope of this paper , and is remained for future work . at time being , we can analyze consequences of our approach for the following versions of @xmath41 dependence at @xmath45 : + function @xmath41 is saturated at some level of the order of @xmath42 . for concreteness we can take @xmath46 @xmath41 is again proportional to @xmath3 for elevations much larger than @xmath42 : @xmath47 but with the proportionality constant @xmath48 .
if so , we can also study the case @xmath49 even though such a condition may not be realizable in nature . in that case
our analysis of the limit @xmath49 has only a methodological character : it allows to derive an approximate analytic solution for all the objects of interest as functions of @xmath40 that is also valid for the outer scale of turbulence not exceeding a value of the order of @xmath42 .
it should be noticed that traditional turbulent closures ( including ours ) can not be applied for strongly stratified flows with @xmath50 ( may be even at @xmath51 ) .
the main reason is that these closures are roughly justified for developed _
vortical _ turbulence , in which the eddy - turnover time is of the order of its life time ; in other words , there are no well defined quasi - particles " or waves .
this is not the case for stable stratification with @xmath50 , in which the brunt - visl frequency @xmath52 is larger then the eddy - turnover frequency @xmath53 .
it means that for @xmath54 there are weakly decaying kelvin - helmoholtz internal gravity waves ( with characteristic frequency @xmath55 and decay time above @xmath56 ) , propagating on large distances , essentially effecting on tbl , as pointed out by zilitinkevich , ( 2002 ) .
we concentrate in our paper on self - consistent description of the lower part of the atmospheric tbl , in which turbulence has vortical character and consequently , large values of @xmath13 do not appear .
we relate large values of @xmath13 in the upper part of tbl with contributions of the internal gravity waves to the energy and the energy flux in tbl , to the momentum flux , and to the production of ( vortical ) turbulent energy . due to their instability in a shear flow , the waves can break and create turbulent kinetic energy .
all these effects are beyond the framework of our paper .
their description in the upper potential - wave " tbl and intermediate region with the combined vortical - potential " turbulent velocity field is in our nearest agenda . to make the paper more transparent for wide audience , not necessarily experts in atmospheric tbl
, we attempt to present the material in a self - contained manner , and organized it as follows . in sect .
[ s : balance]a we use the oberbeck - boussinesq approximation and apply the standard reynolds decomposition ( into mean values and turbulent zero - mean fluctuations of the velocity and temperature fields ) to derive equations for the mean values and balance equations for all relevant second - order correlation functions . in sect .
[ s : balance]b we demonstrate that the resulting balance equations exactly preserve ( in the non - dissipative limit ) the total mechanical energy of the system , that consists of the kinetic energy of the mean flow , kinetic energy and potential energy of the turbulent subsystem . in sect .
[ s : closure ] we describe the proposed closure procedure that results in a model of stably stratified tbl , that accounts explicitly for all relevant second - order correlations .
the third order correlations which appear in the theory are modeled in terms of second - order correlations in sects .
[ s : closure]a and b. further simplifications are presented in sects .
[ s : closure]c and d for stationary turbulent flows in a plane geometry outside the viscous and buffer layers . in sect .
[ s : closure]e we suggest a generalization of the standard wall - normalization " to obtain the model equations in a dimensionless form with only one governing parameter , @xmath40 . section [ ss : strat41 ] contains approximate analytical solution of the model .
it is shown that the analytical solution deviates from the numerical counterpart in less than a few percent in the entire interval @xmath57 . the last sect .
[ s : res ] is devoted to a detailed description of our results : profiles of the mean velocity and potential temperature ( sect .
[ s : res]a ) , profiles of the turbulent kinetic and temperature " energies , profiles of the anisotropy of partial kinetic energies ( sect .
[ s : res]b ) , profiles of the turbulent transport parameters @xmath58 and @xmath59 , profiles of the gradient- and flux - richardson numbers @xmath13 and @xmath60 , and the dependence of the turbulent prandtl number pr@xmath61 vs. @xmath62 and @xmath13 , sect .
[ s : res]c . in conclusion sect .
[ s : res]d , we consider the validity of the down - gradient transport concept ( [ dt ] ) and explain why it is violated in the upper part of tbl .
the problem of critical @xmath13 is also discussed .
the aim of this section is to consider the simplified dynamics of a stably temperature - stratified turbulent boundary layer , aiming finally at an explicit description of the height dependence of important quantities like the mean velocity , mean temperature , turbulent kinetic and potential energies , etc . in general
one expects very different profiles from those known in standard ( unstratified ) wall - bounded turbulence .
we want to focus on these differences and propose that they occur already relatively close to the ground allowing us to neglect ( to the leading order ) the dependence of the density on height and the coriolis force .
we thus begin by simplifying the hydrodynamic equations which are used in this section .
first we briefly overview the derivation of the governing equations in the boussinesq approximation .
the system of hydrodynamic equations describing a fluid in which the temperature is not uniform consists of the navier - stokes equations for the fluid velocity , @xmath63 , a continuity equation for the space and time dependent ( total ) density of the fluid , @xmath64 , and of the heat balance equation for the ( total ) entropy per unit mass , @xmath65 , landau and lifshitz , 1987 .
these equations are considered with boundary conditions that maintain the solution far from the equilibrium state , where @xmath66 .
these boundary conditions are @xmath67 at zero elevation , @xmath68 at a high elevation of a few kilometers .
this reflects the existence of a wind at high elevation , but we do not attempt to model the physical origin of this wind in any detail .
the only important condition with regards to this wind is that it maintains a momentum flux towards the ground that is prescribed as a function of the elevation .
similarly , we assume that a stable temperature stratification is maintained such that the heat flux towards the ground is prescribed as well .
we neglect the viscous entropy production term assuming that the temperature gradients are large enough such that the thermal entropy production term dominates . for simplicity of the presentation
we restrict ourselves by relatively small elevations and disregard the coriolis force ( for more details , see wyngaard , 1992 ) . on the other hand
we assume that the temperature and density gradients in the entire turbulent boundary layer are sufficiently small to allow employment of local thermodynamic equilibrium .
in other words , we assume the validity of the equation of state . as a
basic reference state " ( brs ) denoted hereafter by a subscript @xmath69 " we use the isentropic model of the atmosphere , where the entropy is considered space homogeneous . now assuming smallness of deviations of the density and pressure from their brs values and exploiting the equation of state , one obtains a simplified equation , which is already very close to the standard navier - stokes equation in the boussinesq approximation . introducing ( generalized ) potential temperature , one results in the well - known system of hydrodynamic equations in the boussinesq approximation .
close to the ground , where one can neglect the dependence of the density on height , the system reads : @xmath70here @xmath71 is the convection time derivative , @xmath72 deviation of pressure from brs , @xmath73 is the density in brs , @xmath74 is the buoyancy parameter ( @xmath75 , @xmath76 , @xmath10 is the gravity acceleration and @xmath77 is the thermal expansion coefficient , which is equal to @xmath78 , reciprocal molecular temperature , for an ideal gas ) , @xmath79 is the deviation of the potential temperature from brs value , @xmath80 kinematic viscosity and @xmath81 is the kinematic thermal conductivity . to develop equations for the mean quantities and correlation functions
one applies the reynolds decomposition : @xmath82 .
here the average @xmath83 stands for an averaging over a horizontal plane at a constant elevation .
this leaves the average quantities with a @xmath84 dependence only . substituting in eqs .
( [ goveq ] ) one gets equations of motion for the mean velocity and mean temperature profiles @xmath85 here @xmath86 is the mean convection derivative .
the total ( molecular and turbulent ) momentum and thermal fluxes are @xmath87 where @xmath88 is the reynolds stress tensor describing the turbulent momentum flux , and @xmath89 is the turbulent thermal flux . in order to derive equations for these correlation functions ,
one considers the equations of motion for the fluctuating velocity and temperature : [ fluct]@xmath90 the whole set of the second order correlation functions includes the reynolds stress , @xmath14 , the turbulent thermal flux , @xmath15 , and the temperature energy " @xmath91 , which is denoted and named by analogy with the turbulent kinetic energy density ( per unit mass and unit volume ) , @xmath92 . using ( [ fluct ] ) one gets the following balance equations " : [ corr]@xmath93 here we denoted the dissipations of the reynolds - stress , heat - flux and the temperature energy by @xmath94 the last term on the lhs of each of eqs .
( [ corr ] ) describes _ spatial _ flux of the corresponding quantity . in models of wall
bounded unstratified turbulence it is known that these terms are very small almost everywhere .
we do not have sufficient experience with the stratified counterpart to be able to assert that the same is true here .
nevertheless , for simplicity we are going to neglect these terms .
it is possible to show that the accounting for these terms does not influence much the results .
note that keeping these terms turns the model into a set of differential equations which are very cumbersome to analyze .
this is a serious uncontrolled step in our development , so we cross our fingers and proceed with caution . since these terms are neglected we do not provide here the explicit expressions for @xmath95 , @xmath96 , and @xmath97 . the first term on the rhs of the balance eq .
( [ corra ] ) for the reynolds stresses is the energy production tensor " @xmath98 , describing the production of the turbulent kinetic energy from the kinetic energy of the mean flow , proportional to the gradient of the mean velocity : [ defs1 ] @xmath99 the second term on the rhs of eq .
( [ corra ] ) , @xmath100 , will be referred hereafter to as the energy conversion tensor " .
it describes the conversion of the turbulent kinetic energy into potential energy .
this term is proportional to the buoyancy parameter @xmath101 and the turbulent thermal flux @xmath15 : @xmath102the next term in the rhs of eq .
( [ corra ] ) is known as the pressure - rate - of - strain tensor " : @xmath103 in incompressible turbulence its trace vanishes , therefore @xmath104 does not contribute to the balance of the kinetic energy . as we will show in sec .
[ sss : prs ] , this tensor can be presented as the sum of three contributions ( zeman , 1981 ) , @xmath105 in which @xmath106 is responsible for the nonlinear process of isotropization of turbulence and is traditionally called the return - to - isotropy " , @xmath107 is similar to the energy production tensor ( [ defs1a ] ) and is called isotropization of production " .
a new term , appearing in the stratified flow , @xmath108 , is similar to the energy conversion tensor ( [ defs1b ] ) and will be refereed to as the isotropization of conversion " .
consider the balance of the turbulent thermal flux @xmath15 , eq .
( [ corrb ] ) .
the first term in the rhs , @xmath109 , describes the source of @xmath15 and , by analogy with the energy - production tensor , @xmath110 , is called thermal - flux production vector " . like @xmath110 , eq .
( [ defs1b ] ) , it has the contribution , @xmath111 , proportional to the mean velocity gradient : @xmath112 and two additional contributions , related to the temperature gradient and to the temperature energy " , @xmath113 , and the buoyancy parameter .
one sees , that in contrary to the oversimplified assumption ( [ dth ] ) , the thermal flux in such a turbulent media can not be considered as proportional to the temperature gradient .
it has also a contribution proportional to the velocity gradient and even to the square of the temperature fluctuations .
moreover , the rhs of the flux - balance eq .
( [ corrb ] ) has an additional term , the pressure - temperature - gradient vector " which , similarly to the pressure - rate - of - strain tensor ( [ rof ] ) , can be divided into three parts ( zeman , 1981 ) : @xmath114as we will show in sec .
[ sss : prs ] the first contribution , @xmath115 is responsible for the nonlinear flux of @xmath15 in the space of scales toward smaller scales , similarly to the correlation @xmath116 , which is responsible for the flux of kinetic energy @xmath117 toward smaller scales .
the correlation @xmath118 may be understood as the nonlinear contribution to the dissipation of the thermal flux . correspondingly we will call it renormalization of the thermal - flux dissipation " and will supply it with a superscript
@xmath119 " .
the next two terms in the decomposition ( [ defs1f ] ) are @xmath120 and @xmath121 .
they describe the renormalization of the thermal - flux production terms @xmath122 and @xmath123 , accordingly .
the total mechanical energy of temperature stratified turbulent flows consists of three parts with densities ( per unit mass ) : @xmath124 , where @xmath125 is the density of kinetic energy of the mean flow , @xmath126 is the density of turbulent kinetic energy and @xmath127 is the density of potential energy , associated with turbulent density fluctuation @xmath128 , caused by the ( potential ) temperature fluctuations @xmath18 , and @xmath129 . the balance eq . for @xmath130 follows from eq .
( [ mean ] ) : [ bale]@xmath131 + \tau_{ij}\nabla_{\!\ !
j } \ , u_i\,,\end{aligned}\ ] ] with the help of identity : @xmath132 and definition ( [ def - tau ] ) .
the terms on the lhs of this eq . , proportional to @xmath80 and @xmath133 respectively , describe the dissipation and the spatial flux of @xmath134 .
the term [ source @xmath135 on the rhs of eq .
( [ baletm ] ) describes the external source of energy , originating from the boundary conditions described above , and @xmath136 describes the kinetic energy out - flux from the mean flow to turbulent subsystem .
the balance eq . for the turbulent kinetic energy follows directly from eq .
( [ corra ] ) : @xmath137/2=- \tau_{ij } \nabla_{\!\ !
j } \ , u_i + \beta f_z\ . \]]on
the lhs of eq .
( [ balett ] ) one sees the dissipation and spatial flux terms . the first term on the rhs originates from the energy production , @xmath138 , defined by eq .
( [ defs1a ] ) .
this term has an opposite sign to the last term on the rhs of eq .
( [ baletm ] ) and describes the production of the turbulent kinetic energy from the kinetic energy of the mean flow .
the last term on the rhs of eq .
( [ balett ] ) originates from the energy conversion tensor @xmath139 , eq .
( [ defs1b ] ) , and describes the conversion of the turbulent kinetic energy into potential one . according to the last of eqs .
( [ corr ] ) , one gets the balance equation for the potential energy @xmath140 ; multiplying eq .
( [ corrc ] ) for @xmath113 by @xmath141:@xmath142/s_\theta = -\beta f_z\ .\ ] ] the rhs of this eq . [ coinciding up to a sign with the last term on the rhs of eq .
( [ balett ] ) ] is the source of potential energy ( from the kinetic one ) . in the sum of the three balance equations , the conversion terms ( of the kinetic energy from the mean to turbulent flows and of the turbulent kinetic energy to the potential one )
cancel and one gets the total mechanical energy balance : @xmath143 + { { \bm{\nabla}}}\ , [ \mbox{flux } e]= [ \mbox{source } e_{{{\mathcal{k}}}}]\ .\ ] ] this equation exactly respects the conservation of total mechanical energy in the dissipation - less limit , irrespective of the closure approximations .
this is because the energy production and conversion terms are exact and do not require any closures , while the pressure - rate - of - strain tensor , that requires some closure , does not contribute to the total energy balance .
in this section we describe the proposed closure procedure that results in a model of stably stratified tbl . in developing this model
we strongly rely on the analogous well developed modeling of standard ( unstratified ) tbl .
the final justification of this approach can be done only in comparison to data from experiments and simulations .
we will do below what we can to use the existing data , but we propose at this point that much more experimental and simulational work is necessary to solidify all the steps taken in this section . the correlation functions @xmath144 and @xmath145 , defined by eqs .
( [ defs1c ] ) and ( [ defs1f ] ) , include fluctuating part of the pressure @xmath146 .
the poisson s equation for @xmath146 follows from eq .
( [ fluct ] ) : @xmath147 $ ] .
the solution of this equation includes a harmonic part , @xmath148 , which is responsible for sound propagation and does not contribute to turbulent dynamics at small mach numbers .
thus this contribution can be neglected .
the inhomogeneous solution includes three parts @xmath149 $ ] , where @xmath150 and the inverse laplace operator @xmath151 is defined as usual in terms of an integral over the green s function .
correspondingly the correlations @xmath144 and @xmath145 consist of three terms , eqs .
( [ rof ] ) and ( [ defs1f ] ) , in which @xmath152 all of those terms originating from @xmath153 are the most problematic because they introduce coupling to triple correlation functions : @xmath154 and @xmath155 .
thus they require closure procedures whose justification can be only tested a - posteriori against the data .
having in mind to simplify the model in most possible manner , we adopt for the diagonal part of the return - to - isotropy tensor , the simplest rota form ( rotta , 1951 ) [ rof]@xmath156 in which @xmath157 is the relaxation frequency of diagonal components of the reynolds - stress tensor toward its isotropic form , @xmath158 .
the parametrization of @xmath157 will be discussed later .
the tensor @xmath106 is traceless , therefore the frequency @xmath157 must be the same for all the diagonal components of @xmath159 .
on the other hand there are no reasons to assume that off - diagonal terms have the same relaxation frequency .
therefore , following lvov et al .
( 2006a ) we assume that @xmath160with , generally speaking , @xmath161 . moreover
, on the intuitive level , we can expect that off - diagonal terms should decay faster then the diagonal ones , i.e. @xmath162 .
indeed , our analysis of dns results shows that @xmath163 ( lvov et al . , 2006b ) .
the term @xmath164 also describes return - to - isotropy due to nonlinear turbulence self interactions ( zeman , 1981 ) , and may be modeled as : @xmath165 this equation dictates the vectorial structure of @xmath166 , which will be confirmed below .
the rest can be understood as the definition of the @xmath167 as the relaxation frequency of the thermal flux .
its parametrization is the subject of further discussion in sec .
[ ss : closure ] .
the traceless isotropization - of - production " tensor , @xmath107 , has a very similar structure to the production tensor , @xmath98 , eq .
( [ defs1a ] ) , and thus is traditionally modeled in terms of @xmath98 ( pope , 2001):@xmath168the accepted value of the numerical constant @xmath169 ( pope , 2001 ) .
the traceless isotropization - of - conversion " tensor , @xmath108 does not exist in unstratified tbl .
its structure is very similar to the conversion tensor , @xmath170 , eq . ( [ defs1b ] ) .
therefore it is reasonable to model it in the same way in terms of @xmath170 ( zeman , 1981 ) : @xmath171with some new constant @xmath172 .
the renormalization of production terms @xmath173 and @xmath174 are very similar to the corresponding thermal flux production terms , @xmath175 and @xmath176 , defined by eqs .
( [ defofa ] ) .
therefore , in the spirit of eqs .
( [ ip ] ) and ( [ ic ] ) , they are modeled as follows:@xmath177 using this and ( [ pr1 ] ) one finds the sign of @xmath178 : @xmath179to estimate @xmath178 we assume that on the gradient scales the temperature fluctuations are roughly isotropic , and therefore we can estimate @xmath180 . introducing this estimate and integrating by parts leads to @xmath181 . far away from the wall and for large reynolds numbers the dissipation tensors are dominated by the viscous scale motions , at which turbulence can be considered as isotropic . therefore ,
the vector @xmath182 should vanish , while the tensor @xmath183 , eq .
( [ diss ] ) , should be diagonal : [ distau1 ] @xmath184where the numerical prefactor @xmath185 is chosen such that @xmath186 becomes the relaxation frequency of the turbulent kinetic energy . under stationary conditions
the rate of turbulent kinetic energy dissipation is equal to the energy flux through scales , that can be estimated as @xmath187 , where @xmath188 is the outer scale of turbulence .
therefore , the natural estimate of @xmath186 involves the triple - velocity correlator , @xmath189 , exactly in the same manner , as the return - to - isotropy frequencies , @xmath190 and @xmath191 in eqs .
( [ ria ] ) and ( [ rib ] ) .
similarly , @xmath192 in the plane geometry , the equations simplify further .
the mean velocity is oriented in the ( streamwise ) @xmath193 direction and all mean values depend on the vertical ( wall - normal ) coordinate @xmath3 only : @xmath194 , @xmath195 , @xmath196 , @xmath197 , @xmath198 .
therefore @xmath199 , and in the stationary case , when @xmath200 , the mean convective derivative vanishes : @xmath201 .
moreover due to the @xmath202 symmetry of the problem the following correlations vanish : @xmath203 .
the only non - zero components of the mean velocity and temperature gradients are : @xmath204 having in mind eqs . of sec .
[ ss : balance ] and integrating eqs .
( [ mean ] ) for @xmath205 and @xmath206 over @xmath3 , one gets equations for the total ( turbulent and molecular ) mechanical - momentum flux , @xmath207 , and thermal flux , @xmath208 , toward the wall [ tbl - sim ] @xmath209 the total flux of the @xmath2-component of the mechanical moment in @xmath3-direction is @xmath210 . generally speaking ,
@xmath211 depends on @xmath3 . for example , for the pressure driven planar channel flow ( of the half - wight @xmath212 ) @xmath213 . relatively close to the ground , where @xmath214 , the @xmath3 dependence of @xmath215 can be neglected .
in the absence of the mean horizontal pressure drop and spatial distributed heat sources @xmath216 and @xmath208 are @xmath3-independent , and thus equal to their values at zero elevation , as indicated in eqs .
( [ tbl - sim ] ) after @xmath217"-sign .
notice , that in our case of stable stratification both vertical fluxes , the @xmath2-component of the mechanical momentum , @xmath218 , and the thermal flux , @xmath219 , are directed toward the ground , i.e. negative . for the sake of convenience , we introduce in eqs .
( [ tbl - sim ] ) notations for their ( positive ) zero level absolute value : @xmath220 and @xmath221 .
recall that in the plain geometry @xmath222 .
nevertheless one can write an equation for @xmath223 : @xmath224 which describes a turbulent correction ( @xmath225 ) to the hydrostatic equilibrium .
actually , this equation determines the profile of @xmath226 , that does not appear in the system of balance equations ( [ tbl - sim ] ) .
consider first the balance eqs .
( [ corra ] ) for the diagonal components of the reynolds - stress tensor in algebraic model ( which arises when we neglect the spatial fluxes ) : @xmath227where @xmath228 .
the lhs of these equations includes the dissipation and return - to - isotropy terms . on the rhs we have the kinetic energy production and isotropization of production terms ( both proportional to @xmath229 ) together with the conversion and isotropization of conversion terms , that are proportional to the vertical thermal flux @xmath36 .
the horizontal component of the thermal flux , @xmath37 , does not appear in these equations .
system ( [ taup ] ) allows to find anisotropy of the turbulent - velocity fluctuations and to get the balance eqs . for the turbulent kinetic energy with the energy production and conversion terms on the rhs : [ tau1 ] @xmath230 e{_{\scriptscriptstyle\rm{k } } } \\
\nonumber & & -(3 -2c{_{\scriptscriptstyle\rm{ip } } } + c{_{\scriptscriptstyle\rm{ic}}})\ , \beta f_z / \gamma{_{\scriptscriptstyle\rm{ri } } } \big\ } \ , , \\ 3\tau_{yy } & = & 2 \big\ { [ ( c{_{\scriptscriptstyle\rm{ip}}}-1)\gamma_{uu}/\gamma{_{\scriptscriptstyle\rm{ri } } } + 1 ] e{_{\scriptscriptstyle\rm{k } } } \\
\nonumber & & -(c{_{\scriptscriptstyle\rm{ip } } } + c{_{\scriptscriptstyle\rm{ic } } } ) \beta f_z/\gamma{_{\scriptscriptstyle\rm{ri } } } \big\ } \ , , \\
3\tau_{zz } & = & 2 \big\ { [ \left ( c{_{\scriptscriptstyle\rm{ip}}}-1 \right ) \gamma_{uu}/\gamma{_{\scriptscriptstyle\rm{ri } } } + 1 ] e{_{\scriptscriptstyle\rm{k } } } \\
\nonumber & & - ( c{_{\scriptscriptstyle\rm{ip } } } -2c{_{\scriptscriptstyle\rm{ic } } } -3 ) \ , \beta f_z/\gamma{_{\scriptscriptstyle\rm{ri}}}\big\ } \ , , \\ \label{bale } \gamma_{uu}e{_{\scriptscriptstyle\rm{k } } } & = & -\tau_{xz}s_{_u}+ \beta f_z \,,\ ] ] equation ( [ bale ] ) includes the only non - vanishing tangential ( off - diagonal ) reynolds stress @xmath35 and has to be accompanied with an equation for this object : @xmath231 this equation manifests that the tangential reynolds stress @xmath35 , that determines the energy production [ according to eq .
( [ bale ] ) ] , influences , in its turn , on the value of the streamwise thermal flux @xmath37 , which therefore effects on the turbulent kinetic energy production .
as we mentioned , in the plain geometry @xmath232 . equations ( [ corrb ] ) for the @xmath37 and @xmath36 in this case take the form : @xmath233 in which the rhs describes the thermal - flux production , corrected by the isotropization of production terms . the last eq .
( [ corrc ] ) for @xmath234 , represents the balance between the dissipation ( lhs ) and production ( rhs ) : @xmath235 at this point we follow a tradition in modeling of all the nonlinear inverse time - scales by dimensional estimates ( kolmogorov , 1941 ) : @xmath236remember that @xmath237 is the outer scale of turbulence " .
this scale equals to @xmath3 for @xmath238 , where @xmath42 is the obukhov length ( definition is found below ) .
detailed analysis of experimental , dns and les data ( see lvov et al . ,
2006 , and references therein ) shows that for unstratified flows , @xmath239 , the anisotropic boundary layers exhibits values of the reynolds stress tensor that can be well approximated by the values @xmath240 , @xmath241 . in our approach
this dictates the choice @xmath242 .
also we can expect that @xmath243 is almost not affected by buoyancy .
this gives simply @xmath244 .
if so , eqs .
( [ tau1 ] ) with the parametrization ( [ freqs ] ) can be identically rewritten as follows : [ simple ] @xmath245 for completeness we also repeated here the parametrization ( [ freqs ] ) of @xmath186 .
finally we present the version of the balance eqs . for the thermal flux ( [ sim - f - x ] ) , ( [ sim - f - z ] ) , and for the temperature energy " , ( [ sim - et ] ) , after all the simplified assumptions : @xmath246 the analysis of the balance eqs .
( [ simple ] ) is drastically simplified if they are presented in a dimensionless form .
traditionally , the conventional wall units " are introduced via the wall friction velocity @xmath247 , and the viscous length - scale @xmath248 . a wall unit for the temperature @xmath249
is defined via the thermal flux at the wall and friction velocity .
subsequently , @xmath250 , @xmath251 , @xmath252 , @xmath253 , @xmath254 , @xmath255 , etc .
then the governing eqs .
( [ goveq ] ) take the form:@xmath256 these eqs . include two dimensionless parameters : the conventional prandtl number pr@xmath257 , and @xmath258 the obukhov length @xmath42 measured in wall units : @xmath259 , @xmath260 .
we used here the modern definition of the obukhov length , which differs from the old one by the absence of the von - krmn constant @xmath261 in its denominator ( monin and obukhov , 1954 ) . outside of the viscous sub - layer , where the kinematic viscosity and kinematic thermal conductivity can be ignored , @xmath258 is the only dimensionless parameter in the problem , which separates the region of weak stratification , @xmath262 , and the region of strong stratification , where @xmath263 .
given the generalized wall normalization we introduce objects with a superscript @xmath264 " in the usual manner : @xmath265 in the turbulent region , governed by @xmath258 only , eqs . ( [ tbl - sim ] )
simplify to @xmath266 , @xmath267 . outside of the viscous region , where eqs .
( [ simple ] ) were derived , the problem has only one characteristic length , i.e. the obukhov scale @xmath269 .
correspondingly , one expects that the only dimensionless parameter that governs the turbulent statistics in this region should be the ratio of the outer scale of turbulence , @xmath188 , to the obukhov length - scale @xmath42 , which we denote as @xmath270 . indeed , introducing @xmath268-objects " : @xmath271 and using eqs .
( [ plus - m ] ) one rewrites the balance eqs .
( [ simple ] ) as follows : [ mm]@xmath272 these equations are the main result of current sec .
[ s : closure ] .
it may be considered as minimal model " for stably stratified tbl , that respects the conservation of energy , describes anisotropy of turbulence and all relevant fluxes explicitly and , nevertheless is still simple enough to allow comprehensive analytical analysis , that results in an approximate analytical solution ( with reasonable accuracy ) for the mean velocity and temperature gradients @xmath273 and @xmath274 , and all second - order ( cross)-correlation functions .
as expected , the only parameter appearing in the minimal model ( [ mm ] ) is @xmath275 .
the outer scale of turbulence , @xmath188 , does not appear by itself , only via the definition of @xmath275 ( [ ddagb ] ) .
therefore our goal now is to solve eqs .
( [ mm ] ) in order to find five functions of only one argument @xmath276 : @xmath277 , @xmath278 , @xmath279 , @xmath280 and @xmath281 .
after that we can specify the dependence @xmath282 and then reconstruct the @xmath283-dependence of these five objects .
this subsection is devoted to an analytical and numerical analysis of the minimal - model ( [ mm ] ) .
an example of numerical solution of eqs .
( [ mm ] ) ( with some reasonable choice of the phenomenological parameters ) is shown in fig .
[ f : s ] .
nevertheless , it would be much more instructive to have approximate analytical solutions for all correlations that will describe their @xmath275-dependence with reasonable accuracy .
the detailed cumbersome procedure of finding these solutions is skipped here , but a brief overview is as follows .
the eqs . ( [ mm ] ) can be reformulated as a polynomial equation of ninth order for the only unknown @xmath284 .
an analysis of its structure helps to formulate an effective interpolation formula ( [ inter ] ) , discussed below .
hence , we found the solutions of eqs .
( [ mm ] ) at neutral stratification , @xmath285 , corrected up to the linear order in @xmath275 .
its comparison with the existing dns data resulted in an estimate for the constants @xmath286 , and @xmath287 .
then , we considered the region @xmath288 . even though such a condition may not be realizable in nature , from a methodological point of view , as we will see below , it enables to obtain the desired analytical approximation .
the @xmath288 asymptotic solution with corrections , linear in the small parameter @xmath289 , were found .
now we are armed to suggest an interpolation formula [ inter ] @xmath290 that coincides with the exact solutions for @xmath285 and for @xmath291 , including the leading corrections to both asymptotics , linear in @xmath275 , and @xmath292 . moreover , in the region @xmath293 , eq .
( [ intera ] ) accounts for the structure of the exact polynomial . as a result ,
the interpolation formula ( [ intera ] ) is close to the numerical solution with deviations smaller than 3% in the entire region @xmath294 , see upper middle panel on fig .
[ f : s ] . together with eq .
( [ 9-eqc ] ) it produces a solution for @xmath295 , that can be written as @xmath296 where @xmath297 and @xmath261 is the von - krmn constant .
this formula gives the same accuracy @xmath298 , see upper left panel in fig .
[ f : s ] .
we demonstrate below that the proposed interpolation formulae describe the @xmath275-dependence of the correlations with a very reasonable accuracy , about @xmath299 , for any @xmath300 , see black dashed lines in figs .
[ f : s ] .
unfortunately , a direct substitution of the interpolation formula ( [ inter ] ) into the exact relation for @xmath301 obtained from the system ( [ mm ] ) works well only for small @xmath302 , in spite of the fact that the interpolation formula is rather accurate in the whole region .
we need therefore to derive an independent interpolation formula for @xmath301 . using expansions for small @xmath303 and large @xmath304 we suggest @xmath305
in which @xmath306 and @xmath307 satisfies @xmath308with @xmath309 equation ( [ inter1a ] ) is constructed such that the leading and sub - leading asymptotics for small and large @xmath310 coincide with the first two terms in the exact expansions at `` almost '' neural stratification and extremely strong stratification . as a result ,
( [ inter1a ] ) approximates the exact solution with errors smaller then 5% for @xmath311 and @xmath312 and with errors
smaller than 10% for any @xmath310 , see lower left panel in figure [ f : s ] . substituting the approximate eqs .
( [ inter ] ) into the exact relations ( [ mm ] ) , one gets approximate solutions @xmath313 and @xmath314 with errors smaller than 10% , see rightmost panels in figure [ f : s ] . + in principle , integrating the mean shear @xmath315 and the mean temperature gradient @xmath316 , one can find the mean velocity and temperature profiles .
unfortunately , to do so we need to know @xmath317 and @xmath316 as functions of the elevation @xmath3 , while in our approach they are found as functions of @xmath62 .
remember , that the external parameter @xmath188 is the outer scale of turbulence that depends on the elevation @xmath3 .
the importance of an accounting for the proper physically motivated dependance of @xmath188 on @xmath3 for an example of channels and pipes has been recently shown by lvov et al .
( 2008 ) . for the problem at hands
, we can safely take @xmath318 if @xmath319 , however when @xmath320 the function @xmath41 is not found theoretically although it was discussed phenomenologically with support of observational , experimental and numerical data .
it is traditionally believed that for @xmath321 the scale @xmath188 saturates at some level of order @xmath42 [ see , e.g. eq . ( [ sless ] ) ] .
the resulting plots of @xmath322 are shown on figure [ f : means ] , left panel . even taking @xmath323
one gets a very similar velocity profile , see figure [ f : means ] , right panel .
with @xmath323 we found an analytical expression for the mean - velocity profile using the interpolation eq .
( [ interb ] ) for @xmath315 : @xmath324+\frac{z } { l_1 } \ . \ ] ] here @xmath325 is the roughness length .
the resulting mean velocity profiles have logarithmic asymptotic for @xmath326 and a linear behavior for @xmath320 in agreement with meteorological observations .
usually the observations are parameterized by a so - called log - linear approximation ( monin and obukhov , 1954 ) : [ ll - apr ] @xmath327 which is plotted in figure [ f : means ] by dotted lines .
one sees some deviation in the region of intermediate @xmath3 .
the reason is that the real profile [ see , e.g. eq . ( [ vel ] ) ] has a logarithmic term that saturates for @xmath45 , while in the approximation ( [ ll - aprv ] ) this term continues to grow . to fix this one
can use eq .
( [ vel ] ) ( with @xmath328 for simplicity ) , or even its simplified version @xmath329 this approximation is plotted as a dashed line on figure [ f : means ] for comparison .
one sees that the approximation ( [ ll - aprv2 ] ) works much better than the traditional one .
thus we suggest eq .
( [ ll - aprv2 ] ) for parameterizing meteorological observations .
the temperature profiles in our approach look similar to the velocity ones : they have logarithmic asymptotic for @xmath330 and linear behavior for @xmath331 .
correspondingly , they can be fitted by a log - linear approximation , like ( [ ll - aprv ] ) , or even better , by an improved version of it , like eq .
( [ ll - aprv2 ] ) .
clearly , the values of constants will be different : @xmath332 , @xmath333 , etc .
the computed profiles of the turbulent kinetic and temperature energies , horizontal thermal flux profile and the anisotropy profiles are shown on figure [ f : s ] in the middle and right panels .
the anisotropy profiles , lower middle panel , saturate at @xmath334 , therefore they are not sensitive to the @xmath3-dependence of @xmath41 ; even quantitatively one can think of these profiles as if they were plotted as a function of @xmath335 .
another issue is the profiles of @xmath336 ( upper middle panel ) and of @xmath337 and @xmath314 ( rightmost panels ) , that are @xmath338 for @xmath339 ( if realizable ) . with the interpolation formula ( [ sless ] )
the profiles of the second order correlations have to saturate at levels corresponding to @xmath340 .
this sensitivity to the @xmath3-dependence of @xmath41 makes a comparison of the prediction with experimental data very desirable . in our notations the turbulent viscosity and thermal conductivity , turbulent prandtl number , the gradient- and flux - richardson numbers are [ num ] @xmath341 with eqs .
( [ nu ] ) and ( [ chi ] ) we introduce also two dimensionless functions @xmath342 and @xmath343 that are taken as @xmath344-independent constants in the down - gradient transport approximation ( [ dt ] ) described in the introduction .
we will show , however , that these functions have a strong dependence on @xmath344 , going to zero in the limit @xmath345 as @xmath346 .
therefore this approximation is not valid for large @xmath344 even qualitatively .
+ as we mentioned in the introduction , the concept of the down - gradient transport assumes that the momentum and thermal fluxes are proportional to the mean velocity and temperature gradients , see eqs .
( [ dt ] ) : @xmath347 where @xmath58 and @xmath59 are effective turbulent viscosity and thermal conductivity , that can be estimated by dimensional reasoning .
equations ( [ dt ] ) , giving this estimates , include additional physical arguments that vertical transport parameters should be estimated via vertical turbulent velocity , @xmath23 , and characteristic vertical scale of turbulence , @xmath24 .
the relations between the scales @xmath348 in different @xmath349-directions in anisotropic turbulence can be found in the approximation of time - isotropy , according to which @xmath350 here @xmath53 is a characteristic isotropic frequency of turbulence , that for concreteness can be taken as the kinetic energy relaxation frequency @xmath351 .
the approximation ( [ ti ] ) is supported by experimental data , according to which in anisotropic turbulence the ratios @xmath352 ( @xmath353 ) are larger then the ratios @xmath354 that are close to unity . with this approximations @xmath58 and @xmath59 can be estimated as follows:@xmath355 where , according to the approximation of down - gradient transport , the dimensionless parameters @xmath25 and @xmath26 are taken as constants , independent of the level of stratification . in order to check how the approximation ( [ appr ] )
, ( [ est1 ] ) works in the stratified tbl for both fluxes , one can consider eqs .
( [ appr ] ) as _ definitions _ of @xmath58 and @xmath59 and eqs .
( [ est1 ] ) as _ definitions _ of @xmath25 and @xmath25 .
this gives [ defc ] @xmath356 recall , that in this paper the down - gradient approximation is not used at all .
instead , we are using exact balance equations for all relevant second order correlations , including @xmath35 and @xmath37 . substituting our results in the rhs of the definitions ( [ defc ] ) we can find , how @xmath25 and @xmath26 depend on @xmath357 that determines the level of stratification in our approach .
the resulting plots of the ratios @xmath358 and @xmath359 are shown in the leftmost panel in figure [ f : res ] .
one sees that the @xmath360 and @xmath361 can be considered approximately as constants only for @xmath362 . for larger
@xmath363 both @xmath360 and @xmath361 rapidly decrease , more or less in the same manner , diminishing by an order of magnitude already for @xmath364 .
for larger @xmath62 one can use the asymptotic solution according to which @xmath365 this means that both functions vanish as @xmath366 : @xmath367 where numerical prefactors account for the accepted values of the dimensionless fit parameters .
the physical reason for the strong dependence of @xmath368 and @xmath26 on stratification is as follows : in the rhs of eq .
( [ txz ] ) for the momentum flux and eq .
( [ sim - f - z ] ) for the vertical heat flux there are two terms . the first ones , proportional to @xmath369 and velocity ( or temperature ) gradients correspond to the approximation ( [ appr ] ) , giving ( in our notations ) @xmath370const and @xmath371const , in agreement with the down - gradient transport concept . however , there are second contributions to the vertical momentum flux @xmath372 and to the vertical heat flux , that is proportional to @xmath373 . in our approach
both contributions are negative , giving rise to the _ counter - gradient fluxes_. what follows from our approach , is that these counter - gradient fluxes cancel ( to the leading order ) the down - gradient contributions in the limit @xmath345 . as a result , in this limit
the effective turbulent diffusion and thermal conductivity vanish , making the down - gradient approximation for them ( with constant @xmath25 and @xmath26 ) irrelevant even qualitatively for @xmath374 .
in our picture of stable temperature - stratified tbl , the turbulence exists at any elevations , where one can neglect the coriolis force .
moreover , the turbulent kinetic and temperature energies increase as @xmath375 for @xmath376 , see figure [ f : s ] . at the same time , the mean velocity and potential temperature change the @xmath377-dependence from logarithmic lo linear , see figure [ f : means ] and ( modified ) log - linear interpolation formula ( [ ll - aprv2 ] ) .
correspondingly , the shear of the mean velocity and _ the mean temperature gradient _ saturate at some elevation ( and at some @xmath62 ) , and @xmath13 _ saturates as well_. this predictions agree with large eddy simulation by zilitinkevich and esau ( 2006 ) , where @xmath13 can be considered as saturating around 0.4 for @xmath378 .
notice that the turbulent closures of kind used above can not be applied for strongly stratified flows with @xmath50 ( may be even at @xmath51 ) .
there are two reasons for that .
the first one was mentioned in the introduction .
namely , for @xmath50 the brunt - visl frequency @xmath379 , @xmath380 , is larger then the eddy - turnover frequency and therefore there are weakly decaying kelvin - helmoholtz internal gravity waves which , generally speaking , have to be accounted for in the momentum and energy balance equations . the second reason , that makes the results very sensitive to the contribution of internal waves follows from the fact that vortical turbulent fluxes vanish ( at fixed velocity and temperature gradients ) . therefore even relatively small contributions of different nature to the momentum and thermal fluxes may be important .
the final conclusion is that the tbl modeling at large level of stratification requires an accounting for turbulence of the internal waves together with the vortical turbulence . definitely , new observations , laboratory and numerical experiments with control of internal wave activity are very likely .
vl kindly acknowledge the possibility to give an invited lecture on the problems , discussed in this paper , at the international conference `` turbulent mixing and beyond '' , which was held august 2007 at the abdus salam international center for theoretical physics , trieste , italy .
vl and or also acknowledge the support of the transnational access programme at risc - linz , funded by the european commission framework 6 programme for integrated infrastructures initiatives under the project science ( contract no .
let us look more carefully at the approximation ( [ freqs ] ) , which is @xmath381the dimensional reasoning that leads to this approximation is questionable for problems having a dimensionless parameter @xmath344 .
generally speaking , all constants " @xmath382 and @xmath383 in eq .
( [ app - appr ] ) can be any functions of @xmath344 .
presently we just hope that a possible @xmath344 dependence of these functions is relatively weak and does not affect the qualitative picture of the phenomenon . moreover , even the assumption ( [ distau1a ] ) that the dissipation of the thermal flux @xmath384 is proportional to the thermal flux and the assumption ( [ distau1b ] ) that the dissipation of @xmath113 , @xmath385 are also questionable . formally speaking , one can not guarantee that the triple cross - correlator @xmath386 that estimates @xmath387 , can be ( roughly speaking ) decomposed like @xmath388 , i.e really proportional to @xmath389 as it stated in eq .
( [ distau1a ] ) .
theoretically , one can not exclude the decomposition @xmath390 , i.e. a contribution to @xmath391 .
similarly , the dissipation @xmath392 in the balance ( [ corrc ] ) of @xmath113 , that is determined by the correlator ( [ distau1b ] ) , is @xmath393 , as it follows from the decomposition @xmath394 and is stated in eq .
( [ distau1b ] ) .
this correlator allows , for example , the decomposition @xmath395 , i.e. contribution to @xmath396 .
this discussion demonstrates , that the situation with the dissipation rates is not so simple , as one may think and thus requires careful theoretical analysis that is in our agenda for future work .
our preliminary analysis of this problem shows that all fitting constants are indeed functions of @xmath344 .
fortunately , they vary within finite limits in the entire interval @xmath397 .
therefore we propose that the approximations used in this paper preserve the qualitative picture of the phenomenon . once again
, the traditional down - gradient approximation does not work even qualitatively because corresponding constants " @xmath25 and @xmath26 vanish in the limit @xmath345 .
hunt , j. c. r. , stretch , d. d. , and britter , r. e. , 1988 : length scales in stably stratified turbulent flows and their use in turbulence models . in : _
i.m.a . conference on `` stably stratified flow and dense gas dispersion '' _
( j. s. puttock , ed . ) , clarendon press , 285 - 322 .
luyten , p. j. , carniel , s. , and umgiesser , g. , 2002 : validation of turbulence closure parameterisations for stably stratified flows using the provess turbulence measurements in the north sea , _ j. sea research _ , * 47 * , 239 - 267 .
stretch , d. d. , rottman , j. w. , nomura , k. k. , and venayagamoorthy , s. k. , 2001 : transient mixing events in stably stratified turbulence , in : _
14th australasian fluid mechanics conference _ , adelaide , australia , 10 - 14 december 2001 .
zilitinkevich s.s . ,
2002 : third - order transport due to internal waves and non - local turbulence in the stably stratified surface layer , _ quarterly journal of the royal meteorological society _ , * 128 * , 913 - 925 .
zilitinkevich , s.s . ,
elperin , t. , kleeorin , n. , and rogachevskii , i. , 2007 : energy- and flux - budget ( efb ) turbulence closure model for stably stratified flows .
part i : steady - state , homogeneous regimes , _ boundary - layer meteorology _ * 125 * , 167 - 191 .
zilitinkevich , s.s . , and essau , i. : similarity theory and calculation of turbulent fluxes at the surface for the stably stratified atmospheric boundary layer , _ boundary - layer meteorology _ * 125 * , 193 - 205 ( 2007 ) . | we present an extended version of an invited talk given on the international conference turbulent mixing and beyond " .
the dynamical and statistical description of stably stratified turbulent boundary layers with the important example of the stable atmospheric boundary layer in mind is addressed .
traditional approaches to this problem , based on the profiles of mean quantities , velocity second - order correlations , and dimensional estimates of the turbulent thermal flux run into a well known difficulty , predicting the suppression of turbulence at a small critical value of the richardson number , in contradiction with observations .
phenomenological attempts to overcome this problem suffer from various theoretical inconsistencies . here
we present an approach taking into full account all the second - order statistics , which allows us to respect the conservation of total mechanical energy .
the analysis culminates in an analytic solution of the profiles of all mean quantities and all second - order correlations removing the unphysical predictions of previous theories .
we propose that the approach taken here is sufficient to describe the lower parts of the atmospheric boundary layer , as long as the richardson number does not exceed an order of unity . for much higher richardson numbers
the physics may change qualitatively , requiring careful consideration of the potential kelvin - helmoholtz waves and their interaction with the vortical turbulence . * nomenclature * + + [ cols="<,^,<,^,<,^ , < " , ] + + |
as an extension of our previously reported @xcite search for rare and forbidden decays of the @xmath0 charm meson , @xmath24 , we report an analysis using full open - charm data sample for @xmath0 , and also a search for @xmath25 with cleo - c s full @xmath1 data sample . here
, @xmath3 is either @xmath4 or @xmath5 , and charge - conjugate modes are implicit throughout this article .
these decays probe flavor - changing neutral currents ( fcnc ) , in @xmath26 and @xmath27 , and lepton number violations ( lnv ) , in @xmath28 and @xmath29 .
these decays are either highly suppressed or forbidden in the standard model ( sm ) , but can be significantly enhanced by some non - sm physics scenarios @xcite .
standard model short - distance fcnc decays are expected to be of order @xmath30 to @xmath31 @xcite , but long - distance vector - pole induced decays of @xmath0 or @xmath1 @xmath32 ( where @xmath33 is an intermediate vector meson @xmath34 , @xmath35 , or @xmath36 ) are expected to be of order @xmath37 to @xmath38 @xcite . to observe an enhancement in fcnc due to non - sm physics
, we need to search for dielectron mass regions away from the vector poles .
measuring long - distance induced decay itself might be helpful to understand the long - distance dynamics in the @xmath39 sector , such as inclusive @xmath40 decay or exclusive @xmath41 and @xmath42 decays related to extracting cabibbo - kobayashi - maskawa matrix elements @xmath43 .
on the other hand , observation of lnv ( @xmath44 ) decays could be an indication of a majorana nature of neutrinos @xcite .
we have used two sets of open - charm data samples collected by the detector in @xmath45 collisions provided by the cornell electron storage ring ( cesr ) .
the integrated luminosities are @xmath6 pb@xmath7 at the center - of - mass energy @xmath8 gev near the peak of the @xmath46 resonance which decays to @xmath47 pairs , and @xmath11 pb@xmath7 at @xmath12 gev near the peak of @xmath14 pair production .
the @xmath48 gev data set contains @xmath9 @xmath10 pairs and is used to study @xmath24 decays .
the @xmath49 gev data set contains @xmath13 @xmath14 pairs , and is used to study @xmath25 decays .
the remainder of this article is organized as follows .
the detector is described in sec .
[ sec : detector ] .
event selection criteria are described in sec .
[ sec : event_selectuon ] .
features of background processes , our suppression strategy , and signal sensitivity are discussed in sec .
[ sec : analysis ] .
results are presented as plots and tables in sec .
[ sec : results ] .
systematic uncertainties associated with the branching fractions and their upper limits are discussed in sec .
[ sec : systematic_uncertainty ] .
finally , a summary of our results with systematic uncertainties is provided in sec .
[ sec : summary ] .
the detector @xcite is a general - purpose solenoidal detector equipped with four concentric components : a six - layer vertex drift chamber , a 47-layer main drift chamber , a ring - imaging cherenkov ( rich ) detector , and a cesium iodide electromagnetic calorimeter , all operating inside a 1 tesla magnetic field provided by a superconducting solenoidal magnet .
the detector provides acceptance of @xmath50% of the full @xmath51 solid angle for both charged particles and photons .
the main drift chamber provides specific - ionization ( @xmath52 ) measurements that discriminate between charged pions and kaons .
the rich detector covers approximately @xmath53% of @xmath51 and provides additional separation of pions and kaons at momentum above @xmath54 mev .
hadron identification efficiencies are approximately @xmath55% with misidentification rates of a few percent @xcite .
electron identification is based on a likelihood variable that combines the information from the rich detector , @xmath52 , and the ratio of electromagnetic shower energy to track momentum ( @xmath56 ) .
typical electron identification efficiency is well over @xmath57% on average with the pion fake rate less than @xmath58% and the kaon fake rate less than a percent @xcite .
a geant - based @xcite monte carlo ( mc ) simulation is used to study efficiencies of signal and background events .
physics events are generated by evtgen @xcite , tuned with improved knowledge of charm decays , and final - state radiation ( fsr ) is modeled by photos @xcite .
nonresonant fcnc and lnv signal events are generated according to phase space .
signal candidates are formed from sets of well - measured drift chamber tracks consistent with coming from the nominal interaction point . charged pions and kaons are identified from the tracks with momentum greater than @xmath59 mev and with @xmath60 , where @xmath61 is the angle between the track and the beam axis .
electron candidates are required to be above @xmath62 mev with @xmath63 to ensure that @xmath56 is well measured . at @xmath8 gev , for each signal candidate of the form @xmath24 ( where @xmath3 is either @xmath4 or @xmath5 ) , two kinematic variables are computed to define a signal region : the energy difference @xmath64 and the beam - constrained mass difference @xmath65^{1/2 } - m_{d^+}$ ] , where @xmath66 is the four - momentum of the signal @xmath0 candidate , @xmath67 is the beam energy , and @xmath68 is the nominal @xcite mass of the @xmath0 meson . to improve the resolution of the kinematic variables , we recover bremsstrahlung photon showers within @xmath69 mrad of the direction of the electron candidates .
we define a signal box for further analysis as @xmath70 , which corresponds to about @xmath71-standard deviations of the kinematic variables .
because the expected contribution from the resonant decay @xmath72 is within our sensitivity , we further subdivide @xmath26 candidates into two channels : resonant @xmath73 and nonresonant @xmath26 for the fcnc search .
if the dielectron invariant mass @xmath74 of the signal candidate is within @xmath75 mev of the nominal @xcite mass of the @xmath36 meson , we treat it as a resonant @xmath73 candidate and exclude it from the @xmath26 candidates . similarly , at @xmath12 gev , for each signal candidate of the form @xmath25 , the following two variables are computed to define a signal region : the mass difference @xmath76 and the recoil mass ( against the signal candidate ) difference @xmath77^{1/2 } - m_{d^{\ast + }
_ s}$ ] , where @xmath78 is the invariant mass of the signal candidate , @xmath79 is the nominal @xcite mass of the @xmath80 , @xmath81 is the total four - momentum of the @xmath82 beam taking the finite beam crossing angle into account , @xmath83 is the four - momentum of the signal candidate with @xmath84^{1/2}$ ] , and @xmath85 is the nominal @xcite mass of the @xmath86 . the same bremsstrahlung recovery is performed and the @xmath87 channel is subdivided into resonant @xmath88 and nonresonant channels .
the signal box is defined as @xmath89 for further analysis .
the broad recoil mass window @xmath90 mev is required to allow both primary and secondary ( from @xmath91 or @xmath92 ) @xmath1 candidates to be selected .
backgrounds are dominantly from events with real electrons , particularly from @xmath93 semileptonic decays .
the majority of combinatorial background events are from double charm semileptonic decays , typically 4 or less charged particles in the event with large missing energy due to the missing neutrinos .
hadronic decays involving @xmath94-conversion and @xmath95 ( @xmath96 , @xmath35 ) dalitz decay , or accompanied by another charm semileptonic decay , can mimic the @xmath2 signal , as well .
because of the low probability of hadrons being misidentified as electrons @xcite , background from @xmath47 decays to 3-body charged - particle hadronic decays ( such as @xmath97 , @xmath98 , @xmath99 , @xmath100 ) are negligible after two electrons are identified , and they do not peak at the signal region due to the wrong mass assignments for the hadrons misidentified as electrons .
that is , @xmath47 backgrounds are predominantly associated with the semileptonic decays and non-@xmath47 ( @xmath101 continuum , @xmath102-pair , radiative return , or qed events ) backgrounds are associated with the @xmath94-conversion and dalitz decays .
all of these backgrounds are nonpeaking or peak away from the signal regions .
our background suppression criteria tuning procedure for @xmath24 channels is detailed in our previous article @xcite .
we have used the same background rejection criteria with the four kinematic variables to reject the above - mentioned backgrounds in @xmath0 channels and revised the criteria to accommodate the @xmath1 channels .
the other side total energy @xmath103 is the sum of energies of all particles other than those making up the signal candidate .
we use this variable to reject events associated with semileptonic decays , mainly for double charm semileptonic decays , in which the visible other side energy would be small due to the undetectable missing neutrinos .
we reject candidates if @xmath104 gev for @xmath26 , @xmath105 gev for @xmath106 , @xmath107 gev for @xmath87 , and @xmath108 gev for @xmath27 .
for the lnv modes , we reject candidates if the number of tracks in the event is 4 or fewer and @xmath109 gev .
semileptonic events involving @xmath110 in the final state can mimic the signal in @xmath111 channels .
we have used the invariant mass @xmath112 to veto these events .
we veto the candidate when the charged pion in the signal candidate combined with any other unused oppositely charged track satisfies @xmath113 mev , where @xmath114 is the nominal @xcite mass of the @xmath115 .
real electrons from @xmath94-conversion and dalitz decays are suppressed by using the dielectron invariant mass squared @xmath116 computed from the signal electron positron pair , or @xmath117 computed using one signal side electron ( positron ) combined with any oppositely charged unused track .
we veto candidates if @xmath118 or @xmath119 . for @xmath1 ,
we have required the solo photon from @xmath86 decays to @xmath120 to be explicitly reconstructed to further suppress underlying nonstrange - charmed meson backgrounds at @xmath12 gev , by requiring the recoil mass of the signal candidate plus solo photon @xmath121 to be within @xmath122 mev of the nominal @xcite @xmath80 mass .
regardless of whether the signal @xmath1 candidate is the primary or secondary @xmath1 , for the decay @xmath123 , the mass of the system recoiling against the @xmath1 plus @xmath94 should peak at the @xmath1 mass .
the analysis was done in a blind fashion .
before we opened the signal box , all above - mentioned criteria were optimized using mc events with a sensitivity variable which is defined as the average upper limit one would get from an ensemble of experiments with the expected background and no signal , @xmath124 where @xmath125 is the expected number of background events , @xmath126 is the observed number of events , @xmath127 is the 90% confidence coefficient upper limit on the signal , @xmath128 is the poisson probability , @xmath129 is the number of @xmath0 or @xmath1 , and @xmath130 is the signal efficiency .
in addition to the signal mc samples , four types of background mc samples are utilized to optimize the background suppression criteria : @xmath131 times the data sample for open - charm ( @xmath47 , @xmath132 , @xmath133 , @xmath134 @xmath135 , and @xmath136 ) , @xmath137 times the data sample of noncharm @xmath138 continuum ( @xmath101 ) , @xmath102-pair , and radiative return to the @xmath139 . to normalize background mc events to match the expected number of the data events , we have used integrated luminosity and cross sections for each process . for @xmath24 events at @xmath140 mev
, we have used @xmath141 nb @xcite , @xmath142 nb @xcite , @xmath143 nb @xcite , @xmath144 nb , where @xmath145 is the @xmath102 velocity . ] , and radiative return to the @xmath139 @xmath146 nb @xcite . for @xmath25 events at @xmath147 mev
, we have used @xmath148 nb @xcite ( and used other open - charm cross sections from the same reference ) , @xmath149 nb @xcite , @xmath150 nb , and radiative return to the @xmath139 @xmath151 nb @xcite .
we have found that the agreements between data and mc simulated events are excellent in various kinematic variables used in the background suppression , giving us confidence in our optimization procedure using our mc samples .
possible systematic uncertainties due to the data and mc differences are assessed in sec .
[ sec : systematic_uncertainty ] .
scatterplots of @xmath152 vs @xmath153 and @xmath154 vs @xmath155 for signal candidates with all background suppressions applied are shown in figs .
[ fig:2d - de - dmbc ] and [ fig:2d - dm - dmrecoil ] . except for the @xmath88 channels
, we find no evidence of signals , and we calculate @xmath57% confidence level upper limits ( ul ) on the branching fractions based on _ poisson processes with background _ @xcite ( e.g. section 28.6.4 _ poisson processes with background _ therein ) as summarized in table [ table : summary - final-2 ] : @xmath156 for @xmath0 and @xmath1 @xmath157 channels , we find weak evidence of signals with significance @xmath158 for the @xmath0 and @xmath159 for the @xmath1 , so both branching fractions and upper limits are shown in table [ table : summary - final-2 ] . [ cols="<,^,^,^,^,^,>",options="header " , ] uncertainty in the number of @xmath0 ( @xmath1 ) is estimated by adding contributions from uncertainties in integrated luminosity @xcite @xmath160% and the production cross section @xcite @xmath161% ( @xmath162% for @xmath1 @xcite ) in quadrature .
we assign relative uncertainties of @xmath163% to the number of @xmath0 and of @xmath164% to the number of @xmath1 .
there are several sources which can contribute to uncertainty in the signal efficiency estimation , as listed in table [ table : systsummary ] . by adding contributions from tracking @xcite , particle identification ( pid ) @xcite , fsr @xcite , background suppression , and mc statistics in quadrature we found total uncertainties in the signal efficiency for each channel range from @xmath71% to @xmath165% .
we use the number of background events estimated by the mc simulation rather than using the sidebands in data . the mc samples , being @xmath137-@xmath131 times larger , have higher precision .
we have evaluated possible systematic bias caused by the use of mc events rather than the data sideband by using alternative background shapes , and by comparing the mc predicted number to that interpolated from the data sideband .
we found no indication of systematic bias ; all deviations are adequately explained as statistical fluctuations due to the data statistics .
we conclude that our mc events reproduce the features of the data backgrounds well .
we took the statistical uncertainty in the mc simulated number of backgrounds as the systematic uncertainty in the expected number of background , as summarized in table [ table : systsummary ] .
with the complete samples of open - charm data , corresponding to integrated luminosities of @xmath6 pb@xmath7 at @xmath8 gev containing @xmath9 @xmath10 pairs and @xmath11 pb@xmath7 at @xmath12 gev containing @xmath13 @xmath14 pairs , we have searched for rare ( fcnc ) and forbidden ( lnv ) decays of @xmath0 and @xmath1 mesons of the form @xmath2 , where @xmath166 is either a charged pion or a charged kaon .
we found no evidence of signals and set upper limits on branching fractions at the @xmath57% confidence level as summarized in table [ table : summary - final-2 ] .
systematic uncertainties in the signal efficiency , the number of @xmath0 ( or @xmath1 ) events , and the expected number of background events are incorporated by decreasing the numbers used for those quantities by 1 standard deviation of the systematic uncertainty on those quantities .
these results are the most stringent limits on fcnc and lnv for the @xmath0 and @xmath1 @xmath167 decays to date and the limits in the dielectron channels are comparable to those in the dimuon channels @xcite , but are still a few orders of magnitude larger than the sm expectation @xcite in fcnc decays .
this leaves some room for possible enhancement @xcite in both fcnc and lnv decays induced by non - sm physics .
we have separately measured branching fractions of the resonant decays @xmath168 and @xmath169 due to their large expected contributions to @xmath111 channels .
the significance of our measured branching fractions is poor at @xmath158 standard deviations for @xmath0 and @xmath159 standard deviations for @xmath1 , so we have also included upper limits in table [ table : summary - final-2 ]
. our measured branching fractions of these decays are consistent with the products of known world average @xcite branching fractions , @xmath170 \times [ ( 2.97 \pm 0.04 ) \times 10^{-4 } ] = ( 1.8 \pm 0.2 ) \times 10^{-6}$ ] and @xmath171 \times [ ( 2.97 \pm 0.04 ) \times 10^{-4 } ] = ( 1.3 \pm 0.1 ) \times 10^{-5}$ ] .
we gratefully acknowledge the effort of the cesr staff in providing us with excellent luminosity and running conditions .
d. cronin - hennessy thanks the a.p . sloan foundation .
this work was supported by the national science foundation , the u.s .
department of energy , the natural sciences and engineering research council of canada , and the u.k .
science and technology facilities council .
q. he _ et al .
_ ( cleo collaboration ) , phys . rev .
lett . * 95 * , 221802 ( 2005 ) .
g. burdman , e. golowich , j. hewett and s. pakvasa , phys .
d * 66 * , 014009 ( 2002 ) .
s. fajfer , s. prelovsek and p. singer , phys .
d * 64 * , 114009 ( 2001 ) .
s. fajfer and s. prelovsek , phys .
d * 73 * , 054026 ( 2006 ) .
s. fajfer , n. kosnik and s. prelovsek , phys .
d * 76 * , 074010 ( 2007 ) .
a. ali , a. v. borisov and n. b. zamorin , eur .
j. c * 21 * , 123 ( 2001 ) .
a. atre , t. han , s. pascoli and b. zhang , jhep * 0905 * , 030 ( 2009 ) .
r. a. briere _ et al .
_ ( cesr - c and cleo - c taskforces , cleo - c collaboration ) , cornell university , lepp report no
. clns 01/1742 ( 2001 ) ( unpublished ) .
y. kubota _ et al .
_ ( cleo collaboration ) , nucl . instrum .
methods phys .
a * 320 * , 66 ( 1992 ) .
d. peterson _
et al . _ ,
instrum .
methods phys .
res . , sect .
a * 478 * , 142 ( 2002 ) .
m. artuso _ et al .
_ , nucl .
instrum .
methods phys .
res . , sect .
a * 502 * , 91 ( 2003 ) .
s. dobbs _ et al .
_ ( cleo collaboration ) , phys . rev .
d * 76 * , 112001 ( 2007 ) .
d. m. asner _
( cleo collaboration ) , phys .
d * 81 * , 052007 ( 2010 ) .
d. besson _ et al .
_ ( cleo collaboration ) , phys . rev .
d * 80 * , 032005 ( 2009 ) .
r. brun _ et al .
_ , geant 3.21 , cern program library long writeup w5013 ( unpublished ) 1993 .
d. j. lange , nucl .
instrum .
methods phys .
res . , sect .
a * 462 * , 152 ( 2001 ) .
e. barberio and z. was , comput .
. commun . * 79 * , 291 ( 1994 ) . c. amsler
_ et al . _
( particle data group ) , phys .
b * 667 * , 1 ( 2008 ) .
j. z. bai _ et al . _
( bes collaboration ) , phys .
lett . * 88 * , 101802 ( 2002 ) .
m. benayoun , s. i. eidelman , v. n. ivanchenko and z. k. silagadze , mod .
a * 14 * , 2605 ( 1999 ) .
d. cronin - hennessy _ et al .
_ ( cleo collaboration ) , phys .
d * 80 * , 072001 ( 2009 ) .
r. m. barnett _ et al .
_ ( particle data group ) , phys .
d * 54 * , 1 ( 1996 ) . | we have searched for flavor - changing neutral current decays and lepton - number - violating decays of @xmath0 and @xmath1 mesons to final states of the form @xmath2 , where @xmath3 is either @xmath4 or @xmath5 .
we use the complete samples of open - charm data , corresponding to integrated luminosities of @xmath6 pb@xmath7 at the center - of - mass energy @xmath8 gev containing @xmath9 @xmath10 pairs and @xmath11 pb@xmath7 at @xmath12 gev containing @xmath13 @xmath14 pairs .
no signal is observed in any channel , and we obtain @xmath15 confidence level upper limits on branching fractions @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , and @xmath23 . |
image states @xcite are a special class of weakly bound surface electronic states in which an electron outside a dielectric or conductor polarises the surface and is then attracted to the resulting `` image charge '' .
asymptotically the potential varies like @xmath6 so that a band gap preventing penetration of electrons into the crystal leads to a rydberg - like series of states which in the case of a purely coulombic image force at a planar metal surface arise at energies @xmath7/n^2+\hbar^2k^2/2m$ ] , @xmath8 where @xmath9 is the electron wave vector parallel to the surface and @xmath10 the electron mass .
deviations from this behaviour reflect the influence and response of the surface - dependant electronic and atomic structure , which may therefore be investigated by studying image states .
examples of theoretical and experimental work include the systematics of image states binding and dispersion on clean surfaces @xcite , image state on overlayers @xcite , at stepped metal surfaces @xcite , exchange splitting of image states at ferromagnets @xcite , as well as image states at surface nanostructures @xcite .
in recent years there has also been considerable interest in the dynamics of image state electron @xcite as model electronic excitations at surfaces .
one aspect of the physics of image state electrons that has yet to be addressed is the influence of the spin orbit interaction @xmath11 @xcite which has recently been found to have a significant effect on other surface state electron levels at the surfaces of conductors with high atomic number @xcite . at first sight
the spin - orbit interaction might be expected to be negligible .
the mathematical analogy that can be drawn between the schrdinger equation describing the electrons moving in the coulomb - like image potential and that of @xmath12-electrons in the hydrogen atom enable the image state wavefunctions to be written as @xmath13 where @xmath14 is the normalised radial hydrogenic wavefunction and @xmath15 a pauli spinor : @xmath16 . using these wavefunctions to diagonalise
the spin - orbit perturbation @xmath17 in the subspace of degenerate image state levels gives a spin - orbit splitting of @xmath18\frac{k}{n^3}\ ] ] where @xmath19 is the fine structure constant .
this is well below the current resolution of inverse photoemission , two - photon photoemission or scanning tunnelling spectroscopy .
however , it has previously been recognised that a more significant contribution to the spin - orbit splitting of `` crystal derived '' surface states arises from the brief time spent by the electron in the vicinity of the nuclei of the surface atoms , where the gradient contribution to the spin - orbit interaction is @xmath20 . in this paper
we report on calculations that we have performed to quantify the magnitude of the spin - orbit splitting that arises from the penetration of the image state wave function into the crystal at surfaces of ir , pt and au .
these are described in section [ section : rlkkr ] . in section [ section :
phase ] we describe the modification of the phase accumulation model for image state energetics to include the effects of the spin orbit interaction .
finally , we summarise and discuss our findings .
to calculate the spin - orbit splitting of image states we use a recently developed code that implements relativistic multiple - scattering theory .
the theory behind this method is essentially that described by halilov _
@xcite so we do not reproduce it in detail here .
the basic idea is that the electronic structure is found from the single - particle green function corresponding to the dirac hamiltonian @xmath21 @xcite .
thus spin - orbit effects are treated non - perturbatively . using scattering techniques the green function
is determined for the special case of semi - infinite crystals with two - dimensional in - plane translational periodicity , treating intralayer scattering within an angular momentum representation and interlayer scattering in a plane wave representation .
our calculations use 25 and 19 two - dimensional reciprocal lattice vectors to describe the interlayer scattering for the ( 001 ) and ( 111 ) surfaces respectively , and partial waves up to @xmath22 @xcite .
the semi - infinite substrate means that continuum and surface - localised states are clearly distinguished in the wave vector resolved local density of states , found from the imaginary part of the green function . as in the non - relativistic version of the code @xcite the electronic structure
is found self - consistently using the local density approximation to density functional theory .
we use the atomic sphere approximation for the crystal potential ( including dipole contributions ) , with the potential in the three outermost atomic layers allowed to vary in response to the presence of the surface . since the local density approximation does not lead to an image - like surface barrier , and hence does not support image states , once self - consistency has been achieved we replace the self - consistent barrier with a parameterised model barrier , for which we use the `` jjj '' potential @xcite @xmath23 the fitting parameters @xmath24 , @xmath25 and @xmath26 were fixed by starting with values quoted by smith _ et al . _
@xcite , who fitted to first - principles slab calculations , and then adjusted slightly to place the @xmath2 image state at @xmath27 , @xmath28 , close to values found experimentally .
the procedure does not uniquely fix the parameters , but we found that different combinations that gave the same value for @xmath28 resulted in almost identical image state dispersion curves .
note that our results are for the @xmath29 unreconstructed surfaces of the materials studied .
in several cases the surfaces undergo complex surface reconstructions ( e.g. au(001 ) and pt(001 ) adopt a @xmath30 reconstruction but may be prepared in the @xmath29 structure see @xcite ) .
= 160 mm in figure [ fig : splits ] we illustrate the dispersion curves that we find . for both pt(111 ) and au(001 ) the spin - orbit interaction can clearly be seen to split the @xmath2 image state , and whilst a splitting exists for the @xmath31 and higher states it is much smaller .
the inset in the figure [ fig : splits]a illustrates the variation of the splitting with wave vector @xmath32 , the near - linear variation corresponding to a rashba - like dispersion @xcite @xmath33 from curves such as these we extract rashba - parameters @xmath1 by a least - square fit using wave vectors @xmath34 .
these values are tabulated in table [ table : results ] for the image states at the ( 111 ) and ( 001 ) surfaces of ir , pt and au . at au(111 )
the vacuum level lies outside of the projected band gap so that the image states in this case exist as resonances .
we have not therefore included results for this case , but instead give the results that we find for the spin orbit splitting of the occupied surface state that occurs within the band gap at this surface .
the dispersion of this state is shown also in figure [ fig : splits ] , and agrees well with previous work , validating our procedure .
we find the wave vector splitting at the fermi energy is @xmath35 , compared with experimental values of @xmath36 @xcite , @xmath37 @xcite and @xmath38 @xcite , and a theoretical value of @xmath36 @xcite that have been reported previously .
.[table : results]calculated image state energies including the spin - orbit interaction at @xmath39 and @xmath40 surfaces of ir , pt and au : @xmath41 .
@xmath42 values in brackets are from experiment . [ cols="^,^,^,^,^,^,^ " , ] the results in table [ table : results ] indicate that the spin - orbit splitting of @xmath2 image states at ir , pt and au surfaces is an order of magnitude smaller than that of the au(111 ) shockley surface state , and that of the @xmath31 states smaller still .
these trends reflect the differing extents to which the corresponding wave functions penetrate the crystal and experience the spin - orbit interaction at the ion cores . at the ( 001 ) surfaces @xmath43 which might be expected given the increasing atomic number ( @xmath44 ) , but at the ( 111 ) surface @xmath45 , pointing to a more complicated `` band - structure '' effect . in figure
[ fig : splits ] the splitting of the au(001 ) @xmath2 image state is also seen to be affected as it disperses towards the band edge of continuum levels .
we return to this later .
the `` standard model '' for understanding image state energies is the phase accumulation model @xcite in which surface states are envisaged as one - dimensional waves trapped by multiple reflection from the surface barrier and the crystal . in a region of constant potential ( see figure [ fig : phase ] ) between barrier and crystal ( which may be infinitesimal in width ) the electron wave function can be expressed in terms of forward and backward travelling plane waves @xmath46 the two components are related at the barrier reference plane @xmath47 by the barrier reflection coefficient @xmath48 , @xmath49 and at the crystal reference plane @xmath50 by the crystal reflection coefficient @xmath51 , @xmath52 which together with ( [ eqn : wave ] ) give rise to the condition for a surface state to exist : @xmath53 for energies below the vacuum level and coincident with the crystal band gap the reflection probability at both crystal and barrier are unity and the reflection coefficients may be written in terms of phases : @xmath54 , @xmath55 .
the surface state condition then becomes @xmath56 which is a condition on the round - trip phase accumulated by the electron wave .
the phases in ( [ eqn : phase ] ) increase with energy .
the crystal phase increases from 0 to @xmath57 as the energy sweeps across the band gap , changing most rapidly near the band edges . towards the bottom of the gap it is
this variation in @xmath58 which will determine whether or not ( [ eqn : phase ] ) is satisfied , so that any surface state that does arise is usually referred to as `` crystal - derived '' . on the other hand
, @xmath59 increases more and more rapidly as the energy approaches the vacuum energy , varying to a good approximation as @xmath60 in combination with ( [ eqn : phase ] ) this yields a rydberg - like series of image states @xmath61 where the quantum defect @xmath62/\pi)/2 $ ] may usually be considered constant over the range of energies at which the image states are found .
= 80 mm we now consider the extension of this model to include the spin orbit interaction .
introducing the electron spin in to the wavefunction in ( [ eqn : wave ] ) @xmath63 and reflection from the crystal is now described by a matrix @xmath64 allowing for the possibility of spin - flip upon reflection . with a similar matrix used to describe scattering from the barrier ,
the condition for a surface state becomes @xmath65=0 .
\label{eqn : det}\ ] ] the four reflection coefficients in ( [ eqn : rc ] ) are not independent for example flux conservation requires that @xmath66 within a gap , and for a non - magnetic crystal @xmath67 . in the non - magnetic case and for a planar potential @xmath68 the electron wave functions @xmath69 are found from the hamiltonian @xmath70 which may be diagonalised by a rotation in spin space @xmath71 with @xmath72 where @xmath73 is the angle of the electron wave vector .
the hamiltonian @xmath74 does not mix spin - up and spin - down channels and so in this representation the reflection matrix describing scattering from the crystal is diagonal : @xmath75 since the spin - orbit interaction is negligible in the barrier the barrier reflection matrix is also diagonal , and for a non - magnetic surface @xmath76 so that the surface state condition ( [ eqn : det ] ) becomes @xmath77 leading to the round - trip phase condition @xmath78 where we have introduced @xmath79 appropriate to energies within a gap .
the surface states now come in spin - split pairs , with spin orientations that may be deduced from the spinors that are obtained by rotating back in to the original reference frame the spin - up and -down eigenspinors @xmath80 of the primed frame : @xmath81 thus @xmath82 and we find that the spins lie in the surface plane and perpendicular to @xmath83 , to the left ( right ) for @xmath84 ( @xmath85 ) . rotating back to the original spin frame gives the reflection matrix ( [ eqn : rc ] ) as @xmath86 in figure [ fig : tan ] we illustrate the variation in @xmath87 at the au(001 ) surface calculated from the reflection matrix found using the relativistic multiple - scattering method of section [ section : rlkkr ] . for @xmath9 along @xmath88 ( see inset in figure [ fig : tan ] ) equation ( [ eqn : rc ] ) @xmath89 .
the relativistic multiple - scattering calculations include the atomic structure of the surface and the crystal potential is not one - dimensional , but the reflection coefficient for specular reflection behaves in a very similar manner to that of a one - dimensional crystal , especially for small @xmath32 where the wave function varies only slowly across the surface .
in particular we find that the angular variation predicted by equation ( [ eqn : rc ] ) is satisfied to within a percent or so .
it is evident from figure [ fig : tan ] that the magnitude of the spin - orbit induced phase change is small , and away from the band edges @xmath90 is approximately independent of energy at au(001 ) and linear in @xmath32 : @xmath91 k$ ] . to first order
the round - trip phase condition ( [ eqn : phase2 ] ) is satisfied at energies @xmath92 where @xmath93 neglecting the energy dependence of @xmath90 which is small compared to the other phases . for all the surfaces that we have studies we have found that @xmath90 does not change sign across the band gap , and the denominator in ( [ eqn : energies ] ) is positive .
hence the model predicts a series of spin - orbit split states with identical spin orderings , which we have confirmed is also the case in the relativistic multiple scattering calculations described in section [ section : rlkkr ] . in particular with @xmath94 the surface states that exist are split with the lower of each pair of levels having the spin pointing to the right of @xmath9 , as shown in figure [ fig : spin ] .
this is in agreement with the spin assignments shown in henk _ et al . _
@xcite for the au(111 ) surface state , but disagrees with those given in ref .
@xcite .
= 50 mm when the gap contains image states , over the narrow range of energies within which the image states are found the energy - dependence of the round - trip phase is dominated by the variation in the barrier phase ( [ eqn : barphase ] ) and then @xmath95 thus the spin - splittings exhibit the same scaling as the lifetime broadening @xcite , in each case the behaviour ultimately originating in the variation with @xmath96 of the wave function overlap with the crystal .
we also see from ( [ eqn : de ] ) that the linear - in-@xmath32 behaviour of @xmath97 arises from similar behaviour in @xmath90 . since @xmath90 is small @xmath98 to a good approximation , and hence @xmath99 .
thus the spin - orbit splitting is directly related to the spin - flip scattering rate , which could therefore be determined from experimental values of image state splittings . at au(001 ) , @xmath100 mev ( table [ table : results ] ) , so combining equations ( [ eqn : ryd ] ) and ( [ eqn : de ] ) gives @xmath101 and a splitting of 22 mev ( figure [ fig : splits ] ) at @xmath102 @xmath103 yields @xmath104 , which agrees with the value found from the multiple - scattering calculations shown in figure [ fig : tan ] .
finally , we note that equation ( [ eqn : de ] ) will not hold near band edges where the energy - dependence of the crystal phase @xmath58 can not be neglected .
this is the origin of the anomalous dispersion shown for the @xmath2 state at au(001 ) in figure [ fig : splits ] .
to summarise , we have investigated the effect of the spin - orbit interaction on image state electrons at the ( 111 ) and ( 001 ) surfaces of ir , pt , and au .
non - perturbative calculations that use relativistic multiple - scattering theory with self - consistent potentials and a parameterised surface barrier predict rashba - like dispersion of the image state bands with splittings for @xmath2 that are a factor 10 - 20 times smaller than that of the au(111 ) shockley state , for which our results are in good agreement with experiment and previous theory . extending the phase accumulation model to include spin - orbit scattering
, we find that the splittings scale as @xmath4 , where @xmath5 is the quantum defect , and are directly related to the spin - flip scattering rate at the surface .
the largest image state splittings we find are at au(001 ) for which @xmath105 mev at @xmath102 @xmath103 .
this is larger than some exchange splittings of image states that have previously been resolved ( e.g. @xmath106 mev at ni(111 ) @xcite and @xmath107 mev at ni(001 ) @xcite ) by exploiting the spin resolution available in spin - resolved inverse photoemission , suggesting that the spin - orbit splitting may also be observable with an appropriate experimental set - up .
the splittings are comparable to lifetime broadenings of late - transition and noble metal image states @xcite , indicating that account of spin - orbit effects may be necessary when determining image state lifetimes from lineshape analysis for 5@xmath108 metals .
99 echenique p m and pendry j b 1978 j. phys .
c * 11 * 2065 straub d and himpsel f j 1986 phys . rev .
b * 33 * 2256 padowitz d f , merry w r , jordan r e and harris c b 1992 phys . rev
. lett .
* 3583 ischer r , schuppler s , fisher n , fauster th and steinmann w 1993 phys . rev .
lett . * 70 * 654 wallauer w and fauster th 1996 phys .
b * 54 * 5096 smadici s , mocuta d and osgood r m 2004 phys .
b * 69 * 035415 passak f and donath m 1992 phys .
lett . * 69 * 1101 nekovee m , crampin s and inglesfield j e 1993 phys .
lett . * 70 * 3099 de rossi s , ciccacci f and crampin s 1996 phys .
lett . * 77 * 908 fischer r , fauster th , and steinmann w 1993 phys .
b * 48 * 15496 ortega j e , himpsel j f , haight r and peale d r 1994 phys . rev .
b * 49 * 13859 hill i g and mclean a b 1999 phys .
lett . * 82 * 2155 kasperovich v , wong k , tikhonov g , and kresin v v 2000 phys .
85 * 2729 echenique p m , pitarke j m , chulkov e v and rubio a 2000 chem . phys . * 251 * 1 wahl p , schneider m a , diekhner l , vogelgesang r and kern k 2003 phys . rev .
lett . * 91 * 106802 rhie h -s , link s , drr h a , eberhardt w and smith n v 2003 phys . rev .
b * 68 * 033410 boger k , weinelt m and fauster th 2004 phys .
* 92 * 126803 messiah a 1962 _ quantum mechanics _
vol 2 ( amsterdam : north holland ) lashell s , mcdougall b a and jensen e 1996 phys .
rev.lett . * 77 * 3419 rotenberg e , chung j w and kevan s d 1999 phys .
lett . * 82 * 4066 halilov s v , tamura e , meinert d , gollisch h and feder r 1993 j. phys .
: condens .
matter * 5 * 3859 crampin s 1994 phys .
b * 49 * 14035 crampin s 1993 j. phys . : condens .
matter * 5 * 4647 jones r o , jennings p j , and jepsen o 1984 phys .
b * 29 * 6474 smith n v , chen c t and weinert m 1989 phys .
b * 40 * 7565 drube r , dose v and goldmann a 1988 surf .
* 197 * 317 rashba e i 1960 sov .
phys . solid state * 2 * 1109 bychkov y a and rashba e i 1984 jetp lett . *
39 * 78 link s , drr h a , bihlmayer g , blgel s , eberhardt w , chulkov e v , silkin v m , and echenique p m 2001 phys .
b * 63 * , 115420 kinoshita i , anazawa t and matsumoto y 1996 chem .
lett . * 259 * 445 nicolay g , reinert f , hfner s and blaha p 2001 phys . rev .
b * 65 * 033407 reinert f 2003 j. phys .
: condens .
matter * 15 * s693 kliewer j , berndt r , chulkov e v , silkin v m , echenique p m and crampin s 2000 science * 288 * 1399 ciccacci f , de rossi s , taglia a and crampin s 1994 j. phys .
: condens .
matter * 6 * 7227 straub a and himpsel f j 1984 phys .
lett . * 52 * , 1922 mugarza a _ et al .
_ 2002 phys .
b * 66 * 245419 henk j , ernst a and bruno p 2003 phys .
b * 68 * 165416 starke k , ertl k , and dose v 1992 phys . rev .
b * 45 * 6154 | we quantify the effect of the spin - orbit interaction on the rydberg - like series of image state electrons at the ( 111 ) and ( 001 ) surface of ir , pt and au . using relativistic multiple - scattering methods we find rashba - like dispersions with @xmath0 with values of @xmath1 for @xmath2 states in the range @xmath3 . extending the phase - accumulation model to include spin - orbit scattering we find that the splittings vary like @xmath4 where @xmath5 is the quantum defect and that they are related to the probability of spin - flip scattering at the surface .
the splittings should be observable experimentally being larger in magnitude than some exchange - splittings that have been resolved by inverse photoemission , and are comparable to linewidths from inelastic lifetimes . |
motivated by the success of cold atomic physics , recently , there has been a growing interest in producing cold and ultracold molecules in molecular physics @xcite .
cold molecules could be produced by sympathetic cooling @xcite or deceleration of supersonic beams @xcite .
these methods are general ; however , it is still hard to attain ultracold molecules(@xmath11mk ) . despite the difficulty of directly cooling molecules , numerous experimental efforts have been made to generate ultracold alkali molecules through associating ultracold atoms implying the feshbach resonance @xcite or photoassociation @xcite techniques . in general , the ultracold molecules created by these two approaches are distributed in various excited states . over the last few years ,
a lot of theoretical and experimental investigations have demonstrated , via the stimulated raman adiabatic passage ( stirap ) technique @xcite , that excited ultracold molecules could be transferred to molecular ground states @xcite-@xcite .
the stirap technique applies two specially designed sequential laser pulses to produce a coherent population trapping ( cpt ) state for the purpose of realizing a complete population transfer between two quantum states of an atom or a molecule @xcite . in stirap experiments , when two pulsed lasers interact with an atom or a molecule , the corresponding hamiltonian for the atom or molecule is time dependent .
theoretical and experimental investigations for the stirap have shown that it is critical to keep the atomic or molecular system adiabatically staying in the instantaneous cpt eigenstate of the time - dependent hamiltonian @xcite-@xcite .
therefore , any dynamical deviations from the cpt state must be small enough to achieve a good stirap .
in photoassociation of ultracold molecules combined with stirap , the adiabaticity analysis of the cpt state gets more complicated , because the physical process of atom - molecule transfer is intrinsically nonlinear .
two nonlinear resources exist .
one is due to merging of two atoms , and the other arises from collisions between atoms and molecules .
the nonlinearity could make the system unstable , such that a small deviation from the cpt state might be amplified , and consequently the transfer efficiency from atoms to molecules will be decreased @xcite .
so adiabaticity and stability are crucial for achieving an efficient molecular generation .
ignoring collisional interactions between atoms and molecules , adiabatic conditions for the cpt state in photoassociation process have been intensively investigated @xcite-@xcite ; in particular , reference @xcite presents a general approach to treat the adiabatic condition through a linearization procedure . however , these investigations neglect the effect of nonlinear instability caused by collisional interactions between atoms and molecules @xcite . quite recently , references.@xcite extend the approach in @xcite including dynamical instabilities , and obtain improved adiabatic conditions .
in this paper we study the dynamics of applying stirap to photoassociate molecules from an atomic condensate .
we emphasize the relations between dynamical instability aroused from intra - atomic collision , the adiabatic condition of a cpt state and the transfer efficiency from atoms to molecules .
we obtain an analytical formula for a cpt state including the atom - atom collisional effect and obtain the collision - induced frequency shift between two quantum states .
further , even in dynamically unstable regimes , we find that the adiabatic realization of the cpt state is available , and a very efficient atom - molecule transfer could be achieved when the single - photon detuning can compensate the collision - induced frequency shift .
our paper is organized as follows . in section 2
, we present a two - color @xmath2-type model in the mean - field treatment and study the stability and adiabaticity properties for the corresponding cpt state . in particular , we derive an analytical form for the adiabatic condition in the case of effective single - photon resonance . in section 3 , we numerically test our theoretical results and discuss the experimental feasibility .
finally , a summary is presented in section 4 .
model.eps our scheme of photoassociating molecules applies a two - color @xmath2-type configuration @xcite as shown in figure [ model ] .
first , a pair of ultracold atoms initially occupied in the ground state @xmath3 is associated with and excited to a molecular state @xmath4 via a pump laser with a one - photon detuning @xmath5 from the state @xmath4 .
subsequently , another control laser with a detuning @xmath5+@xmath6 ( @xmath6 , the two - photon detuning ) from the molecular excited state @xmath4 takes the excited molecules to the molecular ground state @xmath7 .
the rabi frequencies by the two lasers , respectively denoted as @xmath8 and @xmath9 , are treated as real numbers in our following discussions . in the interaction picture ,
the hamiltonian for the atom - molecule system is given by @xcite : @xmath10 where @xmath11 are the annihilation(creation ) field operators for the state @xmath12 @xmath13 .
in equation ( [ hamn ] ) , @xmath14 with @xmath15-wave scattering length @xmath16 is the pseudo - potential for the atom - atom interaction .
atom - molecule and molecule - molecule interactions are ignored for the purpose of extracting physics induced by two - body @xmath15-wave collisions .
actually , our later numerical simulations show that by adding the neglected collisional effects , our main results will not change . in our analysis , we are concerned with a spatially uniform system at zero - temperature , and
thus the kinetic and trapping potential terms within equation ( [ hamn ] ) have been safely removed . using the mean - field method that replace @xmath17 with @xmath18 , where @xmath19 is the density of the total particle number , from equation ( [ hamn ] )
, we obtain a set of coupled heisenberg motional equations for @xmath20 , given by @xmath21 where @xmath22 is introduced phenomenologically for describing the spontaneous emission of the excited state @xmath4 , @xmath23 and @xmath24 . without an intra - atomic collisional term in equation ( [ dy1 ] ) ,
a cpt state has already been obtained @xcite .
now , including the collision , how the cpt state is modified is worthy of investigating @xcite . to obtain a cpt state including intra - atomic collisional effect , we assume @xmath25 , @xmath26 , @xmath27 , where @xmath28 is the chemical potential and @xmath29 are for population distributions in the cpt state .
we also neglect the spontaneous loss @xmath22 and treat all the other parameters as time - independent . inserting the ansatz for the cpt state into equations ( [ dy1])-([dy3 ] ) and considering the total particle number normalization , @xmath30 , we obtain the population distributions given by a new vector @xmath31^{t}$ ] , where @xmath32 under the condition of an effective two - photon resonance @xmath33 with @xmath34 . here , @xmath35 is the collision - induced frequency shift .
the chemical potential corresponding to equation ( [ dark_popu ] ) is @xmath36 .
equation([dark_popu ] ) shows that all the atoms @xmath37 could be converted into ground - state molecules @xmath38 under the two - photon resonance when @xmath39 approaches to @xmath40 .
more interestingly , equation ( [ dark_popu ] ) indicates that when @xmath41 is satisfied , the cpt state does not rely on the @xmath42 coefficient .
in contrast , previous works on other models for photoassociating molecules shows that the atom - atom collision could easily lead to the deviation of the system from a cpt state , and thus the final produce of molecules in the ground state could be low @xcite . in section [ cpt distributions ] , we have derived a cpt state with atom - atom collision .
however , this collision could lead to dynamical instability of the atom - molecule coupled system .
in this situation , whether the cpt state could be adiabatically followed is worthy of further exploring . in this section ,
we start to investigate an adiabatic theory for the cpt state .
first , we add small perturbations @xmath43 to the cpt state distributions @xmath44
, i.e. , @xmath45 and study the time evolution of the perturbations . inserting equations ( [ linearized1])-([linearized3 ] ) into equations ( [ dy1])-([dy3 ] ) and taking into account the time dependence of the rabi frequencies @xmath46 whose time dependence is ignored in section [ cpt distributions ] , we obtain a set of linearized coupled equations for the vector @xmath47 ^{t}$ ] with its conjugate component @xmath48:@xmath49 in equation ( [ psi_vec_eq ] ) , the coefficient matrices @xmath50 and @xmath51 are given by @xmath52{ccc}2u_{aa}\phi_{a}^{2 } & -\omega_{p}\phi_{a } & 0\\ -\omega_{p}\phi_{a } & -\delta_{u } & -\frac{\omega_{s}}{2}\\ 0 & -\frac{\omega_{s}}{2 } & 0 \end{array } \right ] , \mathbf{b}=\left [ \begin{array } [ c]{ccc}2u_{aa}\phi_{a}^{2 } & 0 & 0\\ 0 & 0 & 0\\ 0 & 0 & 0 \end{array } \right]\ ] ] with @xmath53 , @xmath54^{t}$ ] is a driving source term , and @xmath55 is a 3@xmath563 matrix associated with the loss of the excited state denoted by the only nonzero element @xmath57 . for the convenience of the following analysis , we introduce a vector @xmath58^{t}$ ] , and rewrite the motional equation ( [ psi_vec_eq ] ) as,@xmath59 where @xmath60 , \mathbf{d}=\left [ \begin{array } { cc } \mathbf{f } & 0\\ 0 & \mathbf{f}\end{array } \right ] , \mathbf{\dot{\lambda}}\left ( t\right ) = \left [ \begin{array } [ c]{c}\mathbf{\dot{\lambda}}(t)\\ \mathbf{\dot{\lambda}}(t ) \end{array } \right].\ ] ] second , we further write the characteristic equation for the matrix @xmath61 as @xmath62 , where @xmath63 is the eigenvalue and @xmath64^{t}$ ] the @xmath65th eigenvector . vectors @xmath66 and @xmath67 containing the familiar bogoliubov @xmath68-@xmath69 parameters should fulfill the normalized condition of @xmath70 , where vectors @xmath71 ^{t}$ ] and @xmath72 ^{t}$ ] . by solving the corresponding secular equation of matrix @xmath73
, we obtain eigenvalues @xmath63 , which are a pair of zero frequency mode @xmath74 and two pairs of nonzero excited modes @xmath75 given by@xmath76 with @xmath77 , @xmath78 , @xmath79
. we can prove that without atom - atom collision(@xmath80 ) , @xmath81 are real and the cpt state is dynamically stable . however , when the atom - atom collision is present , in the spaces of @xmath82 or @xmath83 , @xmath81 are not real and thus the cpt state solution becomes unstable . in experiments , @xmath42 and @xmath84
are two controllable parameters .
thereby , we study the dynamical stability for the cpt in ( @xmath42 , @xmath84 ) space . to get an intuitive picture for the stability ,
we plot a stability diagram in figure [ stability_dia ] , where a pair of gaussian pulses is adopted with the pulse width @xmath85 : @xmath86 \label{gaussian}\ ] ] where @xmath87 and @xmath88 are the central positions and peak strengths , respectively . in our calculations ,
a pair of fixed laser pulses is applied : @xmath89s@xmath90 , @xmath91s , @xmath92s , @xmath85=4@xmath28s .
figure [ stability_dia ] shows unstable regions , respectively , labeled by ( @xmath82 ) , ( @xmath83 ) , and stable regions labeled by .
strict mathematical results for these regions are listed below : region is for @xmath82 , given by@xmath93 region is for @xmath83 , given by@xmath94where @xmath95 is the real root derived from a cubic equation @xmath96 .
region is for dynamical stable regimes .
unstable_areasnn.eps after deciding the stability properties , we proceed to study the adiabaticity of the cpt state .
we turn our attention to eigenstates of the matrix @xmath97 .
the eigenstate with @xmath74 corresponding to the goldstone mode takes a non - normalized form of@xmath98 since the goldstone mode is degenerate , we need to introduce a new vector @xmath99 complimentary to @xmath100 given by@xcite @xmath101 where @xmath102 is determined from a normalization condition @xmath103 and @xmath104 is given in equation ( [ eta_pm ] ) . by solving equations ( [ mq_eq ] ) and ( [ normalized ] ) simultaneously , we find @xmath99 has the following structure:@xmath105 where elements @xmath106 are@xmath107 and the coefficient @xmath69 is @xmath108 .
all of the base vectors including @xmath100 , @xmath99 , @xmath109 corresponding to @xmath110 create a complete 6@xmath566 space in the sense that they obey the following biorthonormality relations:@xmath111 @xmath112 where @xmath104 and @xmath113 are two metric matrices defined as@xmath114{cc}\mathbf{i } & 0\\ 0 & -\mathbf{i}\end{array } \right ] , \eta_{-}=\left [ \begin{array } [ c]{cc}0 & \mathbf{i}\\ -\mathbf{i } & 0 \end{array } \right ] \label{eta_pm}\ ] ] and @xmath115 being a 3@xmath563 unit matrix . equipped with this totally complete space , we are able to expand @xmath116 in the form of:@xmath117 inserting equation ( [ superp_chi ] ) into equation ( [ chi_eq ] ) and using biorthonormality relations equations ( 17 ) and ( 18 ) , one could obtain the dynamic equations for @xmath118 : @xmath119 in deriving equation ( [ goldstonen ] ) , we have assumed that the vectors @xmath100 and @xmath99 are time independent , so that the source terms @xmath120 will vanish .
thereby , @xmath121 and @xmath122 would be decoupled from the motional equation for the vector @xmath116 .
apparently , here we only need to focus on the values @xmath123 and @xmath124 for the purpose of adiabatic condition .
the dynamical behavior for @xmath125 is governed by@xmath126 in the adiabatic limit , @xmath127 , projections @xmath128 can be treated as time - independent values since @xmath129 and @xmath22 are usually sufficiently large .
now , we estimate @xmath128 up to the first order of @xmath130 , and obtain the following set of linearly coupled equations:@xmath131{l } \left ( \begin{array } [ c]{cccc}i\omega_{1}^{\ast}+\gamma g_{11 } & \gamma g_{12 } & { 0 } & -\gamma f_{12}\\ \gamma g_{12}^{\ast } & i\omega_{2}^{\ast}+\gamma g_{22 } & \gamma f_{12 } & 0\\ 0 & -\gamma f_{12}^{\ast } & \gamma g_{11}^{\ast}-i\omega_{1 } & \gamma g_{12}^{\ast}\\ \gamma f_{12}^{\ast } & 0 & \gamma g_{12 } & \gamma g_{22}^{\ast}-i\omega_{2}\end{array } \right ) \left ( \begin{array } [ c]{c}c_{1}\\ c_{2}\\ c_{1}^{\ast}\\ c_{2}^{\ast } \end{array } \right ) \\ \nonumber = \left ( \begin{array } [ c]{c}-\mathbf{w}_{1}^{\dag}\mathbf{\dot{\lambda}}(t)\\ -\mathbf{w}_{2}^{\dag}\mathbf{\dot{\lambda}}(t)\\ -\mathbf{w}_{1}^{t}\mathbf{\dot{\lambda}}(t)\\ -\mathbf{w}_{2}^{t}\mathbf{\dot{\lambda}}(t ) \end{array } \right ) \end{array } \label{coupled_excit}\ ] ] where coefficients @xmath132 and @xmath133 are denoted by @xmath134 and@xmath135 as for @xmath136 , @xmath137 is replaced by @xmath138 in equation ( [ source_term ] ) . compared with the results of equation ( 42 ) in @xcite ,
equation ( [ coupled_excit ] ) has been extended to include unstable parameter regions . in @xcite
, an adiabatic parameter @xmath139 is introduced to measure the quality of adiabaticity . in general , @xmath140 is required to keep good adiabaticity of the cpt state . after solving equation ( [ coupled_excit ] )
, we could calculate the adiabatic parameter @xmath141 .
however , the general analytical formula is very complicated for analysis . in the following ,
we calculate the adiabatic parameter in a special case of @xmath142 . in the derivation of adiabatic parameter @xmath141 , we are pleased to see if @xmath84 is tuned to 0 , i.e. on effective one - photon resonance , equation ( [ chi_eq ] )
can be further simplified to @xmath143 where the vector @xmath144 is described as @xmath145 ^{t}$ ] with elements @xmath146(@xmath147 ) . in this case , we find the coupling coefficient matrix @xmath148 turns out to be@xmath149{ccc}0 & -\omega_{p}\phi_{a } & 0\\ -\omega_{p}\phi_{a } & 0 & -\frac{\omega_{s}}{2}\\ 0 & -\frac{\omega_{s}}{2 } & 0 \end{array } \right ) \label{matrix_p}\ ] ] surprisingly , in contrast to @xmath61 in equation ( [ chi_eq ] ) , now @xmath148 becomes 3@xmath563 dimensional and symmetrical , and its eigenvalues and eigenvectors are all analytically solvable , which are@xmath150 and @xmath151 ^{t},\label{eigenvectors1 } \\
\mathbf{w}_{\pm}^{\prime } & = \frac{1}{\sqrt{2\omega_{s}\omega_{eff}}}\left [ 2\phi_{a}\omega_{p},\pm\sqrt{\omega_{s}\omega_{eff}},\omega_{s}\right ] ^{t}.\label{eigenvectors2}\ ] ] more remarkable , the eigenvectors become complete and orthogonal . as a result , any bare vector @xmath152 in equation ( [ psi_vec_eq ] )
can be directly expanded by @xmath153 , i.e. @xmath154 , leading to a new set of linearly coupled equations,@xmath155{cc}\mathbf{c } & \mathbf{d}\\ -\mathbf{d } & -\mathbf{c}\end{array } \right ) \left ( \begin{array } [ c]{c}\mathbf{\dot{c}}\\ \mathbf{\dot{c}}^{\ast}\end{array } \right ) + \left ( \begin{array } [ c]{cc}\mathbf{\gamma } & 0\\ 0 & \mathbf{\gamma}\end{array } \right ) \left ( \begin{array } [ c]{c}\mathbf{\dot{c}}\\ \mathbf{\dot{c}}^{\ast}\end{array } \right ) = -\left ( \begin{array } [ c]{c}\mathbf{\pi}\\ \mathbf{\pi}\end{array } \right ) .
\label{n_coupled_eq}\ ] ] here vectors
@xmath156 ^{t}$ ] , @xmath157 ^{t}$ ] and other coupled coefficient matrices are@xmath158{cc}\kappa_{- } & \kappa_{0}\\ \kappa_{0 } & \kappa_{+}\end{array } \right ) , \mathbf{d=}\left ( \begin{array } [ c]{cc}\kappa_{0 } & \kappa_{0}\\ \kappa_{0 } & \kappa_{0 } \end{array } \right ) , \mathbf{\gamma}=\left ( \begin{array } [ c]{cc}\frac{\gamma}{2 } & -\frac{\gamma}{2}\\ -\frac{\gamma}{2 } & \frac{\gamma}{2}\end{array } \right)\ ] ] with the corresponding elements @xmath159 in obtaining equation ( [ n_coupled_eq ] ) , we consider @xmath160 for the adiabatic limit . in addition , the zero modes @xmath161 and @xmath162 are left decoupled due to the weak coupling strengths and the eliminated source terms .
based on the above analysis , the adiabatic parameter @xmath141 can be solved analytically@xmath163 with @xmath164 .
equation([ann ] ) agrees with equation ( 11 ) in @xcite in the case of @xmath165 . in equation ( [ ann ] ) , only the parameter @xmath166 is related to the atom - atom collisional strength .
evidently , when @xmath167 which could be met by using strong pump laser such that @xmath168 , the adiabatic parameter @xmath141 is nearly immune to atom - atom collisional effect .
in this subsection , we numerically study the relation of the stability and adiabaticity for the cpt state with atom - atom collision . in figure
[ adiabaticity_pic ] , we plot the distributions of the adiabatic parameter @xmath141 defined in equation ( [ r_parameters ] ) in @xmath169 parameter space at time @xmath170 .
the unstable and stable regions , whose symbols , and are the same as those used in figure [ stability_dia ] , are also shown in figure [ adiabaticity_pic ] .
black curves stand for the boundary of unstable regions at @xmath171 .
it is interesting to see that in the region near the boundary of the stable region and unstable region , @xmath141 is very small ( @xmath10.1 ) no matter how strong the atom - atom collision @xmath42 is . and
, @xmath141 at the right corner of the upper unstable region is even lower than those close to this corner but in the stable region .
so , in our scheme , good adiabaticity could even be achieved in the dynamical unstable regions .
adiabaticity_insta.eps it is widely accepted that the @xmath141 values in unstable regions are in general larger than those in stable regions and hence good adiabaticity could be hardly achieved in the dynamical unstable parameter regions @xcite .
so , our result is counterintuitive . to understand this result
, we introduce @xmath172im@xmath173(@xmath174 and @xmath175 are conjugate with each other ) in which i m denotes the imaginary part of a complex number .
@xmath2 measures unstable growth rate .
we plot the time evolution of @xmath2 in figure [ omega_com_pic ] , respectively , in unstable region with parameters @xmath176mhz , @xmath177mhz(figure [ omega_com_pic](a ) ) , and in unstable region with parameters @xmath176mhz , @xmath178mhz(figure [ omega_com_pic](b ) ) .
figure [ omega_com_pic ] shows that the unstable growth rates in both regions have a gaussian - like shape with a very small time width .
the peak value of the unstable growth rate in the unstable region is smaller than that in the unstable region , so the adiabatic parameter in the unstable region is smaller than that in the unstable region . in figure
[ omega_com_pic ] , we also plot the time evolutions of the pump and control laser intensities @xmath179 .
when two lasers overlap in time domain , the unstable growth rate is nonzero .
thus , if the overlapping time of the two lasers is short enough , the unstable growth of the deviation from the cpt state could be negligible and then a good adiabaticity could be maintained in this situation .
omega_com.eps in this section , we further investigate the relations of the atom - molecule transfer efficiency to the stability and adiabaticity of a cpt state . the heisenberg mean - field equations ( [ dy1])-([dy3 ] ) are numerically solved using the same parameters ( @xmath180 ) as those used in figure [ stability_dia ] .
figure [ population_dy ] shows the final atom - molecule transfer efficiency in ( @xmath181 ) parameter space , which are obtained when intensities of two pulsed optical fields are essentially zero after a long time . comparing figure [ population_dy ] with figure [ adiabaticity_pic ]
, we find that in unstable region , the transfer efficiency is always low ( @xmath110@xmath182 ) , and the high transfer efficiency ( @xmath18380@xmath182 ) is achieved in the region with very good adiabaticity ( the unstable region and the region that is part of the stable region but is also close to region ) . in figure
[ population_dy ] , a nearly vertical white line close to @xmath184 indicates the maximum transfer efficiencies according to different @xmath42 values .
the details of the transfer efficiency along this line versus @xmath42 are plotted in the inset .
obviously , by employing the effective one - photon resonance condition , the atom - molecule transfer efficiency decreases very slowly , just from 88@xmath182 at @xmath80 to 87@xmath182 at @xmath185mhz .
that is quite a small decrease compared with a large change of @xmath42 values .
actually , this is due to insensitivity of the adiabaticity to the collisional strength between atoms , as discussed at the end of section 2 .
dynamics0520.eps now we proceed to study how the atom - atom collision and the effective one - photon resonance could affect the atom - molecule transfer processing .
figure [ population_opt ] shows the time evolution of ground - state atom population(upper ) , excited - state molecule population(middle ) and ground - state molecule population(bottom ) for one case with an effective one - photon resonance ( @xmath142 , solid ) , and for other two cases with @xmath186(dashed ) and 2.0mhz(dotted ) .
a state very close to the cpt state ( @xmath187 ) is almost produced for different detuning schemes .
but , since a good adiabatic condition of our cpt state could be achieved with effective one - photon detuning even in dynamical unstable regions , figure [ population_opt ] shows that the excited molecule population for the case with effective one - photon is much lower than those without this condition .
a good adiabaticity could guarantee a high atom - molecule transfer efficiency when the atom - atom collision presents .
the left and right panels of figure [ population_opt ] plot a comparison of the atom - molecule transfer processing for two quite different atom - atom collisional strengths @xmath42=2.0mhz and 5.0mhz .
when @xmath42 is relatively low ( 2.0mhz ) , the final atom - molecule transfer efficiency is not sensitive to single photon detuning .
in contrast , when @xmath42 increases to 5.0mhz , the final transfer efficiency becomes very sensitive to it : without effective single - photon resonance , the final efficiency is just around 10@xmath182 , and a robust atom - molecule transfer ( efficiency@xmath188@xmath182 ) can be achieved when the effective single photon detuning is employed .
in fact , the cpt state in our scheme already includes the collisional effect .
when the collision strength @xmath42 is relatively weak , the cpt state with collisions is almost close to those without collisions .
when the collisional strength is strong enough , collision effect has strong influences on the stability of the cpt state and thus the final atom - molecule transfer efficiency becomes sensitive to the detuning .
fortunately , the collision nonlinearity could be compensated by using the effective single photon resonance and thus we can still achieve a much higher atom - molecule transfer efficiency than those not satisfying this condition .
populations .[population_opt ] before ending , it is necessary for us to verify the experimental feasibility of our scheme , especially the parameters we have used in the calculations .
table [ table1 ] organized from @xcite enables to show the background @xmath15-wave scattering lengths @xmath189 and collisional strengths @xmath42 of some usual alkali - metal atoms .
@l*15c species & @xmath190g@xmath191 & @xmath192nm@xmath191 & @xmath193(mhz ) + @xmath194 li & 543.25(5 ) & 3.122 & 0.415 + & 834.149 & -74.348 & -9.880 + @xmath195 na & 853 & 3.381 & 0.117 + & 907 & 3.323 &
0.115 + @xmath196 k & 202.10(7 ) & 9.208 & 0.184 + & 224.21(5 ) & 9.208 & 0.184 + @xmath197 rb & 155.0 & -23.442 & -0.220 + @xmath198 rb & @xmath199 & 5.318 & 0.049 + @xmath200 cs & 19.90(3 ) & 8.625 & 0.052 + & 47.97(3 ) & 47.890 & 0.287 + table [ table1 ] provides the parameters @xmath189 and @xmath42 characterizing the background scattering potential for experimentally relevant zero - energy resonances . explicitly , almost all the collisional strengths @xmath201 values are much smaller than 10.0mhz(except @xmath194li at @xmath202 g ) .
thereby , results discovered from figures [ population_dy ] and [ population_opt ] are universal and can be available with any atomic species . in our other numerical simulations ( not shown ) , we use laser pulses with different intensities and widths , and even take other collisions into account , including collisions between atom - molecule and molecule - molecule .
pleasantly , we find that the results are still perfectly consistent with figure [ population_dy ] and other texts .
finally , we note that the effective one- and two - photon resonances , which are the keys to obatin a high atom - molecule transfer efficiency , can be realized with current technique of frequency chirped pump and control laser pulses @xcite . in experiments ,
it s hard to exactly maintain the effective single photon resonance .
actually , figure [ adiabaticity_pic ] also shows that when @xmath203mhz , @xmath141 is much smaller than 0.1 , very good adiabaticity and a high atom - molecule transfer efficiency could be obtained .
in summary , we have investigated the atom - molecule transfer within a @xmath2-type configuration including atom - atom collision .
we first obtain a cpt state including the atom - atom collisional effect under the effective two - photon resonance condition and find the collision induced frequency shift between two quantum states .
further , we study the stability and adiabaticity of the cpt state .
we find that instability could be essentially induced when the pump and control lasers overlap and thus when the overlapping time of the pump and control lasers is short so that the unstable growth of the deviation from the cpt state could be neglected , good adiabaticity of the cpt state can be obtained in some unstable regions , not just in stable ones .
employing the effective single photon resonance , our numerical simulations show that a rather high atom - molecule transfer efficiency ( @xmath18380@xmath182 ) could be achieved .
finally , the feasibility of our scheme for future experiments is discussed .
our findings may provide a new way to efficiently photoassociate ground state molecules from a cold atomic gas with strong atom - atom collisional interaction .
this work was supported by the national basic research program of china ( 973 program ) under grants no .
2011cb921602 and no .
2011cb921604 , and the national natural science foundation of china under grant no . 11104076 , no .
11034002 and no . 10874045 and the specialized research fund for the doctoral program of higher education no 20110076120004 .
99 doyle j , friedrich b , krems r v and masnou - seeuws f 2004 eur .
j. d * 31 * 149 . | we investigate the relationship between stability , adiabaticity and transfer efficiency in a @xmath0-type atom - molecule coupling system via a nonlinear stimulated raman adiabatic passage .
we find that only when the pump and control lasers overlap in time domain , the coherent population trapping ( cpt ) state could become unstable .
if the overlapping time of the two lasers is short so that unstable growth of the deviation from the cpt state is negligible , then good adiabaticity of the cpt state could be maintained even in the unstable region . in this case
, a high atom - molecule transfer efficiency could be obtained by chirping applied laser pulses to elegantly compensate the frequency shift induced by intra - atomic collision .
our results could be useful for efficiently photoassociating ground - state molecules from a cold atomic gas with strong atom - atom collisional interaction . |
the baryon asymmetry of the universe may be expressed in terms of the ratio of the baryon density @xmath0 to the entropy density @xmath1 of the universe .
the agreement between astrophysical observations and primordial nucleosynthesis calculations constrains this ratio to the range at @xmath2 c.l .
@xcite : @xmath3 which is in good agreement with the inferred from cosmic microwave background data , particularly that provided by wmap @xcite .
almost 40 years ago , sakharov explained how this baryon asymmetry could be obtained via a microphysical mechanism incorporating interactions that violate baryon and/or lepton number , violate c and cp , and drop out of equilibrium in the early universe @xcite .
since such interactions are present in the electroweak theory and its extensions , the baryon asymmetry might have originated at any energy above the electroweak scale .
one of the most attractive and simple ways to realize the sakharov mechanism is high - scale leptogenesis through the out - of - equilibrium decays of a heavy majorana neutrino , since it arises naturally within the seesaw mechanism for generating light neutrino masses @xcite . however , high - scale leptogenesis and the seesaw mechanism are difficult to test directly , since the natural scale of the interactions of the heavy majorana neutrino is in the range @xmath4 gev to @xmath5 gev .
this is impossible to reach by accelerators , except indirectly via renormalization effects on light particle masses , for example .
baryogenesis at the electroweak scale itself is certainly a most attractive option@xcite , because it may be tested directly , if it can be achieved economically .
however , this option is now excluded in the standard model , because the unsuccessful lep searches require the higgs boson to be too heavy for the electroweak phase transition to have been first - order , and the effective amount of cp violation is in any case very small @xcite . on the other hand ,
an electroweak mechanism may be possible in the minimal supersymmetric extension of the standard model ( mssm ) , but at the price of a certain number of rather restrictive conditions @xcite .
it is therefore worthwhile to look for other scenarios for electroweak baryogenesis .
an additional problem with any supersymmetric model that postulates high temperatures in the early universe is that of the gravitino abundance @xcite .
thermal production of gravitinos at high temperatures in the very early universe might have exceeded the bounds imposed by primordial nucleosynthesis and the cosmic microwave background @xcite .
therefore , in the context of supersymmetry , it is doubly interesting to look for alternative baryogenesis scenarios that operate at lower energy scales below about a tev , which might be tested directly in the foreseeable future .
it has recently been proposed that soft supersymmetry - breaking breaking terms comprising bilinear and trilinear scalar couplings involving the right - handed sneutrino fields could be responsible for leptogenesis , an option referred to as soft leptogenesis @xcite .
in contrast to standard leptogenesis from heavy majorana ( s)neutrino decay , in soft leptogenesis the cp violation needed can be provided within the framework of a single sneutrino generation . however , even if this mechanism of leptogenesis could evade the gravitino problem , it is still difficult to probe in laboratory experiments , because of the high mass scale of the right - handed sneutrino . in this paper , we propose an alternative leptogenesis mechanism via _ decays of a left - handed sneutrino at a low energy scale_. this scenario may be realized in supersymmetric models with non - zero majorana neutrino masses that lead to mixing and mass splitting between the left - handed sneutrino @xmath6 and the corresponding antisneutrino @xmath7 .
as was shown in @xcite , the neutrino mass and the sneutrino mass splitting may be due to related @xmath8 interactions , linked by supersymmetry breaking .
as we show below , the @xmath9 system may exhibit oscillatory behaviour analogous to the @xmath10 system .
in addition , soft supersymmetry breaking sectors provide new sources of cp violation in @xmath6-@xmath7 mixing which can generate a lepton asymmetry in the sneutrino decays .
we explore in this paper how the three sakharov conditions @xcite for generating the observed baryon asymmetry of the universe can be fulfilled in this restrictive framework . in the sense
that supersymmetry breaking is the source of leptogenesis , our scenario is similar to soft leptogenesis , but differs in that it operates at the electroweak scale .
if the scenario proposed does become a candidate for leptogenesis , it could in principle be tested in collider experiments via a sneutrino oscillation signal , though this may be difficult in practice .
the lepton number is tagged in the decay of sneutrino by identifying the charge of the outgoing lepton , and a same - sign dilepton signal may be observable when the sneutrino - antisneutrino pairs decay into charged leptons .
however , measurement of a lepton asymmetry would probably need more events than are likely to be provided at the lhc or ilc .
the outline of this paper is as follows . in the next section ,
we review how to generate the sneutrino mass splitting in a supersymmetric model with a heavy right - handed majorana neutrino superfield and discuss how small the mass splitting can be . in section iii
, we examine how leptogenesis via the decay of the left - handed sneutrino can be realized and in section iv the cp asymmetry generated from the decay of the light sneutrino is considered . in section
v , we discuss possible wash - out processes , which may be harmless for our leptogenesis scenario , and present our conclusions .
we first review in more detail how to generate the sneutrino mass splitting in a supersymmetric model with non - zero majorana neutrino masses , which could arise from a right - handed neutrino superfield .
the relevant superpotential is @xmath11 where @xmath12 are superfields for lepton doublets , heavy right - handed neutrinos and higgs fields , respectively .
the @xmath13 terms are the same as in the mssm .
the relevant terms in the soft supersymmetry - breaking scalar potential are given by @xmath14 where @xmath15 and @xmath16 are the light and heavy sneutrinos , respectively .
from now on , we consider a single generation of @xmath17 and @xmath18 .
after the electroweak symmetry is spontaneously broken , the light neutrino acquires a mass via the seesaw mechanism : @xmath19 , where @xmath20 with @xmath21 .
defining @xmath22 one can separate the sneutrino mass - squared matrix into two blocks and then , to first order in @xmath23 , the two light sneutrino eigenstates are @xmath24 and @xmath25 , with the following masses - squared , @xmath26 where the mass - squared difference @xmath27 is of order @xmath23 . here , we can remove the phases in the superpotential ( [ superpotential ] ) by rotating superfields and the relative phase between @xmath28 and @xmath29 with a r - rotation , but there remains an unremovable phase , which makes @xmath30 complex . the sneutrino mass splitting
is then easily computed by using the relation @xmath31 @xcite , where @xmath32 is the average of the light sneutrino masses and @xmath33 we note that @xmath34 and @xmath35 are of the same order as the electroweak scale , whereas @xmath36 and @xmath29 are soft - supersymmetry breaking parameters associated with the @xmath37 singlet superfield @xmath18 , and may be much larger than the electroweak scale . as one can see from ( [ sp1 ] ) , if @xmath38 , the sneutrino mass splitting is significantly enhanced , whereas it is of the same order as the neutrino mass when @xmath39 .
this mass splitting may in principle be probed through the sneutrino - antisneutrino oscillation which would result in a same - sign dilepton signal .
however , to have an observable rate for the same - sign dilepton signal , the ratio of the mass splitting to the sneutrino decay width should be large , namely , @xmath40 , and the sneutrino branching ratio into a charged lepton should also be large .
we now examine how leptogenesis via the decay of the light sneutrino can be realized in this framework . as is well known , for a successful mechanism of leptogenesis
, we need lepton - number- and cp - violating interactions that should be out of equilibrium , so that the asymmetry generated is not automatically suppressed . in this scenario , as mentioned before , _ sneutrino - antisneutrino mixing is a direct manifestation of lepton - number violation _ , and the _ soft supersymmetry - breaking terms may well provide a suitable new source of cp violation_. we must then impose the _ out - of - equilibrium condition for the decay of the light sneutrino _ , which is given by @xmath41 where @xmath42 is the decay rate of the sneutrino , @xmath43 is the hubble constant at the decay epoch , and @xmath44 is the planck scale .
the parameter @xmath45 is the effective number of massless degrees of freedom , which takes the value @xmath46 for the mssm with one generation of right - handed neutrinos . as well as the out - of - equilibrium condition ( [ out1 ] )
, we must require that the electroweak _ sphaleron interactions are still in thermal equilibrium _ at the time the lepton asymmetry is generated , so that they can convert the lepton asymmetry partially into a baryon asymmetry of the universe .
the temperature at which the sphaleron interactions freeze out depends on how the electroweak phase transition occurs .
it is known that the sphaleron interactions freeze out at the critical temperature of electroweak phase transition if it is strongly first - order , whereas the freeze - out temperature may become lower than the critical temperature if the transition is second - order or weakly first - order . in the latter case ,
the sphaleron interactions freeze out at the temperature at which the sphaleron transition rate @xmath47 becomes equal to the expansion rate of the universe@xcite .
it has been found in numerical simulations that the sphaleron interactions are effective as long as @xmath48 gev .
therefore , for a successful mechanism of leptogenesis , we require 510 ^ -4 _ 1.310 ^ -4()^2 , [ out2 ] where the first condition follows from the sphaleron equilibrium condition and the second from ( [ out1 ] ) for @xmath49 .
as seen in ( [ out2 ] ) , a scenario for leptogenesis can be successful only when @xmath50 gev .
in order to discuss the decays , mixing and cp - violating asymmetries for light sneutrinos , we first integrate out the heavy right - handed neutrino superfield , following which the superpotential for the light sneutrino is : @xmath51 where @xmath52 denote the light neutrino , the right - handed charged lepton and the charged higgs superfields , respectively , and @xmath53 stands for the yukawa couplings of the lepton sector , which are given by @xmath54 .
the soft supersymmetry - breaking terms involving the light sneutrinos @xmath15 are ; @xmath55 we note that the effects of the right - handed neutrino superfields are absorbed into the seesaw term in ( [ seesaw2 ] ) and the sneutrino mass splitting @xmath30 discussed previously .
the sneutrino interaction lagrangian is then given by @xmath56 where we have considered a single sneutrino generation and @xmath57 after the electroweak symmetry is broken . it should be noted that the higgs field @xmath58 is decomposed into the physical higgs sector and the goldstone boson sector . selecting the physical higgs sector in @xmath58 is equivalent to replacing @xmath58 by @xmath59 .
its yukawa coupling therefore becomes @xmath60 , where @xmath61 is the charged lepton mass . in the lagrangian for a single generation of sneutrino
, there is a physical cp - violating phase . with superfield rotations and an r - rotation we can eliminate cp phases in the parameters @xmath30 , @xmath62 and @xmath63 , but
then the cp - violating phase in @xmath64 can not be removed . as observed above in ( [ out2 ] ) , the mechanism of low - energy leptogenesis proposed here could be successful only if @xmath65 is of order @xmath66 ev , so it is important to discuss how such a low decay rate might be achieved .
if @xmath67 , which is the ordering of masses generally expected in the mssm with universal soft supersymmetry - breaking scalar masses ( the cmssm ) , the dominant sneutrino decays are those into two - body final states . as shown in @xcite , the typical size of @xmath65 for such two - body decays
is @xmath68(gev ) , which is too large to achieve successful leptogenesis . on the other hand ,
if @xmath69 and no two - body sneutrino decay channels are open , three - body sneutrino decays will dominate .
this mass ordering is possible if the gravitino is the lightest supersymmetric particle , and is also allowed by cosmology and the standard accelerator constraints @xcite .
the following chargino- and neutralino - mediated three - body decays are then generally dominant : @xmath70 and @xmath71 , assuming that decays into final states containing lighter sneutrinos can be neglected .
assuming that the lightest neutralino is dominated by its bino component , the rates for the chargino- and neutralino - mediated sneutrino decays ( see fig . [
fig : threebody](a ) for an illustration ) are given by @xcite ( _ ll^-^+_r _ ) & = & , + ( _ l_l_r^^ ) & = & , where the @xmath72 are gaugino mass parameters , @xmath73 and @xmath74 . .
_ , title="fig : " ] .
_ , title="fig : " ] in order to obtain @xmath75 ev , one must require that the @xmath6 and @xmath76 be nearly degenerate .
as an example , for @xmath77 , @xmath78 and @xmath79 , we indeed find @xmath75 ev .
this near - degeneracy between the @xmath76 and @xmath6 is not possible in the cmssm with universal soft scalar masses @xmath80 and gaugino masses @xmath81 at the gut scale , in which the mass difference between the @xmath76 and @xmath6 is approximately @xmath82 . according to the constraints on the parameter space @xmath83 from the cosmological observation , @xmath84 and muon @xmath85 ,
the favoured region of @xmath81 is somewhat large , namely @xmath86 gev , and the @xmath76 is considerably lighter than the @xmath6 @xcite in the cmssm .
thus , we need a non - universal boundary condition on the soft scalar masses at the gut scale .
there is no known theoretical or phenomenological reason why the soft superymmetry - breaking @xmath76 and @xmath6 masses should not differ at the gut scale . at the electroweak scale , @xmath87 and @xmath88
can be written as m__r & & m_0r^2 + 0.15m_1/2 ^ 2+^2_w2m_z^2 , + m _ & = & m_0l^2 + 0.54m_1/2 ^ 2 + 1/22m_z^2 , and we can make the physical @xmath87 and @xmath88 nearly degenerate by choosing the soft scalar masses @xmath89 and @xmath90 appropriately .
we now consider mass mixing and cp violation in the @xmath91 system .
since it is analogous to the @xmath92 and @xmath93 systems , we can calculate mixing and cp asymmetries using the formulae derived previously for these meson systems @xcite .
the evolution of the system is determined by a hamiltonian @xmath94 where , to leading order in the soft terms , @xmath95 where @xmath96 is the total @xmath15 decay width .
the dominant part of the off - diagonal component @xmath97 can be obtained by considering the imaginary part of the loop diagram shown in fig .
[ fig : threebody](b ) . the analytic expression for @xmath97
is given by @xmath98 where @xmath99 the eigenvectors of the hamiltonian @xmath43 are then given by @xmath100 where the parameters @xmath101 and @xmath102 are related by @xmath103 it is appropriate for cosmology to consider an initial state at @xmath104 with equal densities of @xmath15 and @xmath105 . at a later time
@xmath106 , the state will have evolved into @xmath107 here @xmath108 and we have neglected @xmath109 with respect to @xmath110 .
we can then compute the total integrated lepton asymmetry , defined by @xmath111}{\sum_f\int^{\infty}_{0}dt[\gamma(\widetilde{\nu}(t)\rightarrow f)+\gamma(\widetilde{\nu}^{\ast}(t)\rightarrow f)+\gamma(\widetilde{\nu}(t)\rightarrow \bar{f})+\gamma(\widetilde{\nu}^{\ast}(t)\rightarrow \bar{f})]}\\ & = & \frac{1}{2}\left(\left|\frac{q}{p}\right|^2-\left|\frac{p}{q}\right|^2\right ) \frac{\int^{\infty}_0dt|g_{-}|^2}{\int^{\infty}_0dt(|g_+|^2+|g_-|^2)},\end{aligned}\ ] ] where @xmath112 stands for a final state with lepton number equal to one and @xmath113 is its conjugate . calculating ( [ qop ] ) in the limit @xmath114 , we find @xmath115 we notice that @xmath64 is the only complex parameter in this expression .
performing the time integral , we find @xmath116 thus , we obtain the following final expression for the cp asymmetry : @xmath117 the baryon asymmetry is then given by @xcite @xmath118 the first factor takes into account the reprocessing of the @xmath119 asymmetry by sphaleron transitions , with the number of higgs doublets @xmath120 equal to 2 , and @xmath121 is the equilibrium sneutrino density in units of the entropy density for temperatures much larger than @xmath122 .
therefore , we obtain @xmath123 the efficiency factor @xmath124 includes effects caused by the sneutrino density being smaller than the equilibrium density and wash - out due to the decays not being completely out of equilibrium . the value of @xmath124 must be obtained by integrating the relevant boltzmann equations , but may plausibly be of order unity . for the value of @xmath125 given in ( 1 ) , we see that that our mechanism should yield @xmath126 .
let us estimate how we can achieve this desirable amount of @xmath127 .
we infer from recent experimental results that a neutrino mass of order @xmath128 ev is possible .
then , we see from ( 5 ) that @xmath129 in the case that the parameters @xmath130 are of the same order as @xmath88 . in this case , the size of the quantity in ( [ ofact ] ) is of order one .
we can also take the parameter @xmath131 to be the same order of @xmath88 .
moreover , we observe that the value of @xmath97 can be enhanced by taking large @xmath132 .
numerically , we find a magnitude of @xmath127 in ( [ cpasym ] ) of order @xmath133 when we take @xmath134 and @xmath135 , and a larger value of @xmath127 could be obtained if needed .
finally , we note that the large sneutrino - antisneutrino mixing which is expected because @xmath136 might lead to an observable like - sign dilepton signal .
this could provide an characteristic collider signature of our scenario . on the other hand ,
the amount of cp violation is likely to be too small to be observable in the near future .
the scenario proposed here needs further investigation . in particular , we recall that successful leptogenesis requires a departure from thermal equilibrium for wash - out processes , as well lepton - number - violating decays .
the dominant wash - out processes are scattering reactions mediated by sneutrinos or charginos and neutralinos . since our scenario
calls for relatively heavy charginos and neutralinos , the scattering rates mediated by charginos and neutralinos are suppressed . additionally , scattering reactions mediated by sneutrinos can occur via the vertices involving charginos or neutralinos . however , the rates of these scattering processes are suppressed due to the boltzmann suppression of the number densities of the charginos and neutralinos .
a more accurate estimation of the lepton asymmetry that survives these wash - out processes could be obtained by solving directly the full boltzmann equations for the system with different sparticle masses , as has been done in many detailed works considering high - scale leptogenesis @xcite .
a similarly detailed study goes beyond the scope of this paper , and will be undertaken elsewhere . in conclusion ,
an alternative of leptogenesis through decays of a left - handed sneutrino at a low energy scale has been proposed .
this scenario may be realized in supersymmetric models with non - zero majorana masses of neutrino superfield that lead to mixing and mass splitting between the left - handed sneutrino and the corresponding antisneutrino .
soft supersymmrty breaking sectors provide new sources of cp violation in sneutrino - antisneutrino mixing which can generate a lepton asymmetry in decays of the left - handed sneutrino .
we have shown that the three sakharov conditions @xcite for the observed baryon asymmetry of the universe can be fulfilled in this restrictive framework .
skk is supported in part by bk21 program of the ministry of education in korea and in part by kosef grant no .
r01 - 2003 - 000 - 10229 - 0 .
99 # 1#2#3phys .
* b#1 * , ( # 3 ) # 2 # 1#2#3nucl .
* b#1 * , ( # 3 ) # 2 # 1#2#3phys .
* d#1 * , ( # 3 ) # 2 # 1#2#3phys .
lett . * # 1 * , ( # 3 ) # 2 # 1#2#3mod .
* a#1 * , ( # 3 ) # 2 # 1#2#3phys
. rep .
* # 1 * , ( # 3 ) # 2 # 1#2#3science * # 1 * , ( # 3 ) # 2 # 1#2#3astrophys .
j. * # 1 * , ( # 3 ) # 2 s. eidelman _
[ particle data group ] , . c. l. bennett _ et al .
_ , astrophys
. j. suppl .
* 148 * ( 2003 ) 1 .
a. d. sakharov , zh .
fiz . pisma * 5 * , ( 1967 ) 32 ; jetp lett . * 91b * , ( 1967 ) 27 . m. fukugita and t. yanagida , .
v. a. kuzmin , v. a. rubakov and m. e. shaposhnikov , ; v. a.rubakov and m. s. shaposhnikov , phys .
usp . * 39 * , 461 ( 1996 ) .
k. jansen , nucl .
47 * , ( 1996 ) 196 ; k. rummukainen , m. tsypin , k. kajantie , m. laine and m. shaposhnikov , ; k. rummukainen , k. kajantie , m. laine and m. shaposhnikov , hep - ph/9809435 . m. carena , m. quiros , m. seco and c. e. wagner , , and references therein . j. r. ellis , j. e. kim and d. v. nanopoulos , ; m. l. khlopov and a. d. linde , . m. kawasaki , k. kohri and t. moroi , . y. grossman , t. kashti , y. nir and e. roulet , ; jhep * 11 * 080 ( 2004 ) ; g. dambrosio , g. f. giudice and m. raidal , ; y. grossman , r. kitano and h. murayama , hep - ph/0504160 . y. grossman and h. e. haber , .
m. hirsh , h. v. klapdor - kleingrothaus and s. g. kovalenko , ; e. j. chun , ; k. choi , k. hwang and w. y. song , .
m. shaposhnikov , jetp lett .
* 44 * , 465 ( 1986 ) .
j. ellis , k. a. olive , y. santoso and v. c. spanos , ; ; . for a review ,
see , e.g. p. j. franzini , phys , .
* 173 * ( 1989 ) 1 ; h. r. quinn , phys . rev .
* d 54 * ( 1996 ) 507 .
giudice , a. notari , m. raidal , a. riotto and a. strumia . | we propose an alternative mechanism for leptogenesis at the electroweak scale , through the decays of a left - handed sneutrino .
this scenario may be realized in supersymmetric models with non - zero majorana masses for the neutrino superfield that lead to mixing and mass splitting between the left - handed sneutrino and the corresponding antisneutrino .
soft supersymmetry breaking provides new sources of cp violation in sneutrino - antisneutrino mixing that can generate a lepton asymmetry in decays of the left - handed sneutrino .
we show how the three sakharov conditions for generating the observed baryon asymmetry of the universe can be fulfilled in this restrictive framework
. + cern - ph - th/2005 - 077 , snutp-05 - 005 + may 2005 |
the well understood features in the angular power spectrum of cosmic microwave background ( cmb ) anisotropies , such as acoustic peaks and the damping tail @xcite , allow a useful probe of the cosmology @xcite .
the ability of cmb anisotropies to constrain most , or certain combinations of , parameters that define cold dark matter cosmological models with adiabatic initial conditions has driven a significant number of experiments from ground and space .
these experiments include nasa s map mission and , in the long term , esa s planck surveyor . with the successful detection of the first two or three acoustic peaks at degree
angular scales @xcite , the anisotropy experiments have started to focus on fluctuations on angular scales of arcminutes or less . at these angular scales , fluctuations
are mostly dominated by secondary effects due to the local large scale structure ( lss ) between us and the last scattering surface @xcite .
in addition to generating new anisotropies via scattering and gravitational infall , non - linear effects such as the weak gravitational lensing of the cmb @xcite leave important imprints that can in turn be used to probe cosmology or astrophysics related to the evolution and growth of structure ( e.g. , @xcite ) . the lensing effect on the cmb leads to modifications in the direction of photon propagation while leaving the surface brightness of the cmb unaffected , implying that its contributions are effectively at the second order level in temperature fluctuations .
this second order contribution is probed by the three - point correlation function of the cmb @xcite . here , contributions arise mainly due to the fact that potentials that deflect cmb photons also correlate with other secondary effects from the foreground large scale structure , such as the integrated sachs - wolfe ( isw ; @xcite ) effect or the sunyaev - zeldovich ( sz ; @xcite ) effect . at the four - point level ,
the weak lensing effect itself contributes to the correlation function due to its non - linear mode - coupling nature .
the trispectrum , the four - point correlation function in fourier space , due to lensing was discussed in @xcite and the same trispectrum , under an all - sky formulation , is presented in @xcite . in @xcite , we presented the trispectrum due to lensing alone and the correlation between lensing and secondary effects and focused on the resulting non - gaussian covariance of the power spectrum measurements of the cmb temperature fluctuations . here
, we study an important application of the cmb trispectrum involving a model independent recovery of statistics related to the foreground angular deflections of cmb photons ( or alternatively the projected potentials that generate these deflections ) . in terms of the trispectrum formulated under a flat - sky approximation
, we show explicitly how one can use a higher order statistic such as the power spectrum of squared temperatures to extract the lensing information . the approach described in the present paper
was first introduced by ref . @xcite while other approaches were considered in refs .
@xcite based on cmb temperature and polarization observations .
we discuss the full trispectrum formed by the weak lensing effect on the cmb and extend previous work on this subject by calculating in detail the noise associated with lensing extraction .
we also discuss noise contributions resulting from the cmb trispectrum formed via the correlation between lensing and secondary effects due to the large scale structure .
since the lensing effect modifies the damping tail of cmb anisotropies and the information from the damping tail is used to reconstruct lensing statistics , we suggest that any secondary effect that dominates small angular scale fluctuations can have a significant impact on the lensing extraction .
an important contribution at these scales is the sz effect ; fortunately this contribution can be removed from thermal cmb maps by noting its distinct spectral signature in observations involving many frequencies about the sz null at @xmath0 217 ghz @xcite .
as discussed in @xcite , the lensing extraction is essential for an important application utilizing cmb polarization observations .
it is now well known that the curl ( or b - modes ) of the polarization allows a direct detection of the gravitational wave contribution @xcite .
the amplitude of the angular power spectrum of this polarization signal is proportional to the square of the energy scale of inflation , and thus any detection of the b - mode polarization allows a probe of the inflationary energy scale .
there is one significant source of confusion , however , in the b - mode map which can potentially limit the detection of the gravitational wave signal . as discussed in ref .
@xcite , the weak lensing effect on cmb polarization results in a fractional conversion of the dominant e - mode polarization due to density perturbations to the b - mode .
the contribution from lensing to the b - mode map is such that its power spectrum is at the level of the polarization signal expected from gravitational waves if the energy scale of inflation is at scales of @xmath0 @xmath1 gev .
since we do not know the energy scale of inflation a priori , if it is sufficiently small the lensing effect will easily dominate the contribution from primordial gravitational waves . in order to identify the gravitational wave signature with no confusion
, one must separate the contribution due to lensing reliably . as discussed in @xcite , observations of weak lensing in large scale structure via galaxy ellipticities
do not allow a useful separation of the lensing contribution .
a substantial contribution to the effect comes from potentials at redshift greater than 1 to 2 , the redshift range currently accessible with galaxy lensing surveys .
even if we have a reliable lensing survey out to a redshift of 5 , we will only be able to reduce the lensing signal by a factor of @xmath0 4 , which is hardly of any use given the order of magnitude change in the gravitational wave contribution that is possible if the inflationary energy scale is much less than that related to grand unified theories @xcite .
thus , a better approach is to use a source at the last scattering itself , ie .
cmb anisotropies , to reconstruct the lensing potential .
with such a reconstruction , one can clean the polarization maps and obtain unbiased and model independent estimates of the polarization signal at the last scattering surface . here , we discuss in detail the approach related to the cmb trispectrum and associated non - gaussian noises , ignored previously in lensing reconstruction @xcite and in the application of lensing reconstruction to separate the gravitational wave signature @xcite . under certain conditions ,
we find that these additional noise contributions are an order of magnitude below the dominant gaussian noise contribution . if the noise - bias associated with the gaussian part of the temperature squared power spectrum is removed , then the limiting noise level for lensing extraction is determined by the non - gaussian contribution presented here .
even if the noise - bias is not removed , we note that the proposed statistic allows a high signal - to - noise extraction of the lensing signal and a separation of the confusing lensed e - mode contribution in the b - mode map from those due to inflationary gravitational waves .
in [ sec : generallensing ] , we introduce the basic ingredients for the present calculation .
the cmb anisotropy trispectra due to weak lensing and correlations between weak lensing and secondary effects are derived in
[ sec : lensing ] . in
[ sec : covariance ] , we introduce a statistic involving the power spectrum of squared temperatures , instead of the usual temperature itself , as a probe of the angular power spectrum of lensing deflections and extend the discussion to involve polarization in [ sec : ebtri ] . in [ sec : results ] , we discuss our results and conclude with a summary in [ sec : summary ] .
the appendix contains useful formulae related to additional estimators of lensing based on polarization and a combination of temperature and polarization .
large - scale structure between us and the last scattering surface deflects cmb photons propagating towards us .
since the lensing effect on cmb is essentially a redistribution of photons from large scales to small scales , the resulting effect appears only in the second order @xcite . in weak gravitational lensing , the deflection angle on the sky
is given by the angular gradient of the lensing potential , @xmath2 , which itself is a projection of the gravitational potential , @xmath3 ( see e.g. @xcite ) , @xmath4 the quantities here are the conformal distance or lookback time , from the observer , given by @xmath5 and the analogous angular diameter distance @xmath6 with the expansion rate for adiabatic cdm cosmological models with a cosmological constant given by @xmath7\,.\ ] ] here , @xmath8 can be written as the inverse hubble distance today @xmath9mpc .
we follow the conventions that in units of the critical density @xmath10 , the contribution of each component is denoted @xmath11 , @xmath12 for the cdm , @xmath13 for the baryons , @xmath14 for the cosmological constant .
we also define the auxiliary quantities @xmath15 and @xmath16 , which represent the matter density and the contribution of spatial curvature to the expansion rate respectively .
note that as @xmath17 , @xmath18 and we define @xmath19 .
though we present a general derivation of the trispectrum contribution to the covariance , we show results for the currently favorable @xmath14cdm cosmology with @xmath20 , @xmath21 , @xmath22 and @xmath23 .
the lensing potential in equation ( [ eqn : lenspotential ] ) can be related to the well known convergence generally encountered in conventional lensing studies involving galaxy shear @xcite @xmath24 where note that the 2d laplacian operating on @xmath3 is a spatial and not an angular laplacian .
expanding the lensing potential to fourier moments , @xmath25 we can write the usually familiar quantities of convergence and shear components of weak lensing as @xcite @xmath26 while the two terms @xmath27 and @xmath28 contain differences with respect to radial and wavenumber weights , we note that these differences are artificial and that they cancel under the limber approximation @xcite . the power spectrum of the lensing potential is defined as @xmath29 where @xmath30 is the dirac delta function . expanding the gravitational potential to density perturbations using the poisson equation @xcite @xmath31 and the expansion of a plane wave ,
we can write the power spectrum of lensing potentials as @xmath32 here , @xmath33 and we have introduced the power spectrum of density fluctuations @xmath34 where @xmath35 in linear perturbation theory .
we use the fitting formulae of @xcite in evaluating the transfer function @xmath36 for cdm models . here
, @xmath37 is the amplitude of present - day density fluctuations at the hubble scale ; with @xmath38 , we adopt the cobe normalization for @xmath37 of @xmath39 @xcite , consistent with galaxy cluster abundance ( @xmath40 @xcite ) , note that in linear theory , the power spectrum can be scaled in time , @xmath41 , using the the growth function @xcite @xmath42 in the non - linear regime , one can use prescriptions such as the fitting function by @xcite to calculate the fully non - linear density field power spectrum .
note that an expression of the type in equation ( [ eqn : cllens ] ) can be evaluated efficiently with the limber approximation @xcite . here
, we employ a version based on the completeness relation of spherical bessel functions @xcite @xmath43 where the assumption is that @xmath44 is a slowly - varying function .
a similar approach is also presented in @xcite . using this
, we obtain a useful approximation for the power spectrum of lensing potentials as @xmath45 ^ 2 p\left({l \over{d_a}};{r}\right ) \ , .
\label{eqn : powerform}\ ] ] using equation ( [ eqn : kappa ] ) , we note that the power spectrum of convergence is related to that of the potentials through @xmath46 .
in order to derive weak lensing contributions to the cmb trispectrum , we follow hu in ref .
@xcite and zaldarriaga in ref .
we formulate the contribution under a flat sky approximation . in general
, the flat - sky approach simplifies the derivation and computation through the replacement of summations over wigner symbols through integrals involving mode coupling angles . as discussed in prior papers @xcite , weak lensing remaps temperature projected on the sky through the angular deflections resulting along the photon path by @xmath47 as expected for lensing , note that the remapping conserves the surface brightness distribution of cmb . here
, @xmath48 is the unlensed primary component of cmb , the contribution at the last scattering surface , and @xmath49 is the contribution affected by the gravitational lensing deflections during the transit .
it should be understood that in the presence of low redshift contributions to cmb resulting through large scale structure , the temperature fluctuations also include a secondary contribution which we denote by @xmath50 and , in all cases , a noise component denoted by @xmath51 due to instrumental and detector limitations .
we denote the total contribution involving these through @xmath52 , such that , @xmath53 . since weak lensing deflection angles also trace the large scale structure at low redshifts , secondary effects which are first order in density or potential fluctuations also correlate with the lensing deflection angle @xmath28 .
these secondary effects include the isw and sz and are discussed in @xcite .
taking the fourier transform , as appropriate for a flat - sky , we write @xmath54 where @xmath55
we define the power spectrum and the trispectrum of the @xmath56 fluctuation field , in the flat sky approximation , following the usual way @xmath57 where the subscript @xmath58 denotes the connected part of the above four - point function and the superscript @xmath59 denotes all possibilities including secondary anisotropies and noise .
we make the assumption that primordial fluctuations at the last scattering surface is gaussian , such that statistics are fully described by a power spectrum and the non - gaussian four - point correlator , @xmath60 is zero .
a non - gaussian contribution is generated by lensing effect and other secondary contributions .
the lensed power spectrum , according to the present formulation , is discussed in @xcite and we can write @xmath61 \ , c_l^{\theta}+ { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } c_{| { { \bf l}}- { { \bf l}}_1|}^{\theta}c^{\phi\phi}_{l_1 } [ ( { { \bf l}}- { { \bf l}}_1)\cdot { { \bf l}}_1]^2 \ , .
\label{eqn : ttflat}\end{aligned}\ ] ] thus , the total power spectrum of cmb anisotropies is @xmath62 where we denote the power spectrum of the noise component by @xmath63 when @xmath64 is the fraction of sky surveyed , @xmath65 @xcite is the variance per unit area on the sky when @xmath66 is the time spent on each of @xmath67 pixels with detectors of net @xmath68 , and @xmath69 is the effective beam - width of the instrument .
the @xmath70 represent the power spectrum associated with contributing secondary effects . in the absence of secondary effects , especially in the cases where
secondary contributions can be separated from the primary fluctuations using statistical signatures or distinct information , such as the spectral dependence of the sz , we set @xmath71 .
note that in general , however , important secondary contributions such as the isw can not be separated and will lead to additional noise contributions due to correlations with the lensing potentials .
the calculation of the trispectrum follows similar to the power spectrum . here
, we explicitly show the calculation for one term of the trispectrum and add all other terms through permutations .
first we consider the cumulants involving four temperature terms in fourier space : @xmath72 as written , to the lowest order , we find that contributions come from the first order term in @xmath73 given in equation ( [ eqn : lfactor ] ) .
we further simplify to obtain @xmath74 { \int { d^2 { \bf l}_2 ' \over ( 2\pi)^2 } } { \theta}({{{\bf l}_2 } } ' ) { \phi}({{{\bf l}_2}}-{{{\bf l}_2 } } ' ) \ , ( { { { \bf l}_2}}- { { { \bf l}_2 } } ' ) \cdot { { { \bf l}_2 } } ' { \theta}({{{\bf l}_3 } } ) { \theta}({{{\bf l}_4 } } ) \big > \ ,
\nonumber \\ & = & c_{l_3}^{\theta}c_{l_4}^{\theta}\big < { \phi}({{{\bf l}_1}}+{{{\bf l}_3 } } ) { \phi}({{{\bf l}_2}}+{{{\bf l}_4 } } ) \big > \ , [ ( { { { \bf l}_1}}+ { { { \bf l}_3 } } ) \cdot { { { \bf l}_3 } } ] \ , [ ( { { { \bf l}_2}}+ { { { \bf l}_4 } } ) \cdot { { { \bf l}_4 } } ] + \ , { \rm perm.}\ , , \end{aligned}\ ] ] where there is an additional term through a permutation involving the interchange of @xmath75 with @xmath76 .
introducing the power spectrum of lensing potentials , we further simplify to obtain the cmb trispectrum due to gravitational lensing : @xmath77 + \ , { \rm perm . } \
, , \label{eqn : trilens}\end{aligned}\ ] ] where the permutations now contain 5 additional terms with the replacement of @xmath78 pair by other combination of pairs . the non - gaussian contribution to the trispectrum , through coupling of lensing deflection angle to secondary effects ,
can be calculated with the replacement of , say , @xmath79 and @xmath80 in equation ( [ eqn : tri ] ) by @xmath81 and @xmath82 containing the sources of secondary fluctuations , which will produce an addition contribution involving the @xmath83 contribution associated with @xmath84 .
note that we can no longer consider cumulants such as @xmath85 as the secondary effects are decoupled from recombination where primary fluctuations are imprinted .
however , contributions come from the correlation between @xmath86 and the lensing deflection @xmath28 . here , contributions of equal importance come from both the first and second order terms in @xmath73 written in equation ( [ eqn : lfactor ] ) . first , we note @xmath87 contributions to the trispectrum from the first two terms come through the second order term in @xmath73 , with the two @xmath28 terms coupling to @xmath86 . in the last term , contributions come from the first order term of @xmath73 similar to the pure lensing contribution to the trispectrum . after some straightforward simplifications , we write the connected part of the trispectrum involving lensing - secondary coupling as @xmath88 \left [ { { { \bf l}_4}}\cdot ( { { { \bf l}_2}}+{{{\bf l}_4 } } ) \right ] c^{\theta}_{|{{\bf l}}_1 + { { \bf l}}_3| } + \left [ { { { \bf l}_4}}\cdot ( { { { \bf l}_1}}+ { { { \bf l}_4 } } ) \right ] \left [ { { { \bf l}_3}}\cdot ( { { { \bf l}_2}}+{{{\bf l}_3 } } ) \right ] c^{\theta}_{|{{\bf l}}_1 + { { \bf l}}_4| } \big\ } \nonumber \\ & & \quad \quad + \ , { \rm perm . } \ ,
\label{eqn : trisecfinal}\end{aligned}\ ] ] note that the first two terms come from the first and second term in equation ( [ eqn : trisec ] ) , while the last two terms in above are from the third term . as before , through permutations , there are five additional terms involving the pairings of @xmath78 .
as written in equation ( [ eqn : trilens ] ) , the trispectrum formed by lensing alone depends on the power spectrum of deflection angle .
thus , with the right configuration for the cmb trispectrum , one can hope to extract some information related to lensing , mainly @xmath89 , given some knowledge on cmb anisotropies , such as @xmath90 .
note that a contribution to the cmb trispectrum due to lensing comes from @xmath89 when @xmath91 when @xmath92 and @xmath93 are the two sides of the quadrilateral that form the trispectrum . as illustrated in figure ( [ fig : config ] ) , this is equivalent to the diagonal of the quadrilateral and can be probed easily with a statistic that averages over the configurations for a given length of the diagonal . for this purpose ,
an useful option is to consider the configurations of the trispectrum that contribute to the power spectrum of squared temperatures . to see why such a statistic probe the diagonal of the trispectrum , consider the definition of this squared power spectrum @xmath94 where the fourier transform of the squared temperature
can be written as a convolution of the temperature transforms @xmath95 here , it should be understood that @xmath96 refers to the fourier transform of the square of the temperature rather than square of the fourier transform of temperature , @xmath97 ^ 2 $ ] .
note that the fourier transform of the temperature fluctuation , in the flat sky , is @xmath98 to compute the squared power spectrum of cmb temperature , we take @xmath99 since the lensing power spectrum contributes to the above squared power spectrum via the diagonals of the quadrilateral , the power spectrum of squared temperatures contain terms which are simply proportional to @xmath89 where @xmath100 is the multipole at which the squared power spectrum is measured . as discussed in @xcite , this then allows an extraction of the angular power spectrum of lensing deflections .
there is also a dominant noise contribution to the squared power spectrum .
this noise contribution comes from the lowest order correlations and is present even in the presence of gaussian temperature fluctuations ; thus , we denote it as the gaussian noise contribution . in the presence of the non - gaussian contribution resulting from the lensing effect
, we write these two contributions as @xmath101 with @xmath102 and the non - gaussianity captured by the four - point correlator as @xmath103 . following our previous discussion on the lensing trispectrum , we can now write the full four - point contribution to the squared power spectrum by replacing @xmath104 and @xmath105 in equation ( [ eqn : trilens ] ) with the required configuration from above such that @xmath106 ^ 2 \nonumber \\ & + & { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } { \int { d^2 { \bf l}_2 \over ( 2\pi)^2 } } \big\ { c_{|{{\bf
l}}_1+{{\bf l}}_2|}^{\phi\phi } [ { { \bf l}}_1 \cdot ( { { \bf l}}_1+{{\bf l}}_2 ) c^{\theta}_{l_1 } + { { \bf l}}_2 \cdot ( { { \bf l}}_1+{{\bf l}}_2 ) c^{\theta}_{l_2 } ] [ ( { { \bf l}}+{{\bf l}}_2 ) \cdot ( { { \bf l}}_1+{{\bf l}}_2 ) c^{\theta}_{|{{\bf l}}+{{\bf l}}_2| } - ( { { \bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}-{{\bf l}}_1 ) c^{\theta}_{|{{\bf l}}-{{\bf l}}_1| } ] \nonumber \\ & + & c_{|{{\bf l}}-{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } [ ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \cdot { { \bf l}}_2 c^{\theta}_{l_2 } + ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}-{{\bf l}}_1 ) c^{\theta}_{|{{\bf l}}-{{\bf l}}_1| } ] [ -({{\bf l}}-{{\bf l}}_1+{{\bf
l}}_2 ) \cdot { { \bf l}}_1 c^{\theta}_{l_1 } + ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}+{{\bf l}}_2 ) c^{\theta}_{|{{\bf l}}+{{\bf l}}_2| } ] \big\ } \ , .
\nonumber \\\end{aligned}\ ] ] as written , and discussed above , one of the terms in the non - gaussian contribution to the squared power spectrum is directly proportional to @xmath107 . to extract this angular power spectrum of lensing deflections ,
consider the approach suggested by @xcite involving the quadratic statistics formed by the divergence of the temperature - weighted gradients . to understand the mechanism involved ,
consider a filtered version of the squared power spectrum such that @xmath108 where @xmath109 is a filter that acts on the cmb temperature .
this filtering can be either in real space or fourier , or multipolar , space . for simplicity
, we have introduced filters that are applicable in fourier space . to extract the optimal filter in the presence of the gaussian noise alone , we ignore the additional noise contribution from the second and third terms of the trispectrum involving integrals over @xmath110 and @xmath111 .
we can , thus , write the filtered version of the squared power spectrum as @xmath112 \nonumber \\ & + & c_l^{\phi\phi } \left[\int \frac{d^2{{\bf l}}_1}{(2\pi)^2 } w({{\bf l } } , { { \bf l}}_1 ) [ { { \bf l}}\cdot { { \bf l}}_1 c^{\theta}_{l_1 } + { { \bf l}}\cdot ( { { \bf l}}-{{\bf l}}_1 ) c^{\theta}_{|{{\bf l}}-{{\bf l}}_1|}]\right ] \left[\int \frac{d^2{{\bf l}}_1}{(2\pi)^2 } w(-{{\bf l } } , -{{\bf l}}_1 ) [ { { \bf l}}\cdot { { \bf l}}_1 c^{\theta}_{l_1 } + { { \bf l}}\cdot ( { { \bf l}}-{{\bf l}}_1 ) c^{\theta}_{|{{\bf l}}-{{\bf l}}_1|}]\right ] . \nonumber \\\end{aligned}\ ] ] we assume that our filter has the symmetry property @xmath113 and then check to see if this assumption is justified . to be consistent with prior results ,
we also make one more assumption here .
we take the gaussian noise to be the dominant noise contribution and ignore the noise from the two extra terms in the non - gaussian trispectrum that involves integrals over @xmath114 and @xmath115 .
thus , the filter we will derive here is only optimal in the presence of the dominant gaussian noise and is sub - optimal when all noise terms are considered .
we will later add in these two terms as an extra source of noise , with the filter derived by ignoring them applied appropriately . with these assumptions
, we can write the signal - to - noise ratio for the detection of the @xmath28 power spectrum per each multipole moment as , @xmath116\right]^2 } { c_l^{\phi\phi } \left[\int \frac{d^2{{\bf l}}_1}{(2\pi)^2 } w({{\bf l } } , { { \bf l}}_1 ) [ { { \bf l}}\cdot { { \bf l}}_1 c^{\theta}_{l_1 } + { { \bf l}}\cdot ( { { \bf l}}-{{\bf l}}_1 ) c^{\theta}_{|{{\bf l}}-{{\bf l}}_1|}]\right]^2 + \int \frac{d^2{{\bf l}}_1}{(2\pi)^2 } w({{\bf l } } , { { \bf l}}_1)^2 2c_{l_1}^{{\rm t}}c_{|{{\bf l}}-{{\bf l}}_1|}^{{\rm t}}}\right)^2 \ , .\end{aligned}\ ] ] the optimal filter in the presence of gaussian noise is the one that will maximize this ratio . to extract the required filter , we take derivatives with respect to @xmath109 and find that the maximum is satisfied when @xmath117}{2c_{l_1}^{{\rm t}}c_{|{{\bf
l}}-{{\bf l}}_1|}^{{\rm t}}}\,.\ ] ] performing the explicit substitutions @xmath118 and @xmath119 , we see that our optimal filter does possess the desired symmetry .
we introduced this filter in ref .
@xcite when discussing the cmb@xmath120-sz power spectrum as a way to extract the lensing - sz correlation directly using the power spectrum of squared temperature and temperature .
it is not a surprise that the same filter appears now when extracting the lensing - lensing correlations using the power spectrum of squared temperature and squared temperature .
a physical mechanism to achieve this filter in real space is presented in ref .
@xcite by taking the gradient of the temperature , weighting it by a noise filtered cmb map , then taking the divergence of this product . with
the filter applied , we get a direct estimator of @xmath107 . introducing , @xmath121 ^ 2 } { 2c_{l_1}^{{\rm t}}c_{|{{\bf l}}-{{\bf l}}_1|}^{{\rm t } } } \nonumber \\ \label{eqn : nl}\end{aligned}\ ] ] the filtered squared power spectrum is such that @xmath122 where the gaussian noise associated with the extracted @xmath107 is @xmath123 .
we can now add in the noise contribution arising from the two extra terms in the non - gaussian trispectrum which we dropped when deriving the optimal filter in the presence of the gaussian noise only .
this extra noise contribution , neglected in prior studies , is @xmath124 [ ( { { \bf l}}+{{\bf l}}_2 ) \cdot ( { { \bf l}}_1+{{\bf l}}_2 ) c^{\theta}_{|{{\bf l}}+{{\bf l}}_2| } - ( { { \bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}-{{\bf l}}_1 ) c^{\theta}_{|{{\bf l}}-{{\bf l}}_1| } ] \nonumber \\ & + & c_{|{{\bf l}}-{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } [ ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \cdot { { \bf l}}_2 c^{\theta}_{l_2 } + ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}-{{\bf l}}_1 ) c^{\theta}_{|{{\bf l}}-{{\bf l}}_1| } ] [ -({{\bf
l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot { { \bf l}}_1 c^{\theta}_{l_1 } + ( { { \bf l}}-{{\bf l}}_1+{{\bf
l}}_2 ) \cdot ( { { \bf l}}+{{\bf l}}_2 ) c^{\theta}_{|{{\bf l}}+{{\bf l}}_2| } ] \big\ } \ , .
\nonumber \\\end{aligned}\ ] ] with the non - gaussian noise , we can write @xmath125 there is a strong possibility that the noise - bias @xmath123 associated with the gaussian contribution to the squared power spectrum can be removed during the lensing reconstruction .
this would involve the construction of a new estimator of lensing , or filtering scheme , that removes @xmath123 and leaves @xmath126 as the final noise contribution associated with the reconstructed @xmath107 .
the removal of the dominant gaussian noise bias is extremely useful since this allows one to lower the noise contribution associated with @xmath89 by at least an order of magnitude .
we will return later to the impact of such a noise - bias removal on the identification of an inflationary gravitational wave signature in the cmb in the form of a curl ( or b - mode ) contribution to the polarization .
in addition to the additional non - gaussian noise contribution from the lensing trispectrum itself , there is an additional noise contribution resulting from the relevant configuration of the trispectrum due to lensing - secondary correlations .
we do not reproduce this contribution here , but it can be easily calculated by replacing @xmath127 in equation ( [ eqn : trisecfinal ] ) , with @xmath128 . note that this additional noise results from the correlation between potentials that deflect cmb photons and the secondary contributions associated with the isw and sz thermal effects . in the presence of such contributions
, we also note that @xmath129 in equation ( [ eqn : nl ] ) , contains all contributions to noise , including the secondary contribution .
it is well known that certain secondary contributions can be removed based on distinct signatures , such as the spectral dependence of the sz effect @xcite . in such cases , we expect only the lensing - isw correlation to produce a noise contribution .
this effect is small and can be easily ignored . in the presence of the sz contribution
, however , there is a significant noise contribution that can potentially limit the extraction of the lensing power spectrum .
note that the filter is symmetric in @xmath130 and @xmath131 .
we show surfaces of this filter in figure ( [ fig : filter ] ) as a function of @xmath132 and @xmath133 for two values of @xmath100 corresponding to large and small angular scales respectively .
as shown , the filter behaves such that information related to lensing is extracted at arcminute scales via the damping tail of the cmb anisotropies .
additionally , since lensing smooths the acoustic peak structure , further information is also extracted in a subtle way from fluctuations associated with large angular scales .
thus , a simple filter which effectively cuts off the acoustic peak structure but extracts information from small angular scales , such a a step function in multipolar space , does poorly when compared to the filtering scheme proposed above @xcite .
as mentioned before deriving the filter , it should be noted that the proposed filter is only optimal to the extent that gaussian noise contribution is dominant . in the presence of noise contributions from additional four - point correlations or ,
more importantly , additional terms in the lensing trispectrum itself , the filter is sub - optimal .
we failed to find a simple analytical description for an optimal filter that also suppresses the non - gaussian noise contribution from the additional terms in the lensing trispectrum . as we find later , since these terms only contribute at an order of magnitude below the dominant gaussian noise , we do not expect this to be a significant problem in the extraction of the lensing information .
in addition to temperature , lensing also affects the cmb polarization .
we discuss here the best estimator from polarization observables for extracting the lensing contribution following the discussion in ref .
@xcite and provide relevant equations for other estimators in the appendix .
the statistic considered in this section involves the squared map constructed by convolving , in fourier space , the e- and b - mode maps of the polarization .
we denote this by @xmath134 . to understand this combination , we consider the fourier space analogue and , as before , write , @xmath135 \ , .\ ] ] now considering the power spectrum formed by @xmath136 , we write @xmath137 \left[e({{\bf l}}_2 ) b({{\bf l}}'-{{\bf l}}_2 ) + b({{\bf l}}_2 ) e({{\bf l}}'-{{\bf
l}}_2 ) \right ] \rangle \ , .\end{aligned}\ ] ] here , again , we find two sources of contributions , involving a noise contribution due to gaussian fluctuations in the e- and b - maps , @xmath138 \ , , \ ] ] and a non - gaussian contribution due to the lensing effect on polarization .
to calculate the lensing contribution to the polarization , we follow the notation in ref .
@xcite . introducing , @xmath139 where @xmath140 represent combinations of the stokes parameters .
the lensing effect is again to move the photon propagation directions on the sky . in fourier space
, we can consider the e- and b - mode decomposition introduced in ref .
@xcite such that @xmath141 and @xmath142 with this , we can write @xmath143 \phi({{\bf l}}-{{\bf l}}_1 ) ( { { \bf l}}-{{\bf l}}_1 ) \cdot { { \bf l}}_1 \nonumber \\
\tilde b({{\bf l } } ) & = & b({{\bf
l } } ) - { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } \left[e({{\bf l}}_1 ) \sin 2 ( \varphi_{{{\bf l}}_1 } -\varphi_{{\bf l } } ) + b({{\bf l}}_1 ) \cos 2 ( \varphi_{{{\bf l}}_1 } -\varphi_{{\bf l}})\right ] \phi({{\bf
l}}-{{\bf l}}_1 ) ( { { \bf l}}-{{\bf l}}_1 ) \cdot { { \bf l}}_1 \ , , \end{aligned}\ ] ] since primordial @xmath144 , due to the gravitational wave contribution , is small , we can make the useful approximation that @xmath145 \phi({{\bf l}}-{{\bf l}}_1 ) ( { { \bf l}}-{{\bf l}}_1 ) \cdot { { \bf l}}_1 \nonumber \\
\tilde b({{\bf l } } ) & = & - { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } \left[e({{\bf l}}_1 ) \sin 2 ( \varphi_{{{\bf l}}_1 } -\varphi_{{\bf l}})\right ] \phi({{\bf l}}-{{\bf
l}}_1 ) ( { { \bf l}}-{{\bf l}}_1 ) \cdot { { \bf l}}_1 \ , , \end{aligned}\ ] ] under this approximation , the lensed polarization power spectra can be expressed in terms of @xmath107 and the unlensed quantities following ref .
@xcite : @xmath146 \ , c_l^{ee } + \frac{1}{2 } { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } c_{|{{\bf l}}- { { \bf l}}_1|}^{\phi\phi } [ ( { { \bf l}}- { { \bf l}}_1)\cdot { { \bf l}}_1]^2 c^{ee}_{l_1 } [ 1 + \cos(4\varphi_{{{\bf l}}_1 } ) ] \ , , \nonumber \\ \tilde c_l^{bb } & = & \frac{1}{2 } { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } c_{|{{\bf l}}- { { \bf l}}_1|}^{\phi\phi } [ ( { { \bf l}}- { { \bf l}}_1)\cdot { { \bf l}}_1]^2 c^{ee}_{l_1 } [ 1 - \cos(4\varphi_{{{\bf l}}_1 } ) ] \ , .\nonumber \\\end{aligned}\ ] ] the total polarization power spectra is the sum of these lensed spectra and a noise contribution @xmath147 where the noise component is given by equation ( [ e : noise ] ) with a possibly different net @xmath68 for polarization . with the lensing effect on @xmath144 , we can now calculate the trispectrum contribution to the squared power spectrum as @xmath148 ^ 2 \nonumber \\ & + & { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } { \int { d^2 { \bf l}_2 \over ( 2\pi)^2 } } \big\{\frac{1}{4}c_{|{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \left[c_{l_1}^{ee } c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}+{{\bf l}}_2 ) ( { { \bf l}}_1+{{\bf l}}_2 ) \cdot { { \bf l}}_1 - c_{l_2}^{ee } c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}_1+{{\bf l}}_2 ) \cdot { { \bf l}}_2 ( { { \bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}-{{\bf l}}_1 ) \right ] \nonumber \\ & & \quad \quad \times \sin 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \sin 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & \frac{1}{4}c_{|{{\bf l}}-{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}-{{\bf l}}_1+{{\bf
l}}_2 ) \cdot ( { { \bf l}}+{{\bf l}}_2 ) ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}-{{\bf l}}_1 ) - c_{l_1}^{ee } c_{l_2}^{ee }
( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot { { \bf l}}_1 ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot { { \bf l}}_2 \right ] \nonumber \\ & & \quad \quad \times \sin 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \sin 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \big\}\ , , \end{aligned}\ ] ] to extract the lensing contribution , one can again consider a filtering scheme , and following arguments similar to the case with temperature , the derived filter that maximizes the signal - to - noise in the presence of gaussian noise is @xmath149\sin 2 ( \varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}-{{\bf l}}_1 } ) } { c_{l_1}^{ee { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{bb { { \rm t } } } + c_{l_1}^{bb { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{ee { { \rm t } } } } \ , .\ ] ] with this filter applied , the associated noise due to the dominant gaussian part is @xmath123 , where , @xmath150 ^ 2 \sin^2 2 ( \varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}-{{\bf l}}_1 } ) } { 2c_{l_1}^{ee { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{bb { { \rm t } } } + 2c_{l_1}^{bb { { \rm t } } } c_{|{{\bf l}}-{{\bf
l}}_1|}^{ee { { \rm t } } } } \ , .\end{aligned}\ ] ] the extra noise due to the additional terms in the trispectrum is @xmath151 \nonumber \\ & & \quad \quad \times \sin 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \sin 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & \frac{1}{4}c_{|{{\bf l}}-{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}+{{\bf l}}_2 ) ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot ( { { \bf l}}-{{\bf l}}_1 ) - c_{l_1}^{ee } c_{l_2}^{ee } ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot { { \bf l}}_1 ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \cdot { { \bf l}}_2 \right ] \nonumber \\ & & \quad \quad \times \sin 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \sin 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2})\big\ } \ , .\end{aligned}\ ] ]
we have introduced the trispectrum , or four - point correlation function , of cmb anisotropies as a probe of the large scale structure weak lensing effect .
the trispectrum in cmb anisotropies is generated by the non - linear mode coupling behavior of the weak lensing effect on the cmb .
as written in equation ( [ eqn : trilens ] ) , the trispectrum formed by lensing contains a significant contribution from @xmath89 when @xmath152 .
@xmath130 and @xmath153 are the two sides of the quadrilateral that formed the trispectrum in the fourier space as illustrated in figure [ fig : config ] .
contributions thus probe the lensing effect via the diagonal of the trispectrum configuration , and statistics sensitive to the diagonal will provide a practical approach for extracting the lensing angular power spectrum .
a useful statistic for this purpose is the power spectrum formed by squared temperatures introduced earlier .
this approach expands the suggestion in ref .
@xcite to use the squared temperature - temperature power spectrum to study the cmb bispectrum and a filtered version to extract the power spectrum associated with lensing - sz correlation .
the power spectrum formed by squared temperatures probes the trispectrum .
this power spectrum , however , contains a dominant contribution due to the gaussian fluctuations associated with temperature which potentially dominates the contribution arising from the non - gaussian weak lensing effect . to optimally extract the lensing contribution in the presence of this dominant gaussian noise , we have introduced a filter in fourier space .
this filter requires knowledge of the noise contributions to the map as well as the power spectra of prelensed primary anisotropies at the last scattering surface .
though the filter can be easily applied to the fourier transform of the squared temperature map in the fourier space , a useful direct approach is presented in ref .
this real space construction involves the divergence of the temperature - weighted temperature gradients , with each of these temperature estimators first filtered for the excess noise .
we have extended the previous discussions of the extraction of lensing signal to include a discussion of the full trispectrum .
the full trispectrum contains terms which were previously ignored and these terms contribute as additional sources of noise in the lensing reconstruction and can potentially interfere with the extraction of the lensing signal from cmb data .
these additional non - gaussian noise terms correspond to the lensing contribution associated with the second diagonal of the trispectrum configuration shown in figure ( [ fig : config ] ) and the same diagonal under a permutation , while keeping the same vector @xmath154 for the diagonal from which lensing information is extracted . in figure ( [ fig
: tempnoise ] ) , we summarize our results related to the lensing extraction using temperature data alone .
we show the dominant gaussian noise contribution and the associated non - gaussian noise contribution , due to the two additional terms in the trispectrum which involve integrals over @xmath114 and @xmath115 .
we show these noise contributions for temperature maps constructed with an effective detector sensitivity of 1 @xmath155k @xmath156 and varying resolution with beam widths of 5 to 1 arcmin . as shown , non - gaussian noise contributions from additional terms in the trispectrum are roughly an order of magnitude below the dominant gaussian noise contribution .
this conclusion is subject to several conditions . as summarized in figure ( [ fig : sznoise ] ) , the extraction from temperature alone can be significantly impaired by the presence of foregrounds that correlate with the lensing deflection angle .
for example , the sz contribution hinders the separation in two ways .
the first is an enhancement of the gaussian noise signal by the addition of @xmath157 to @xmath129 as indicated by the difference between top dashed and the dot - dashed lines .
the second is a noise contribution due to the non - gaussian trispectrum formed by the correlation of lensing potentials with the sz effect due to the large scale structure .
we show this with the long - dashed line . in the presence of sz
, it is now clear that the lensing extraction is strongly limited .
the other source of correlation is the isw effect .
its effects , through the contribution to gaussian noise , are already included in the @xmath158 defined earlier .
the power spectra calculated using the publicly available cmbfast code @xcite contain the isw signal as well .
the noise due to the correlation between isw and lensing is significantly smaller and is at a level below @xmath159 in figure ( [ fig : sznoise ] ) .
thus , one can easily ignore the additional noise contributions resulting from the isw effect .
furthermore , the sz noise contributions can also be controlled .
one can easily separate the sz contribution from thermal cmb fluctuations based on the frequency spectrum of the sz effect @xcite . in the upcoming planck mission ,
sz separation based on multifrequency data will prevent this contribution from adding significant noise to cmb maps .
we suggest that small angular scales experiments seeking to extract the lensing signal based on temperature anisotropies be designed to contain channels that are optimized for a separation of the sz signal from thermal cmb fluctuations .
in addition to temperature , lensing also modifies polarization data .
thus , one can employ estimators based on polarization , and combinations of polarization and temperature , to extract information related to lensing . in figure ( [ fig : ebnoise ] ) , we illustrate the usefulness of polarization observations for a lensing separation .
the dotted lines are the errors associated with the extraction based on the eb estimator .
this estimator was suggested by @xcite as an estimator that probes the angular power spectrum of lensing on scales smaller than that probed by temperature or other combinations of temperature and polarization .
ee and @xmath160e estimators also probe the lensing deflection power spectrum , and have gaussian and non - gaussian noise contribution in between those of the @xmath161 and eb estimators . while we presented a detailed derivation of the eb estimator
, we write down the trispectrum and related filters for other estimators associated with polarization and combinations of polarization and temperature in the appendix . as shown in figure ( [ fig : ebnoise ] ) , we again find the non - gaussian noise contributions due to additional noise terms in various trispectra due to lensing alone to be roughly an order of magnitude below the dominant gaussian noise contribution .
similar to the @xmath161 estimator , the @xmath160e estimator contains an additional noise term due to the correlation between weak lensing and secondary anisotropies such as the sz effect .
we have presented a derivation of this contribution in the appendix . in both these cases , note that we have only considered the secondary contributions to the temperature .
unlike contributions to the temperature , secondary contributions to the polarization are significantly smaller and are higher order in density fluctuations ; for example , in addition to the density field , these secondary effects also depend on the velocity field @xcite .
this makes the secondary polarization - lensing correlations less significant , similar to the correlation between lensing potentials and the moving baryons responsible for the ostriker - vishniac / kinetic sz effect @xcite .
therefore , we ignore noise contributions from secondary effects associated with lensing extraction based on polarization estimators alone . in general , the cmb data allow high signal - to - noise ratio reconstruction of the lensing effect .
we illustrate the extent to which planck s frequency cleaned temperature maps can be used as a lensing experiment in figure ( [ fig : convergence ] ) . here
, we show the angular power spectrum of convergence for sources at the last scattering , ie . ,
cmb , and at a redshift of 1 ; the latter is consistent with what is expected from weak lensing surveys via galaxy ellipticity data . in the right panel
, we show the cumulative signal - to - noise ratio for planck based on the quadratic statistic discussed here and the gradient - based statistic of @xcite .
the reconstruction based on the squared power spectrum has factor of @xmath0 3 higher cumulative signal - to - noise than the one based on gradients in the temperature map .
this can be understood by noting that the latter method only extracts lensing from the first order expansion of the temperature in terms of the deflection angle , while significant contributions also come from the second order term that is also probed by the proposed estimators . as shown in figure ( [ fig : convergence ] ) , due to planck s beam of @xmath0 5 to 10 arcmin the reconstruction of convergence becomes noise dominated at multipoles of @xmath0 1000 , while one can probe lensing to much smaller angular scales with galaxy information .
the difference between the two curves illustrate the importance of lensing reconstruction for the polarization studies related to gravitational waves .
since cmb photons are affected by potentials out to the last scattering surface , the resulting lensing effect is at least an order of magnitude higher than the lensing effect on sources that are visible from the large scale structure out to a redshift of a few .
thus , it is unlikely that observations of the large scale structure lensing can be used to subtract the lensing contribution to the b - mode polarization sufficiently to permit a confusion free detection of the polarization signal due to gravitational waves .
we refer the reader to @xcite for a further discussion on this application .
we can extend the discussion in @xcite to consider one interesting possibility . as discussed earlier
, one can make a significant improvement in the lensing reconstruction when the gaussian noise bias associated with the temperature squared power spectrum is removed .
in such a scenario , the noise contribution associated with the lensing extraction is determined by the non - gaussianities which are usually an order of magnitude below the dominant gaussian noise contribution .
such a gaussian noise bias removal allows an extremely high signal - to - noise extraction of the lensing signal even for experiments with poor sensitivities that are not optimized for the lensing reconstruction . as shown in figure
( [ fig : convergence ] ) , for planck , the removal of the dominant gaussian noise leads to detection of the lensing angular power spectrum with a cumulative signal - to - noise ratio of @xmath0 700 , which is roughly a factor of 20 above the level expected when the dominant gaussian noise contribution is not removed .
such an extraction not only improves the effectiveness of cosmological studies involving lensing of the cmb @xcite , but also allows a high signal - to - noise separation of the lensed e - mode contribution to the b - mode polarization map of the cmb .
this improves the confusion - free limit on the inflationary energy scale attainable with cmb polarization experiments by another order of magnitude @xcite .
following @xcite , we illustrate the importance of noise - bias removal in figure ( [ fig : inflation ] ) . here , we compare various contributions to the power spectra of cmb polarization , including the lensed e - mode contribution to the b - mode polarization map .
we also show the residual lensing contribution to the b - mode map after the lensing effect has been largely removed using an estimator for lensing deflections .
the estimator employs a cmb temperature map with a resolution of 1 arcmin and a sensitivity of 1 @xmath155k @xmath156 .
the curve labeled @xmath162 is the residual contribution when the lensing extraction is carried out to the noise level determined by the non - gaussian noise contribution , assuming that the dominant gaussian noise contribution can be removed during the lensing extraction .
as shown , such a removal provides an order of magnitude improvement over the residual from the dominant gaussian contribution . in the right panel of figure ( [ fig : inflation ] ) , we show limits on the inflationary energy scale based on lensing extraction with the remaining residuals depicted in the left panel of the same figure .
we refer the reader to @xcite for details related to this calculation . for comparison
, we also show the inflationary energy scale limit when the lensing contribution is ignored ( only instrumental noise ) and the limit when the lensing contribution is treated as an additional contribution to the noise associated with the b - mode map .
as illustrated in these figures , the potential removal of the gaussian noise associated with the lensing reconstruction reduces by an order of magnitude the confusion - free upper bound on the gravitational wave signal in the b - mode polarization .
such an improvement motivates the development of practical schemes to remove the noise - bias associated with the dominant gaussian part of the temperature squared power spectrum . in future work , we will return to this issue and will investigate ways to achieve the required level of noise removal .
we have discussed the four - point correlation function , or the trispectrum in fourier space , of cmb temperature and polarization anisotropies due to the weak gravitational lensing effect by intervening large scale structure .
the trispectrum contains information related to the lensing deflection angle , which for a given configuration of the trispectrum is captured primarily by the diagonal of the quadrilateral that defines the trispectrum in fourier space .
we have devised a statistic involving the power spectrum of squared temperatures that captures this information from the diagonal .
this provides an observational approach to extracting the power spectrum of deflection angles induced by the weak gravitational lensing effect on the cmb .
the method extends previous suggestions on the use of the squared temperature power spectrum to extract the lensing - secondary correlation power spectrum by probing the bispectrum formed by the lensing effect .
the technique presented here is not new and has been discussed previously in the literature by @xcite .
we extend this work by providing a full derivation of the trispectrum terms both due to lensing and lensing - secondary correlations , and by providing a detailed discussion of the filtering required to minimize the dominant noise component due to gaussian temperature fluctuations that impairs the extraction of lensing information .
we also extend previous discussions on the trispectrum and associated weak lensing reconstruction from cmb data by calculating in detail contributions to the noise in the reconstruction .
previous work discussed the dominant gaussian noise , but we include in addition terms associated with non - gaussian contributions due to lensing alone and the correlation of the lensing effect with other foreground secondary anisotropies in the cmb such as the isw effect and the sz effect .
when the sz effect is removed from the temperature maps using its spectral dependence , we find these additional non - gaussian noise contributions to be an order of magnitude lower than the dominant gaussian noise .
when sz is not removed , the extraction of lensing information is significantly suppressed ; we strongly advise the use of multiple frequencies in upcoming small angular scale cmb experiments , if there is an interest in using data from such experiments to conduct a weak lensing study .
the lensing extraction provides a high signal - to - noise ratio detection of the lensing deflection power spectrum . for the planck surveyor , with frequency - cleaned temperature maps
, detection will have a cumulative signal - to - noise of @xmath0 35 , while certain techniques in the literature @xcite , involving the gradients of the temperature map , only give a cumulative signal - to - noise ratio of @xmath0 10 .
since the extraction with planck is severely limited by its beam of 5 to 10 arcmin , we strongly recommend higher resolution observations at the arcminute scale .
if the noise - bias associated with the dominant gaussian part of the temperature squared power spectrum is removed , planck can extract the lensing power spectrum with temperature data alone with a cumulative signal - to - noise ratio around @xmath0 700 .
with information from polarization , the signal - to - noise ratio associated with lensing can be further increased by another factor of a few .
the order of magnitude improvement in the lensing separation motivates the development of practical methods that can remove the dominant gaussian noise contribution to the squared temperature power spectrum .
in general , our study suggests that future cmb observations can be used as a weak lensing experiment akin to current attempts involving lensing reconstruction from galaxy ellipticity data .
this opportunity will help to achieve an ultimate challenge for cmb experiments , a confusion - free direct detection of the gravitational wave signature from inflation .
we thank wayne hu and marc kamionkowski for useful discussions .
this work was supported in part by nsf ast-0096023 , nasa nag5 - 8506 , and doe de - fg03 - 92-er40701 .
kesden acknowledges the support of an nsf graduate fellowship and ac acknowledges support from the sherman fairchild foundation .
here we provide the appropriate formulae for deriving optimal filters for the ee and @xmath163e estimators of @xmath107 .
the fourier transform of the squared e - mode map can be expressed as a convolution of fourier transforms of the e - mode map itself , analogous to equation ( [ eqn : cmbsqr ] ) of [ sec : covariance ] . @xmath164 defining the power spectrum for the squared e - mode map as @xmath165 , we write @xmath166 as before , this power spectrum will consist of a noise component due to gaussian fluctuations in the e - mode map and a non - gaussian contribution due to lensing given respectively by , @xmath167 @xmath168 ^ 2 \nonumber \\ & + & { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } { \int { d^2 { \bf l}_2 \over ( 2\pi)^2 } } \big\ { c_{|{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \left[c_{l_1}^{ee } { { \bf l}}_1 + c_{l_2}^{ee } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \nonumber \\ & & \quad \quad \times \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & c_{|{{\bf l}}-{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{ee } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{ee } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \nonumber \\ & & \quad \quad \times \cos 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \big\ } \ , .\end{aligned}\ ] ] in an analysis similar to that of [ sec : covariance ] and [ sec : ebtri ] , we find that the optimal filter in the presence of the gaussian noise term above is @xmath149\cos 2 ( \varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}-{{\bf l}}_1 } ) } { 2c_{l_1}^{ee { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{ee { { \rm t } } } } \ , .\ ] ] using this filter , the dominant gaussian noise @xmath123 is given by @xmath150 ^ 2 \cos^2 2 ( \varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}-{{\bf l}}_1 } ) } { 2c_{l_1}^{ee { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{ee { { \rm t } } } } \ , , \end{aligned}\ ] ] while the excess noise due to the additional terms in the e - mode trispectrum is @xmath169 \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \nonumber \\ & & \quad \quad \times \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & c_{|{{\bf l}}-{{\bf
l}}_1+{{\bf l}}_2|}^{\phi\phi } \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{ee } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{ee } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \nonumber \\ & & \quad \quad \times \cos 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \big\ } \ , .\end{aligned}\ ] ] here we derive the optimal filter and accompanying noise for the @xmath163e estimator of @xmath107 .
the following equations ( [ e : cmbedef ] ) through ( [ e : nlcmbeng ] ) for this estimator are entirely analogous to equations ( [ e : eedef ] ) through ( [ e : nleeng ] ) of the preceding subsection of this appendix .
an additional complication here is that , unlike in the @xmath170 estimator where @xmath171 and @xmath172 are uncorrelated due to parity , @xmath173 and @xmath171 are correlated with a power spectrum defined by @xmath174 below .
this produces an additional term through the lensing effect in addition to that involving correlations between @xmath173 and @xmath171 separately .
first , we expand in fourier space the product of @xmath173 and @xmath171 maps , @xmath175 \ , , \ ] ] to obtain @xmath176 \left[{\theta}({{\bf l}}_2 ) e({{\bf l}}'-{{\bf l}}_2 ) + e({{\bf l}}_2 ) { \theta}({{\bf l}}'-{{\bf l}}_2 ) \right ] \rangle \ , , \end{aligned}\ ] ] and @xmath177 \ , .\ ] ] the non - gaussian part is @xmath178 \left[1 + \cos 2 ( \varphi_{{{\bf l}}_1}+ \varphi_{{{\bf l}}-{{\bf l}}_1 } ) \right ] \big\}^2 \nonumber \\ & + & \frac{1}{4 } { \int { d^2 { \bf l}_1 \over ( 2\pi)^2 } } { \int { d^2 { \bf l}_2 \over ( 2\pi)^2 } } \big\ { c_{|{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \big\ { \left[c_{l_1}^{{\theta } } { { \bf l}}_1 + c_{l_2}^{{\theta } } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \nonumber \\ & + & \left[c_{l_1}^{ee } { { \bf l}}_1 + c_{l_2}^{ee } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta } } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta } } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & \left[c_{l_1}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) { { \bf l}}_1 + c_{l_2}^{{\theta}e } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \nonumber \\ & + & \left[c_{l_1}^{{\theta}e } { { \bf l}}_1 + c_{l_2}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \big\ } \nonumber \\ & + & c_{|{{\bf
l}}-{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \big\ { \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf
l}}_2 ) \nonumber \\ & + & \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{{\theta}e } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf
l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{{\theta}e } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf
l}}_2 ) \nonumber \\ & + & \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{ee } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta } } ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{{\theta } } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta } } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{{\theta } } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{ee } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf
l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}+{{\bf l}}_2 } -\varphi_{{{\bf l}}_1 } ) \big\ } \big\ } \ , .\nonumber \\\end{aligned}\ ] ] the filter required to suppress the dominant gaussian noise and extract lensing information is @xmath179 \left[1 + \cos 2 ( \varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}-{{\bf l}}_1 } ) \right]}{c_{l_1}^{{\theta}{{\rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{ee { { \rm t } } } + c_{l_1}^{ee { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}{{\rm t } } } + 2c_{l_1}^{{\theta}e { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e { { \rm t } } } } \ , , \ ] ] and the associated gaussian noise is @xmath180 ^ 2 \left[1 + \cos 2 ( \varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}-{{\bf l}}_1 } ) \right]^2}{2c_{l_1}^{{\theta}{{\rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{ee { { \rm t } } } + 2c_{l_1}^{ee { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}{{\rm t } } } + 4c_{l_1}^{{\theta}e { { \rm t } } } c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e { { \rm t } } } } \ , , \ ] ] while the non - gaussian noise is @xmath181 \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) \nonumber \\ & + & \left[c_{l_1}^{ee } { { \bf l}}_1 + c_{l_2}^{ee } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta } } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta } } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & \left[c_{l_1}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) { { \bf l}}_1 + c_{l_2}^{{\theta}e } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \nonumber \\ & + & \left[c_{l_1}^{{\theta}e } { { \bf l}}_1 + c_{l_2}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) { { \bf l}}_2 \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } ( { { \bf l}}+{{\bf l}}_2 ) - c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) ( { { \bf l}}-{{\bf l}}_1 ) \right ] \cdot ( { { \bf l}}_1 + { { \bf l}}_2 ) \big\ } \nonumber \\ & + & c_{|{{\bf
l}}-{{\bf l}}_1+{{\bf l}}_2|}^{\phi\phi } \big\ { \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf
l}}_1+{{\bf l}}_2 ) \nonumber \\ & & \quad \quad \times \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } ( { { \bf l}}+{{\bf
l}}_2 ) - c_{l_1}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \nonumber \\ & + & \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{{\theta}e } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta}e } \cos 2(\varphi_{{{\bf l}}_1 } -\varphi_{{{\bf l}}+{{\bf l}}_2 } ) ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{{\theta}e } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1+{{\bf l}}_2 ) \nonumber \\ & + & \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{ee } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{ee } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{{\theta } } ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{{\theta } } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}-{{\bf l}}_1 } + \varphi_{{{\bf l}}_2 } ) \nonumber \\ & + & \left[c_{|{{\bf l}}-{{\bf l}}_1|}^{{\theta } } ( { { \bf l}}-{{\bf l}}_1 ) + c_{l_2}^{{\theta } } { { \bf l}}_2 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \left[c_{|{{\bf l}}+{{\bf l}}_2|}^{ee } ( { { \bf l}}+{{\bf l}}_2 ) - c_{l_1}^{ee } { { \bf l}}_1 \right ] \cdot ( { { \bf l}}-{{\bf l}}_1 + { { \bf l}}_2 ) \cos 2(\varphi_{{{\bf l}}+{{\bf l}}_2 } -\varphi_{{{\bf l}}_1 } ) \big\ } \big\ } \ , . \nonumber \\\end{aligned}\ ] ] the trispectrum formed from two temperature and two e - mode terms will contain an additional non - gaussian contribution through the coupling of the lensing deflection angle to secondary effects .
a similar contribution occurs in the temperature trispectrum and was discussed at the end of [ sec : lensing ] . as in equation ( [ eqn : trisec ] ) , cumulants such as
@xmath182 will vanish as the secondary effects are decoupled from the primary fluctuations generated at the surface of last scatter .
however , contributions come from the correlation between @xmath86 and the lensing deflection @xmath28 . as before ,
contributions of equal importance come from both the first and second order terms in @xmath73 written in equation ( [ eqn : lfactor ] ) .
we consider this additional non - gaussian contribution for the connected portion of a four - point function in fourier space @xmath183 contributions to the trispectrum from the first two terms come through the second order term in @xmath73 , with the two @xmath28 terms coupling to @xmath86 . in the last term , contributions come from the first order term of @xmath73 similar to the pure lensing contribution to the trispectrum . after some straightforward simplifications ,
we can write the connected part of the trispectrum involving lensing - secondary coupling . we symmetrize with respect to the four indices by adding five terms corresponding to the additional permutations of @xmath184 and multiplying by a factor of 1/6 .
@xmath185 \left [ { { { \bf l}_4}}\cdot ( { { { \bf l}_2}}+{{{\bf l}_4 } } ) \right ] c^{\theta}_{|{{\bf l}}_1 + { { \bf l}}_3| } \cos 2(\varphi_{{{\bf l}}_1 + { { \bf l}}_3 } - \varphi_{{{\bf l}}_1 } ) \cos 2(\varphi_{{{\bf l}}_2 + { { \bf l}}_4 } - \varphi_{{{\bf l}}_2 } ) \nonumber \\ & & \quad \quad + \left [ { { { \bf l}_4}}\cdot ( { { { \bf l}_1}}+ { { { \bf l}_4 } } ) \right ] \left [ { { { \bf l}_3}}\cdot ( { { { \bf l}_2}}+{{{\bf l}_3 } } ) \right ] c^{\theta}_{|{{\bf l}}_1 + { { \bf l}}_4| } \cos 2(\varphi_{{{\bf l}}_1 + { { \bf l}}_4 } - \varphi_{{{\bf l}}_1 } ) \cos
2(\varphi_{{{\bf l}}_2 + { { \bf l}}_3 } - \varphi_{{{\bf l}}_2 } ) + \ , { \rm perm . } \big\ } \nonumber \\\end{aligned}\ ] ] note that the first two terms come from the first and second term in equation ( [ eqn : tetrisec ] ) , while the last two terms are from the third term .
99 p. j. e. peebles and j. t. yu , astrophys .
j. , * 162 * , 815 ( 1970 ) ; r. a. sunyaev and ya.b .
zeldovich , astrophys
. space sci . * 7 * , 3 ( 1970 ) ; j. silk , astrophys .
j. , * 151 * , 459 ( 1968 ) ; see recent review by w. hu & s. dodelson , ann .
, in press , astro - ph/0110414 ( 2001 ) .
g. jungman , m. kamionkowski , a. kosowsky and d.n .
spergel , phys .
d , * 54 * 1332 ( 1995 ) ; j.r . bond , g. efstathiou and m. tegmark , mnras , * 291 * l33 ( 1997 ) ; m. zaldarriaga , d.n .
spergel and u. seljak , astrophys . j. , * 488 * 1 ( 1997 ) ; d.j .
eisenstein , w. hu and m. tegmark , astrophys .
j. , * 518 * 2 ( 1999 ) a. d. miller et al . ,
j. lett .
* 524 * , l1 ( 1999 ) ; p. de bernardis et al . ,
nature * 404 * , 955 ( 2000 ) ; s. hanany et al .
, astrophys .
j. lett .
* 545 * , l5 ( 2000 ) ; n. w. halverson et al .
, astro - ph/0104489 .
see , for example , review by a. cooray , 2002 , in `` 2001 coral gables conference : cosmology and particle physics '' , eds .
b. kursunoglu & a. perlmutter ; american institute of physics conference proceedings , astro - ph/0203048 a. blanchard and j. schneider , a&a , 184 , 1 ( 1987 ) ; a. kashlinsky , astrophys . j. , 331 , l1 ( 1988 ) ; e. v. linder , a&a , 206 , 1999 , ( 1988 ) ; s. cole and g. efstathiou , mnras , 239 , 195 ( 1989 ) ; m. sasaki , mnras , 240 , 415 ( 1989 ) ; k. watanabe and k. tomita , astrophys . j. , 370 , 481 ( 1991 ) ; m. fukugita , t. futumase , m. kasai and e. l. turner , astrophys . j. , 393 , 3 ( 1992 ) ; l. cayon , e. martinez - gonzalez and j. sanz , astrophys .
j. , 413 , 10 ( 1993 ) ; u. seljak , astrophys .
j. , 463 , 1 ( 1996 ) d. n. spergel and d. m. goldberg , phys .
d*59 * , 103001 ( 1999 ) ; d. m. goldberg and d. n. spergel , phys .
d , * 59 * , 103002 ( 1999 ) ;
a. cooray and w. hu , astrophys . j. , * 534 * , 533 ( 2000 ) .
m. zaldarriaga and u. seljak , phys .
d , * 59 * , 123507 ( 1999 ) ; h. v. peiris and d. n. spergel , astrophys .
j. , * 540 * , 605 ( 2000 ) .
j. p. ostriker and e. t. vishniac , astrophys . j. , * 306 * , l51 ( 1986 ) ; e. t. vishniac , astrophys . j. , * 322 * , 597 ( 1987 ) ; s. dodelson and j. m. jubas , astrophys .
j. , * 439 * , 503 ( 1995 ) ; g. efstathiou , in large scale motions in the universe . a vatican study week , ed .
v. c. rubin & g. v. coyne ( princeton : princeton university press ) , 299 ( 1988 ) ; a. h. jaffe and m. kamionkowski , phys .
d , * 58 * , 043001 ( 1998 ) ; | we discuss the four - point correlation function , or the trispectrum in fourier space , of cmb temperature and polarization anisotropies due to the weak gravitational lensing effect by intervening large scale structure .
we discuss the squared temperature power spectrum as a probe of this trispectrum and , more importantly , as an observational approach to extracting the power spectrum of the deflection angle associated with the weak gravitational lensing effect on the cmb .
we extend previous discussions on the trispectrum and associated weak lensing reconstruction from cmb data by calculating non - gaussian noise contributions , beyond the previously discussed dominant gaussian noise .
non - gaussian noise contributions are generated by lensing itself and by the correlation between the lensing effect and other foreground secondary anisotropies in the cmb such as the sunyaev - zeldovich ( sz ) effect .
when the sz effect is removed from temperature maps using its spectral dependence , we find these additional non - gaussian noise contributions to be an order of magnitude lower than the dominant gaussian noise . if the noise - bias due to the dominant gaussian part of the temperature squared power spectrum is removed , then these additional non - gaussian contributions provide the limiting noise level for the lensing reconstruction .
the temperature squared power spectrum allows a high signal - to - noise extraction of the lensing deflections and a confusion - free separation of the curl ( or b - mode ) polarization due to inflationary gravitational waves from that due to lensed gradient ( or e - mode ) polarization .
the small angular scale temperature and polarization anisotropy measurements provide a novel approach to weak lensing studies , complementing the approach based on galaxy ellipticities . |
near a phase - transition critical point , some observed quantities show a universal behavior that is common to a large class of systems , independently of the microscopic details .
a very important universality class is the ising one that is characterized by short - range interactions and a scalar order parameter .
it describes the liquid - vapor transition in fluids , the mixing transition in multicomponent systems , the curie transition in ( anti)ferromagnets with axial anisotropy .
the ising critical behavior has been extensively studied both theoretically and experimentally , see refs .
in particular , the critical exponents , the equation of state , and several amplitude ratios have been determined with good precision .
another important quantity in the theory of critical phenomena is the static structure factor , that can be measured experimentally by determining the intensity of the light scattered by the fluid relative to the intensity of the incident light @xcite . to probe larger wave numbers ,
neutrons are used instead of light . at the critical density of fluids near the gas - liquid critical point or at the critical concentration of binary fluids near the critical mixing point ,
one expects for @xmath2 the general scaling behavior@xcite @xmath3 where @xmath4 , @xmath5 is the correlation length which diverges as @xmath6 , @xmath7 is the momentum - transfer vector , and @xmath8 refers to the two phases , @xmath9 ( resp .
@xmath10 ) corresponding to the high- ( resp .
low- ) temperature phase .
its absolute value is given by @xmath11 where @xmath12 is the wavelength of the radiation ( neutrons ) in the scattering medium and @xmath13 is the scattering angle .
the functions @xmath14 , normalized so that @xmath15 for @xmath16 ( this defines @xmath5 as the second - moment correlation length ) , are universal .
their limiting behavior is well known . for @xmath17 small ,
@xmath14 is approximated by the leading term , the so - called ornstein - zernicke approximation @xmath18 such an approximation well describes the data up to @xmath19 and is routinely used in the analysis of the data with @xmath20 small and of the turbidity for the determination of the correlation length @xcite . on the other hand , for large @xmath17
, @xmath14 shows an anomalous decay controlled by the exponent @xmath21 @xmath22 therefore , the experimental determination of the structure factor for large wavenumbers allows a direct determination of the exponent @xmath21 @xcite . in this paper , we compute the structure factor in the high - temperature phase for small values of @xmath17 by means of monte carlo simulations on lattices @xmath0 , with @xmath23 .
we are able to determine the function @xmath24 with an error of less than 1% ( resp .
2% ) for @xmath25 ( resp .
@xmath26 ) . these numerical results
together with the most recent estimates of the critical exponents @xcite are then used to determine interpolations that are valid for all values of @xmath17 and have the correct large-@xmath17 behavior . for this purpose
, we use a dispersive approach @xcite , which allows us to determine an interpolating form for @xmath24 that agrees with the monte carlo data in the small-@xmath17 region and that well approximates ( within 0.5% ) the experimental results of ref .
@xcite .
these results are then used to compute the turbidity , i.e. the attenuation of the transmitted light intensity per unit optical path length due to the scattering with the sample .
this quantity is routinely measured in experiments , since it allows the determination of the correlation length .
in particular , we compute the deviations from the puglielli - ford expression @xcite , that is based on the ornstein - zernicke approximation . the paper is organized as follows . in sec .
[ sec.2 ] we review the theoretical results for the structure factor . in sec .
[ sec.2.1 ] we define the basic observables and report the behavior of @xmath14 for small and large values of @xmath17 .
estimates of the constants appearing in these expansions are reported in sec .
[ sec.2.2 ] . in sec .
[ sec.2.3 ] we discuss bray s approximation .
first , we discuss the high - temperature phase : we update the estimates of ref .
@xcite by using the most recent results for the critical exponents .
then , we generalize the approximation to the low - temperature phase . in sec .
[ sec.3 ] we discuss our high - temperature monte carlo results which are compared with approximate expressions and with the experimental data of ref . @xcite . in sec .
[ sec.4 ] we compute the turbidity , focusing on the deviations from the puglielli - ford expression @xcite due to the anomalous decay of @xmath24 .
we find that the turbidity is larger than this expression by 1% ( resp .
5% ) for @xmath27 ( resp .
350 ) , where @xmath28 and @xmath29 is the momentum of the incoming radiation .
several theoretical results are available for the structure factor . for @xmath17 small
, one can compute the corrections to the ornstein - zernicke behavior , by writing @xmath30 for large @xmath17 , the structure factor behaves as @xmath31 a behavior predicted theoretically by fisher and langer @xcite and proved in the field - theoretical framework in refs .
@xcite . .
estimates of @xmath32 , @xmath33 , and @xmath34 .
iht denotes the results obtained from the analysis of high - temperature expansions for improved models , ht , lt results obtained from the analysis of high- and low - temperature expansions for the ising model , while @xmath35-exp . " and
@xmath36 @xmath37-exp .
" label the field - theoretical results .
( sc ) and ( bcc ) denote the simple cubic and the body - centered cubic lattice respectively .
unless stated otherwise , field - theoretical results are taken from ref .
@xcite , while the iht estimates are taken from ref . @xcite . for @xmath38 we should also report the monte carlo estimate of ref.@xcite , @xmath39 .
[ cols="^ , > , < , > , < , > , < , > , < " , ] we wish finally to compare our results with the approximate expressions given by ferrell @xcite , which require @xmath40 and @xmath41 , i.e. @xmath42 . by expanding eq .
( [ turbidity - largeq0 ] ) and setting as in ref .
@xcite @xmath43 we obtain @xmath44.\ ] ] in order to compare with ferrell s results , we must compute @xmath45 . since , using the same approximations @xmath46 , we obtain @xmath47 this formula agrees with ferrell s expression once we recognize that @xmath48 since @xmath49 for a purely ornstein - zernicke behavior .
numerically , we predict @xmath50 , which is smaller than ferrell s numerical result 8.4 .
ferrell s expression predicts a turbidity that is somewhat higher than ours .
indeed , his numerical result implies @xmath51 in eq .
( [ turbidity - largeq0 ] ) , and as consequence we would obtain @xmath52 ( resp .
1.085 ) for @xmath53 ( resp . 1000 ) , to be compared with our prediction @xmath54 ( resp .
1.069(3 ) ) .
another expression for the turbidity that takes into account the anomalous decay of the structure factor is given in ref .
it assumes that@xcite @xmath55 , where @xmath56 .
it follows that @xmath57\left[4 - 2 b(\eta-4 ) + b^2 ( \eta^2 + 2 \eta + 8)\right ] - 4 \eta b(1+b ) \over
b^3 \eta ( 2 + \eta ) ( 4 + \eta ) } , \ ] ] where @xmath58 .
such an expression however predicts a turbidity that is too large .
for instance , for @xmath59 it gives @xmath60 , to be compared with our prediction @xmath61 , cf .
table [ table - turb ] .
note the correct turbidity @xmath62 is larger than @xmath63 since @xmath24 decreases slower for @xmath64 than the ornstein - zernicke approximation .
however , this is apparently in contrast with the experimental results for the binary fluid mixture methanol - cyclohexane presented in ref . @xcite .
the work of v.m.m . was supported by the european commission contract hpmf - ct-2000 - 00450 and by ocyt ( spain ) , project no . fpa2001 - 1813 .
a. pelissetto and e. vicari , `` critical phenomena and renormalization - group theory , '' e - print cond - mat/0012164 .
the structure factor directly determines the scattered light intensity in the absence of multiple scattering . however , multiple scattering is important in experiments on fluids @xcite , and it is therefore essential to perform the appropriate corrections to the data @xcite .
note also that we consider here only rayleigh elastic scattering , which is the dominant contribution near the critical point , see , e.g. , ref .
@xcite .
m. e. fisher , j. math .
* 5 * , 944 ( 1964 ) .
m. e. fisher and r. j. burford , phys . rev .
* 156 * , 583 ( 1967 ) .
m. e. fisher and a. aharony , phys .
b * 10 * , 2818 ( 1974 ) . v. g. puglielli and n. c. ford , jr . ,
lett . * 25 * , 143 ( 1970 ) .
d. beysens , a. bourgou , and p. calmettes , phys . rev .
a * 26 * , 3589 ( 1982 ) .
r. f. chang , h. burstyn , and j. v. sengers , phys .
rev . a * 19 * , 866 ( 1979 ) .
r. schneider , l. belkoura , j. schelten , d. woermann , and b. chu , phys .
b * 22 * , 5507 ( 1980 ) .
p. damay , f. leclercq , and p. chieux , phys .
b * 40 * , 4696 ( 1989 ) .
y. izumi , phys .
a * 39 * , 5826 ( 1989 ) .
s. janssen , d. s. schwahn , and t. springer , phys .
* 68 * , 3180 ( 1992 ) . h. sato , n. kuwahara , and k. kubota , phys .
e * 53 * , 3854 ( 1996 ) .
m. lesemann , a. martn , l. belkoura , d. woermann , and e. hoinkis , ber .
phys . chem .
* 101 * , 228 ( 1997 ) .
m. bonetti , c. bagnuls , and c. bervillier , j. chem . phys .
* 107 * , 550 ( 1997 ) .
m. bonetti and p. calmettes , rev .
sci . instrum . * 68 * , 4163 ( 1997 ) ; int .
j. thermophys .
* 19 * , 1555 ( 1998 ) .
p. damay , f. leclercq , r. magli , f. formisano , and p. lindner , phys .
b * 58 * , 12038 ( 1998 ) .
m. bonetti , g. romet - lemonne , p. calmettes , and m .- c .
belissent - funel , j. chem . phys . *
112 * , 268 ( 2000 ) .
m. bonetti , p. calmettes , and c. bervillier , j. chem .
phys . * 115 * , 4660 ( 2001 ) .
m. campostrini , a. pelissetto , p. rossi , and e. vicari ,
25th - order high - temperature expansion results for three - dimensional ising - like systems on the simple cubic lattice , " cond - mat/0201180 .
r. a. ferrell and d. j. scalapino , phys .
* 34 * , 200 ( 1975 ) .
a. j. bray , phys .
b * 76 * , 1248 ( 1976 ) . r. a. ferrell and j. k. bhattacharjee , phys
lett . * 42 * , 1505 ( 1979 ) . m. e. fisher and j. s. langer , phys .
lett . * 20 * , 665 ( 1968 ) .
e. brzin , d. amit , and j. zinn - justin , phys .
* 32 * , 151 ( 1974 ) .
e. brzin , j. c. le guillou , and j. zinn - justin , phys .
* 32 * , 473 ( 1974 ) .
m. campostrini , a. pelissetto , p. rossi , and e. vicari , europhys . lett .
* 38 * , 577 ( 1997 ) ; phys .
e * 57 * , 184 ( 1998 ) .
m. campostrini , a. pelissetto , p. rossi , and e. vicari , phys .
e * 60 * , 3526 ( 1999 ) .
v. agostini , c. carlino , m. caselle , and m. hasenbusch , nucl .
b * 484 * , 331 ( 1997 ) ; m. caselle , m. hasenbusch , and p. provero , nucl .
b * 556 * , 575 ( 1999 ) .
h. b. tarko and m. e. fisher , phys .
* 31 * , 926 ( 1973 ) ; phys .
b * 11 * , 1217 ( 1975 ) . m. combescot , m. droz , and j. m. kosterlitz , phys . rev .
b * 11 * , 4661 ( 1975 ) .
m. e. fisher and s .- y .
zinn , j. phys .
a * 31 * , l629 ( 1998 ) .
r. a. ferrell , physica a * 177 * , 201 ( 1991 ) .
p. calmettes , i. lagus , and c. laj , phys .
* 28 * , 478 ( 1972 ) . in ref .
@xcite , the scaling function @xmath24 is written as @xmath65 , so that the correlation length defined there does not coincide with the usual second - moment one . in order to be consistent with our normalizations ,
we have introduced the correction factor @xmath66 .
d. t. jacobs , s. m. y. lau , a. mukherjee , and c. a. williams , int .
j. thermophys .
* 20 * , 877 ( 1999 ) .
j. g. shanks and j. v. sengers , phys . rev .
a * 38 * , 885 ( 1988 ) .
m. a. anisimov , _ critical phenomena in liquids and liquid crystals _ ( gordon and breach , new york , 1991 ) . | we perform a large - scale monte carlo simulation of the three - dimensional ising model on simple cubic lattices of size @xmath0 with @xmath1 and 256 . we determine the corresponding structure factor ( fourier transform of the two - point function ) and
compare it with several approximations and with experimental results .
we also compute the turbidity as a function of the momentum of the incoming radiation , focusing in particular on the deviations from the ornstein - zernicke expression of puglielli and ford . |
bell nonlocality is one of the most iconic and paradoxical aspects of quantum phenomenology .
the realities of two space - like separated systems can be entangled , such that the choice of whether to measure @xmath0 or @xmath1 on one system can affect the measurement statistics of @xmath2 on the other @xcite .
contextuality generalized these ideas , noting that the essence of non - locality remains in single , localized quantum systems .
the outcome of an observable @xmath2 can depend on whether it is comeasured with @xmath0 or @xmath1 , even if neither observable @xmath0 nor @xmath1 disturbs the measurement statistics of @xmath2 @xcite . unlike nonlocality
, contextuality can be exhibited by classical as well as quantum systems @xcite .
all contextuality experiments so far , have relied on collecting output statistics from a series of joint measurements @xmath3 , @xmath4 , etc . on some quantum state .
while this approach has been immensely successful @xcite , all outcomes in these tests can be replicated by classical systems .
suppose two black boxes are tested for contextuality .
one contains a genuine quantum system that faithfully outputs the required quantum measurement statistics , the other contains a classical computer which merely simulates the mathematics of quantum theory .
current experimental tests of contextuality can not distinguish which of the two is quantum .
this stands in stark contrast to tests of nonlocality where local hidden variables can be excluded due to space like separation of alice and bob s choice of measurement settings and no - signalling .
the inability to exclude hidden variables is a key limitation of current contextuality experiments , which in turn , hinders their use as a resource in device independent scenarios - such as generating certified random numbers @xcite .
this article aims to mitigate this limitation .
we introduce experimental methods to exclude contextual hidden variables through physical principles and apply them to refine experimental tests of contextuality in quantum systems .
our approach rests on the use of an external quantum mechanical degree of freedom . in synthesizing superpositions of different contextuality measurements conditioned on this degree of freedom
, we can imprint the predictions of contextuality into correlations between the system of interest , and our quantum mechanical ancilla .
the outcomes of contextuality experiments can thus be translated into a bipartite setting . in doing so
, we can replace the standard ad hock assumption in contextuality tests - that any participating classical system did not reconfigure its answers based on the measurement settings - with a much more physically motivated assumption of no - signaling .
this puts tests of contextuality on a similar footing to nonlocality tests , and ultimately reduces the current gap between them .
the article is organized as follows .
section [ sec : contextualitytests ] will introduce the formal framework of contextuality .
section [ sec : results ] will outline the specifics of our proposed experiment , together with corresponding proofs of how it can distinguish contextual hidden variables from genuine quantum contextuality .
section [ sec : blackbox ] describes these ideas within the general black box framework .
section [ sec : discusion ] then concludes the paper with discussions .
suppose that a system has @xmath5-observables @xmath6 , including an observable @xmath7 which is compatible with several distinct proper subsets of @xmath8 .
each compatible subset constitutes a measurement context for @xmath7 .
the non - contextual hypothesis is : the outcome of @xmath7 is independent of which subset you choose to measure @xmath7 alongside .
quantum systems with three or more hilbert space dimensions violate his hypothesis - they are innately contextual @xcite .
tests are formulated in terms of non - contextual inequalities ; mathematical inequalities that constrain all non - contextual systems .
an archetypical example is the @xmath5-cycle inequality : @xmath9 where observables @xmath10 have dichotomic @xmath11- outcomes , all addition is done modulo @xmath5 , and any consecutive pair @xmath7 and @xmath12 can be measured simultaneously @xcite .
this inequality is derived under the no - disturbance assumption the marginal distributions @xmath13 is assumed to be independent of the measurement context , i.e. @xmath14 . provided this assumption
is upheld , any non - contextual hidden variable will satisfy . meanwhile quantum systems can violate up to the tsirelson bound @xmath15 if @xmath5 is even , or @xmath16 if @xmath5 is odd @xcite .
contextual hidden variables can also violate this inequality . consider a hidden variable that flips an unbiased coin to generate @xmath17 or @xmath18 at random
, then computes @xmath19
. this hidden variable will violate the @xmath5-cycle inequality up to the arithmetic bound of @xmath20 .
thus contextuality is not unique to quantum systems .
it is also exhibited by classical processes which take the full complement of observables being simultaneously measured @xmath21 and uses this input to generate a context - dependent - outcome @xmath22 for each observable @xmath7 , i.e. a contextual hidden variable .
this observation limits the capacity of standard contextuality tests to guarantee the non - classicality of an untrusted physical system .
an unknown system may violate a noncontextual inequality simply by hosting a computer that executes a suitable contextual hidden variable model . indeed many such automata have already been proposed @xcite .
* framework*. this motivates us to consider the following scenario .
alice wishes to test if a quantum system @xmath23 inside a black box is contextual , while excluding contextual hidden variables .
this may be motivated by a desire to generate quantum random numbers through a certifiable quantum source of contextuality .
her adversary wishes to cheat her by replacing this black - box with an imitation that mimics the measurement results through contextual hidden variables @xcite .
alice is tasked with distinguishing the genuinely quantum system from the imitation and simultaneously checking for contextuality - i.e. the task of discriminating between these two boxes is equivalent to excluding a contextual hidden variable explanations for her experimental data .
alice has an experimental apparatus , which can be pre - configured to measure any pair of observables @xmath24 and @xmath12 .
she also has a source of black boxes ; supplied by an untrusted third party with the capability to generate classical or quantum boxes . in each run
alice uses the source to generate a black box , and then inserts it into the experimental apparatus , without knowing whether it contains a quantum or classical system .
the apparatus then returns two measurement outcomes @xmath25 and @xmath26 , respectively . at the end of the run alice
uses the outcomes as data points for @xmath22 and @xmath27 in her experiment .
we note two properties that are required for the experimental test to conclude a system is contextual : * if an observable @xmath7 is measured twice during a single run , then both answers should be the same .
[ item1 ] * the data will violate the @xmath5-cycle inequality in eqn .. [ item2 ] alice will thus screen her data for compliance with both ( i ) and ( ii ) . criterion ( i ) is often applied to strings of measurements @xmath28 . checking the outcomes of @xmath7 ( and of @xmath12 ) are consistent in any string of measurements @xmath29 , is an operational way of verifying the no - disturbance assumption .
this check is often advocated and used in experimental implementations , where the emphasis is on checking the quantum measurements commute . in the following sections , we develop experimental methods for excluding contextual hidden variables and
use them to refine existing experimental tests of contextuality .
our new refined tests can exclude contextual hidden variables , due to inability to satisfy the aforementioned criteria ( i ) and ( ii ) .
to implement the aforementioned standard contextuality tests , alice needs to choose measurement settings at random .
this involves the use of a reliable source of randomness .
one method , for example , is via the use of an ancillary qubit @xmath30 initiated in state @xmath31 . by measuring this qubit in the standard pauli @xmath32 basis ( i.e. , the @xmath33 basis ) , alice can obtain a random bit @xmath34 that determines which measurement she will perform . for example , alice may use the random bit to decide which of two sets of mutually commuting observables @xmath35 , or @xmath36 , where @xmath37 to measure .
that is she configures her apparatus to co - measure @xmath38 if the random bit if @xmath39 , and @xmath40 if @xmath41 .
basis measurement outcomes of an ancillary qubit @xmath42 , are used as a quantum random number generator ( qrng ) .
if the qrng outputs 0 then the observables @xmath43 are measured , else when the qrng outputs 1 the measurement settings are flipped to @xmath44 . in the revamped protocol ( b ) a control - unitary is used to reverse the order of measurement , so the settings can be flipped retroactively - that is we measure the system we are testing for contextuality , afterwards we fix the settings by measuring the ancilla in the @xmath32 basis .
this figure highlights the tight knit relationship with quantum delayed choice protocols @xcite .
[ fig : schematic],scaledwidth=50.0% ] our proposal is based on reversing this order of events . instead of pre -
selecting the measurement setting , alice will delay this choice until after the system of interest @xmath45 , has been measured . formally , suppose there exists a unitary @xmath46 , that transforms each element of @xmath47 to an element of @xmath48 , i.e. , @xmath49 . for @xmath5-cycle tests based upon the standard observables @xmath50 as in @xcite ,
the compatible subsets @xmath51 and @xmath52 where @xmath53 , satisfy this assumption provided we ensure @xmath47 and @xmath48 have the same number of elements .
consider the following general strategy .
let @xmath23 represent the state of the system of interest , and @xmath54 the state of the ancilla qubit . instead of measuring the system directly
, alice first applies the aforementioned @xmath46 controlled on the @xmath32 basis of the ancilla qubit , i.e. , alice applies @xmath55 . following this process
, alice makes the standard measurements associated with each of the observables in @xmath47 to obtain outcomes @xmath56 .
once these measurement outcomes are obtained , alice generates the random number @xmath57 by a @xmath32 measurement of the ancilla ( see fig . [
fig : schematic ] for details ) .
when @xmath58 , alice treats @xmath56 as the outputs of measuring the elements of @xmath47 and use them to collect statistics on @xmath59 .
otherwise the outputs are treated as if the elements of @xmath48 were measured ; and used to evaluate @xmath60 for elements of @xmath48 .
thus , alice now decides what questions to ask a possibly contextual system of interest _
after _ she receives the answers to the said questions .
the main idea is that the black box s incumbent phd student must now generate outcomes , without knowing exactly what observables are being measured .
such a phd student can no longer execute a contextual hidden variable strategy , which uses the full complement of observables being measured to compute outcomes . in
what follows , we demonstrate appropriate choice of @xmath47 and @xmath48 allows alice to exclude classical contextual hidden variables .
the exact specifics will vary depending on whether the number of observables , @xmath5 , is even or odd .
[ sec : evenncycles ] * when @xmath5 is even * , we execute the protocol above with @xmath61 and @xmath62 for some @xmath63 . in instances where the ancilla measurement outcome @xmath64 , is @xmath65 , we have effectively measured @xmath66 and @xmath67 and thus can collect statistics about @xmath68 . in instances where @xmath41 , similar reasoning allows us to collect information on @xmath69 .
for example , repetition of the protocol when @xmath70 is set to @xmath65 allows estimation of @xmath68 and @xmath71 . reiterating this procedure for @xmath72 each time allows evaluation of each of the correlation terms in eq . .
consider a contextual classical box that attempts to mimic these statistics . without loss of generality
, we can model such a box by assuming it contains an internal memory - a contextual hidden variable @xmath73 .
the box is assumed to have complete knowledge of alice s experimental setup ( i.e , what @xmath47 and @xmath74 are ) , but no knowledge about the outcome of alice s eventual measurement of the ancilla ( we will see in section [ sec : blackbox ] that the latter assumption can be verified experimentally ) . in each run alice
makes a measurement for each observable in @xmath75 . to successfully fool alice
, the box must replicate the quantum output statistics for @xmath25 and @xmath26 . in the following
we show that this is impossible ; a classical box with contextual hidden variables can not consistently satisfy conditions ( i ) and ( ii ) from sec .
[ sec : contextualitytests ] , making it impossible for such a system to mimic genuine quantum contextuality without detection .
_ proof of result .
_ in order to violate the @xmath5-cycle inequality , it is necessary that @xmath76 this holds independently for each @xmath77 . to see this write as @xmath78 , where @xmath79 . because @xmath80 contain @xmath81 correlation terms which satisfy @xmath82 , we are guaranteed that @xmath83 .
hence in order to violate we need @xmath84 .
we take the outcomes @xmath85 , collected from runs where @xmath86 and @xmath87 , in conjunction with the output of the ancilla measurement , @xmath64 , and evaluate @xmath88 we observe that if we define probability distributions @xmath89 and @xmath90 , then @xmath91 is the variational distance between @xmath92 and @xmath93 .
let us assume a contextual hidden variable @xmath73 supplied the outcomes @xmath25 and @xmath26 .
this assignment must be made without knowing @xmath64 , the future outcome of measuring the ancilla qubit in the @xmath32 basis .
we note that @xmath64 equiprobable to be 0 or 1 , since the ancilla is prepared in a @xmath94 state and the control unitary does not effect the @xmath32 statistics of the ancilla ( i.e. the control unitary commutes with a @xmath32 basis measurement of the ancilla , since with respect to the @xmath32 basis of the ancilla the control unitary is @xmath95 , where @xmath96 is the direct sum and @xmath97 is the identity . ) . with no prior knowledge of @xmath64 available when @xmath73 is assigned @xmath98 , and @xmath99 hence the variational distance @xmath100 in eq .
is precisely 0
. a contextual hidden variable can not violate inequality and therefore will not violate the @xmath5-cycle inequality .
of course if the experiment is not repeated enough times , then a classical system could appear to satisfy due to random statistical fluctuations . in practice
this problem effects all contextuality tests , and has a standard resolution .
quantum systems actually violate the @xmath5-cycle inequality up to the tsirleson @xcite bound @xmath101 in sec . [
sec : contextualitytests ] so the quantum statistics ideally satisfy a stricter inequality @xmath102 .
alice will pick an @xmath103 satisfying @xmath104 , and check if the experimental outcomes satisfy : @xmath105 due to the central limit theorem , the probability that statistics generated by a classical hidden variable , will satisfy , dies off exponentially with the number of runs @xmath106 . * when @xmath5 is odd * , the protocol is summarized by fig .
alice intends to evaluate the correlation term @xmath107 for some @xmath108 .
the main difference here is that alice first implements a non - destructive measurement of observable @xmath109 on system @xmath45 , with outcome @xmath25 .
once done , she then implements the procedure in sec .
[ sec : results ] , with @xmath110 and @xmath111 .
that is she introduces an ancillary qubit @xmath42 , and applies the unitary @xmath46 satisfying @xmath112 to @xmath45 , controlled on the @xmath32 basis of @xmath113 . to determine what measurements she has effectively made
, alice completes the protocol by measuring the ancilla in the @xmath32 basis , with outcome @xmath64 . on runs where @xmath114
, alice attributes the outcomes @xmath115 to two consecutive measurements of @xmath109 .
otherwise , alice treats @xmath25 as an outcome of measuring @xmath109 and @xmath26 as an outcome of measuring @xmath116 .
thus , depending on the outcome @xmath64 , alice can collect statistics about either @xmath117 or @xmath107 .
cycle case we use an alternate protocol , where each run allows us to measure @xmath107 for some @xmath118 , or @xmath117 . a non - destructive measurement of the observable @xmath109 on the register is followed by a control unitary @xmath119 where @xmath120 .
finally we repeat the measurement of @xmath109 , and measure the ancilla in the @xmath32 basis generating an outcome @xmath121 .
supposing we start with @xmath122 then post selecting on runs where the ancilla outcome @xmath123 , we can evaluate @xmath71 meanwhile when @xmath58 the data yields @xmath124 .
[ fig3],scaledwidth=30.0% ] analogously to the even @xmath5-cycle case , a classical box with contextual hidden variables must replicate the output statistics of @xmath25 and @xmath26 , for both possible measurement outcomes of the ancilla .
this is not possible .
_ proof of result .
_ in order to satisfy the criteria ( i ) and ( ii ) from sec .
[ sec : contextualitytests ] , it is necessary that @xmath125 this inequality must be satisfied independently for each @xmath126 .
this follows from observing two points .
firstly criterion ( i ) implies @xmath127 . secondly to violate the @xmath5-cycle inequality ( criterion ( ii ) ) , we require @xmath128 for all @xmath108 . we can establish this using proof by contradiction : assume @xmath129 for some @xmath130 and we observe violation of inequality .
this would imply @xmath131 .
but this sum only contains @xmath132 terms each lower bounded by -1 .
this is a contradiction . by mirroring the analysis in the even case ,
specifically eq . -
, we observe @xmath133 again @xmath134 , where @xmath135 and @xmath136 are defined analogously to the even @xmath5 case .
if @xmath25 and @xmath26 are generated by a contextual hidden variable @xmath73 , then using the same argument as in eq .
we conclude @xmath137 is exactly zero provided no prior information about @xmath64 is available to influence the choice of @xmath73 .
hence the contextual box will not satisfy , implying that either condition ( i ) or ( ii ) from sec . [
sec : contextualitytests ] , will be contradicted .
as with all contextuality experiments , the physical implementation of this protocol must satisfy an important caveat . during this implementation alice will need to collect data for all @xmath5 correlation terms @xmath138 .
for each correlation term alice must set up her experiment in figs .
[ fig : schematic]-[fig3 ] using a different @xmath139 gate . if an observable @xmath109 is measured in two different setups
i.e. the setups for @xmath140 and @xmath141 then the actual measurement @xmath109 must be physically identical in both cases .
otherwise any observed dependence of the outcomes @xmath142 on whether we measure @xmath143 or @xmath144 ( i.e. any contextuality ) could be due to imperfect reconstruction of @xmath109 when changing the settings . in practice
this problem is solved in standard contextuality tests by ref . @xcite and others .
we simply refine the contextuality test in ref .
@xcite to ensure we have also fulfilled this basic tenant .
the even case is illustrated in fig .
[ fig2 ] , the odd case will follow the same principles and therefore has not been included here . cycles .
figure ( a ) demonstrates the example @xmath145 , which allows us to measure the correlation terms @xmath146 and @xmath147 .
unitary @xmath148 is applied to the register , controlled on the state of the ancilla @xmath42 .
note @xmath148 is chosen so that it commutes with measurements of @xmath66 , i.e. @xmath149 , and satisfies @xmath150 .
afterwards we measure @xmath66 and @xmath67 on the register , followed by a @xmath32 basis measurement on the ancilla . in ( b )
we need to extend this protocol to get the next correlation term @xmath151 .
however we must ensure that in both ( a ) and ( b ) , the measurements associated with @xmath152 and @xmath2 are physically identical . to do this
we follow the techniques in ref .
we introduce a new control unitary @xmath153 such that @xmath154 while @xmath155 , so that measurements of observable @xmath2 commute with the action of @xmath153 .
note that measurements of @xmath156 will also commute with the @xmath153 gate ( i.e. to measure @xmath157 we postselect on runs where the effect of the @xmath153 gate is the same as an identity transformation ) . in principle , because the only difference between ( a ) and ( b ) is a @xmath153 gate , and this gate does not effect the measurement outcome of @xmath2 , or @xmath156 , these observables should be the same in both setups .
furthermore the action of the two control unitaries is equivalent to a single control gate @xmath158 where @xmath159 while @xmath160 .
iteratively repeating this method along with the techniques in ref .
@xcite allows us to acquire the remaining correlation terms @xmath161 in the @xmath5-cycle inequality .
[ fig2],scaledwidth=50.0% ]
so far we have made use of quantum mechanical terminology .
this is in line with the paper s focus on describing concrete experiments that can rule out contextual hidden variables .
of course , contextuality is often discussed in a completely black - box setting - where the only properties of a system are its output statistics given input questions ; and no assumptions are made that the system is quantum mechanical .
our proposal generalized naturally into this picture .
because the security of this protocol relies on independence of the control from the system being tested for contextuality ; we need to introduce two parties alice and bob , who share a pair of black boxes . indeed , since the quantum protocol required alice and bob to share @xmath5-different entangled resource states
one for each of alice s settings choices @xmath162 , in the @xmath5-cycle test the black box analogue actually requires @xmath5 pairs of boxes , labeled @xmath162 .
we want recast the measurements in figs .
[ fig : schematic]-[fig2 ] as a series of questions alice and bob ask their respective boxes . in each round of the @xmath5-cycle test alice
( and bob ) now select a pair of boxes labeled ` @xmath70 ' .
alice asks her box two sequential dichotomic questions @xmath163 and @xmath164 and gets answers @xmath165 .
afterwards bob asks his box a single question @xmath32 which generates outcome @xmath166 or @xmath167 .
the outcomes @xmath168 can be treated as the outcome from sec .
[ sec : results ] for the aforementioned values of @xmath70 and @xmath5 . * verifying non - determinism of measurement selection . *
recall that our proofs above relied on the crucial assumption that the box does not have access to the value of @xmath64 prior to outputting @xmath25 and @xmath26 .
this relies of the assertion that there is no hidden variable that determines the value of @xmath64 prior to its measurement .
the black - box framework gives us a natural way to explicitly verify this assumption .
on some runs , after selecting their boxes alice and bob may randomly choose to perform a bell test between their box pairs . during
this bell test alice chooses to ask her box one of two possible questions @xmath169 or @xmath170 ; while bob chooses to asks his box either question @xmath171 or @xmath172 .
@xmath171 is selected to be the same question , @xmath32 , bob would have asked in delayed choice contextuality .
the resulting statistics are used to test violation of the chsh inequality @xcite : @xmath173 the violation of this inequality certifies there is no realistic description for outcomes of question @xmath171 .
this demonstrates that there is no hidden variable that predicts alice s decision of whether to ask questions in @xmath47 or @xmath48 .
thus any capacity for the box to gain this knowledge will violate causality . in the specific case of quantum systems , the bell test is performed on the final state of the circuits in figs .
[ fig2]-[fig3 ] .
the control qubit is assumed to be inside bob s box @xmath70 ; and the register inside alice s box ` @xmath70 ' .
@xmath171 represents a @xmath32-measurement , in line with our constraint that it asks the same question as our refined contextuality test .
provided the register was in some pure state @xmath174 directly before the control unitary gate @xmath139 , the resulting bipartite system is able to violate the chsh inequality for any @xmath5 for some choices of @xmath172 , @xmath169 and @xmath170 .
specifically let @xmath172 be a @xmath175 measurement ; such that bob measures the ancilla in either the @xmath175 or @xmath32 basis .
meanwhile define @xmath176 , @xmath177 where @xmath178 , as an effective qubit on alice s system of interest ; with associated pauli operators @xmath179 and @xmath180 ; and let @xmath181 and @xmath182 represent the two measurement settings for alice .
application of these measurement settings allows violation of the chsh inequality up to @xmath183 , for all @xmath184 . for more details
see the appendices .
here we have developed experimental techniques for excluding contextual hidden variables and applied them to refine contextuality experiments . by controlling the measurement settings in the contextuality test on an external ancillary quantum mechanical degree of freedom ,
we present a method for effectively concealing which observable is being measured in the contextuality test .
this will effectively thwart a contextual hidden variable that needs to use prior knowledge of which observables are being measured in order to preassign a contextual outcome .
one of the major stumbling blocks in the use of contextuality as a quantum resource in line with non - locality has been its capacity to be simulated by classical information processing .
how can we use contextuality as a resource if we can replace the same system with a student in a box ? in developing an experimental method to certify a quantum state is contextual , and simultaneously exclude contextual hidden variables , our protocol helps address this problem .
this innovation may thus lead to potential applications for contextual quantum systems - the natural candidate being , for example , the certified generation of random numbers @xcite it would be interesting to see if some of the other problems surrounding contextuality tests could be addressed using similar ideas .
for instance the no - disturbance assumption which is the direct equivalent of no - signalling in bell tests is a hot button topic in contextuality tests .
a key issue is that hidden variable models which do not satisfy the no - disturbance condition ( i.e. the marginal distribution they predict @xmath185 depends on the measurement context @xmath109 ) can violate a noncontextual inequality .
this opens an interesting possibility : could a refined protocols , which effectively hides the measurement context ` - ' in any joint measurement @xmath186`- ' , help address this problem ?
furthermore the topic of contextuality as a resource is controversial , particularly since according to state independent contextuality protocols even the completely mixed state is contextual . by imprinting the predictions of contextuality into correlations between the test system and an external ancilla , we give a new interpretation for contextual statistics , which also links back to the quantum circuit formalism .
this leaves a host of interesting problems which could be addressed in future work .
another point of interest , is the close relationship between our protocol and quantum delayed choice , a class of quantum protocols which have been experimentally demonstrated in the context of optical interferometry @xcite .
the measurement statistics of a photon exiting an interferometer depend on a setting in the interferometer which can be either open and closed . by controlling whether the interferometer is open or closed on an ancillary quantum degree of freedom
, we can effectively toggle the interferometer setting between open and closed after the photon has been measured @xcite .
this innovation , provides a remarkable way to rule out a hidden variable description for the photon s measurement statistics .
while the principles involved are similar to wheeler s delayed choice @xcite , quantum delayed choice is known to be more versatile and have applications to ruling out hidden variables in more general settings such as bell tests @xcite .
quantum delayed choice also furnishes a technique for measuring complementary phenomena using a single configuration of the measurement apparatus @xcite .
our work highlights the natural confluence of these ideas with tests of quantum contextuality .
j.t . would like to acknowledge input and discussions with marek wajs , kavan modi , daniel terno and kihwan kim .
this project / publication was made possible through the support of the john templeton foundation ( grant number 54914 ) and the foundational questions institute .
the opinions expressed in this publication are those of the author(s ) and do not necessarily reflect the views of the john templeton foundation .
the project was also supported by the national research foundation and the ministry of education in singapore through the tier 3 moe2012-t3 - 1 - 009 grant random numbers from quantum processes " as well as the national basic research program of china grant 2011cba00300 , 2011cba00302 , the national natural science foundation of china grants 11450110058 , 61033001 , and 61361136003 .
99 j. s. bell , physics 1 , 195 ( 1964 ) s. kochen & e.p .
specker , j. math .
17 , 59 ( 1967 ) m. kleinmann , o. ghne , j. r. portillo , j. a. larsson , & a. cabello , new journal of physics , 13(11 ) , 113011 .
( 2011 ) b. daki , m. suvakov , t. paterek , & .
brukner , phys .
lett . , 101(19 ) , 190402 ( 2008 ) n. harrigan , t. rudolph & s. aaronson , ( 2008 ) .
[ arxiv:0709.1149 ] r. lapkiewicz , p. li , c. schaeff , n. k. langford , s. ramelow , m. wieniak , & a. zeilinger , nature ( london ) 474 , 490 ( 2011 ) .
v. dambrosio , i. herbauts , e. amselem , e. nagali , m. bourennane , f. sciarrino , & a. cabello , phys .
x. 3 , 011012 ( 2013 ) x. zhang , m. um , j. zhang , s. an , y. wang , d .-
deng , c. shen , l .- m .
duan , & k. kim , phys .
110 , 070401 ( 2013 ) j. ahrens , e. amselem , a. cabello & m. bourennane , sci .
rep . 3 , 2170 ( 2013 ) m. um , x. zhang , j. zhang , y. wang , s. yangchao , d .-
deng , l .- m .
duan & k kim , sci .
rep . 3 , 1627 ( 2012 ) m. araujo , m. t. quintino , c. budroni , m. t. cunha , & a. cabello , phys .
a , 88(2 ) , 022118 ( 2013 ) s. l. braunstein , & c. m. caves , annals of physics , 202(1 ) , 22 - 56 ( 1990 ) a. a. klyachko , m. a. can , s. binicioglu , & a. s. shumovsky , phys .
lett . 101 , 020403 ( 2008 ) b. s. tsirelson , j. sov .
36 , 557 ( 1987 ) a. grudka , k. horodecki , m. horodecki , p. horodecki , m. pawowski , & r. ramanathan , phys .
a , 90 , 032322 ( 2013 ) j. f. clauser , m. a. horne , a. shimony & r. a. holt , phys .
lett . , 23 , 880 ( 1969 ) r. colbeck , & r. renner , nat . phys .
, 8(6 ) , 450 - 453 ( 2012 ) s. pironio , a. acn , s. massar , a. boyer de la giroday , d. n. matsukevich , p. maunz , s. olmschenk , d. hayes , l. luo , t. a. manning & c. monroe , nature 464 , 1021 - 1024 ( 2010 ) r. ionicioiu , & d. r. terno , phys .
lett . , 107(23 ) , 230406 ( 2011 ) r. ionicioiu , t. jennewein , r. b. mann , & d. r. terno nat .
commun . 5 , 3997 ( 2013 ) l. c. cleri , r. m. gomes , r. ionicioiu , t. jennewein , r. b. mann , & d. r. terno , foundations of physics , 44(5 ) , 576 - 587 ( 2014 ) r. ionicioiu , r. b. mann , & d. r. terno , phys .
lett . , 114 , 060405 ( 2015 ) j. s. tang , y. l. li , x. y. xu , g. y. xiang , c. f. li , & g. c. guo , nat .
, 6(9 ) , 600 - 604 ( 2012 ) a. peruzzo , p. shadbolt , n. brunner , s. popescu , & j. l. obrien , science , 338(6107 ) , 634 - 637 ( 2012 ) x. s. ma , s. zotter , j. kofler , r. ursin , t. jennewein , .
brukner , & a. zeilinger , nat .
, 8(6 ) , 479 - 484 ( 2012 ) p.
j. coles , j. kaniewski , s. wehner , nat .
commun . 5 , 5814 ( 2014 ) j.a .
wheeler , in quantum theory and measurement , j.a.wheeler & w.h .
zurek , eds .
( princeton univ . press , princeton , nj , 1984 ) ,
wheeler , in mathematical foundations of quantum mechanics , a.r .
marlow , ed .
( academic , new york , 1978 ) , pp . 948 .
here we elaborate on the bell test in more detail .
we illustrate the idealised scenario where the control qubit is in a state @xmath42 directly before the control unitary gate , while the register is in a pure state @xmath187 . in standard
@xmath5-cycle experiments the system being tested for contextuality is pure which should substantiate this assumption , see for instance @xcite .
however any additional noise in the system being tested for contextuality should simply decrease the predicted bell violation . given these assumptions the state of the circuit after the control unitary @xmath139 is : @xmath188 we use the two subspaces @xmath176 , @xmath177 where @xmath178 , to define an effective qubit on alice s system of interest .
we then equip alice with two observables @xmath181 and @xmath182 defined interms of pauli operators for her effective qubit @xmath179 and @xmath180 . and re - expresses the circuit s final state as @xmath189 let s assume the control qubit goes to bob and the register to alice ; and that bob uses with the standard pauli observables @xmath190 and @xmath191 on his qubit . now alice and bob can test a bell inequality : @xmath192 in terms of the condensed notation @xmath193 , the expectation value on state is : @xmath194 to find the best possible bell violation @xmath195 , we optimize as a function of @xmath196 , @xmath197 . we know
this quantity is greater than ( or equal to ) the value of when @xmath198 , i.e. : @xmath199 now we can use @xmath200 , together with the choice @xmath201 , and @xmath202 to simplify this to @xmath203 for any @xmath204 this is always greater than 2 , provided alice chooses @xmath205 so that @xmath206 . note that when @xmath207 ( i.e. in the limiting case @xmath208 ) the control unitary obviously generates no entanglement and consequently we do not expect any bell inequality violation when @xmath209 .
this is not a problem for our application which should always fall in the regime @xmath210 .
hence alice and bob should be able to violate a bell inequality using the output of the circuits in fig .
[ fig3 ] and [ fig2 ] .
in addition bob may uses @xmath175 and @xmath32-directions as his two measurement settings during the bell test . by using these settings
we can establish the z - outcomes of the control qubit in the contextuality protocol are not realistic .
consequently during any run of the contextuality test , the choice of which two observables are being measured on the register is not predetermined ( because this choice depends on the z - outcome of the control qubit - a non realistic quantity ) .
@xmath212-cycle contextuality tests , are usually performed on @xmath213-level quantum systems in the state @xmath214 , where we have chosen to label the hilbert space basis @xmath215 .
the optimal settings for observables @xmath216 are @xmath217 where any pair @xmath109 and @xmath218 are compatible and @xmath219 for example in the @xmath213-cycle case we have : @xmath220 the theory of euler rotations will give a systematic method for computing @xmath46 , provided we use the representations of @xmath221 given in refs .
@xcite , we provide another appendix on how this representation is related to - .
all other details should follow unaltered .
tests of @xmath222-cycle inequalities based on eq . , are typically implemented in a 3-level quantum system .
the observables are : @xmath223 where the ray @xmath224 for @xmath225 @xcite and @xmath108 .
the state that violates this inequality maximally , and is therefore usually adopted in experimental tests is @xmath226 .
the protocol could be implemented by control - unitary based : @xmath227 where @xmath228 form a basis for the 3-d hilbert space , @xmath229 and @xmath230 .
this is tantamount to a @xmath231 rotation in the plane spanned by @xmath232 and @xmath233 .
all @xmath5-cycle inequalities for even @xmath5 have a one to one mapping onto a chained bell inequalities ( where alice and bob each have one half of a bell pair and @xmath234 observables ) . in this case
our 4-level system basis vectors @xmath235 from sec .
[ sec : evenncycles ] should be identified with a basis for the 2-qubit space according to @xmath236 .
the state that violates the chained bell inequality maximally is @xmath237 and the optimal settings are @xmath238 for an exact correspondence with sec .
[ sec : evenncycles ] note that with the basis identification above @xmath239 in eq . .
similarly @xmath240 , while @xmath241 and @xmath242 .
we highlight that for the chsh inequality @xmath243 using the bell state @xmath244 , the settings prescribed by are @xmath245 the problem of finding @xmath46 such that @xmath246 and @xmath247 in sec . [ sec : evenncycles ] , now inheres a systematic solution from the theory of euler rotations .
this solution is the form @xmath248 , where each single qubit unitary @xmath249 for @xmath250 , has the free parameters , @xmath251 ( rotation angle and axis ) .
parameters @xmath252 are fixed by @xmath253 , while @xmath254 are fixed by @xmath255 . from the form of @xmath109 and @xmath256 in eq .
we observe that it is always possible to set @xmath257 .
once we have found @xmath46 , it is possible , albeit convoluted to rewrite @xmath248 in the basis @xmath258 . | a phd student is locked inside a box , imitating a quantum system by mimicking the measurement statistics of any viable observable nominated by external observers . inside a second box
lies a genuine quantum system .
either box can be used to pass a test for contextuality - and from the perspective of an external observer , be operationally indistinguishable .
there is no way to discriminate between the two boxes based on the output statistics of any contextuality test .
this poses a serious problem for contextuality tests to be used as viable tests for device independent quantumness , and severely limits realistic use of contextuality as an operational resource .
here we rectify this problem by building experimental techniques for distinguishing a contextual system that is genuinely quantum , and one that mimics it through clever use of hidden variables . |
with the advent of radioactive nuclear beams , the properties of nuclei beyond the proton stability line have attracted experimental and theoretical attention in recent years .
special interest is devoted to a unique aspect originating from the fact that protons and neutrons occupy the same orbits in nuclei with @xmath6 ( see @xcite for review ) .
consequently , one expects a strong proton - neutron ( @xmath0 ) interaction because of the large spatial overlaps between proton and neutron single - particle wave functions .
the correlation energies related to the @xmath0 interaction have been extracted from the experimental binding energies @xcite .
a double difference of binding energies has been analyzed with the aim of providing input to semiempirical mass formulas @xcite , and with relation to the clustering of nucleons as elementary modes of excitation in nuclei @xcite .
it has recently been used for study of the p - n interactions @xcite , and discussed in terms of schematic and realistic shell model calculations @xcite .
this approach using the double difference of binding energies may provide details of the @xmath0 interactions , the isoscalar ( @xmath7 ) and isovector ( @xmath8 ) @xmath0 interactions .
the @xmath0 part of the isovector pairing correlation near @xmath5 has been studied in terms of algebraic model @xcite and compared with shell model monte carlo calculations @xcite . on the other hand , there are many discussions of the roles of the isoscalar pairing interactions ( see refs.@xcite for instance ) .
the experimental data indicate @xcite that the symmetry energy accompanied by the so - called wigner energy behaves according to the @xmath9 dependence ( @xmath10 ) .
this form could come from the isospin - invariant hamiltonian .
it has been recently proposed that the wigner energy originates in the isoscalar pairing interaction @xcite .
they pointed out that the wigner energy can not be solely explained in terms of correlations between the @xmath11 isoscalar @xmath0 pairs , and the isoscalar @xmath0 pairs with the other @xmath4 contribute significantly @xcite .
in fact , a recent shell model calculation @xcite with the @xmath12 isovector and @xmath11 isoscalar pairing forces in the @xmath6 @xmath13 shell nuclei can not explain the magnitude of the wigner energy .
on the other hand , another paper @xcite has discussed the fact that the degeneracy of the @xmath14 and @xmath15 states in odd - odd nuclei with @xmath5 is produced by a balance of the symmetry energy and the @xmath12 isovector pairing correlation .
the lowering of the @xmath14 states in @xmath5 odd - odd nuclei , according to our investigation @xcite , is caused by a unique form of isoscalar ( @xmath0 ) pairing force including all @xmath4 components .
this result is consistent with that of satula et al .
@xcite the uniqueisoscalar @xmath0 pairing force , which can be expressed in a simple form including the @xmath9 term , manifests a close relation to the symmetry energy . we shall discuss this matter by a concrete calculation in this paper . the odd - even mass difference ( oemd ) , the extra binding energy of a nucleus relative to its neighbors ,
is known to be an obvious experimental evidence of the pairing correlation @xcite .
the pairing phenomena are well understood in terms of the proton - proton ( @xmath16 ) or neutron - neutron ( @xmath17 ) pair condensate , and described by the bardeen - cooper - schriefer(bcs ) theory @xcite .
the oemd is often interpreted as a measure of the pairing gap ( following the relation @xmath18 on the average ) in medium - heavy and heavy nuclei .
the oemd displays , however a different feature in @xmath6 nuclei such as a special increase at @xmath5 .
an ordinary estimation of the neutron or proton pairing gap from the oemd is not applicable to these nuclei .
on the other hand , it has recently been discussed that the oemd in light nuclei is affected by deformation as well as @xmath12 pairing correlation @xcite . a further investigation of the oemd should be made in @xmath6 nuclei .
we shall discuss the influence of the @xmath0 interactions on the oemd .
the development of recent radioactive nuclear beams facilities provides unstable nuclei beyond the line of proton stability .
experimental and theoretical investigations of proton emitters are increasing .
such phenomena allow a test of the various models on the proton - rich side . for @xmath6 nuclei , one can expect that the @xmath0 interactions also influence the separation energy . since the @xmath0 force is considered to be attractive
, it might increase the separation energy .
in fact , the calculated separation energies by all models without the @xmath0 interaction are smaller than those of experiments at the @xmath5 nuclei .
we need a reliable effective interaction to study the nuclear properties mentioned above .
we have proposed an extension of the @xmath19 model with four forces @xcite , which reproduces quite well the experimental binding energies and energy spectra in @xmath6 nuclei of @xmath2 and @xmath3 subshells .
this model interaction including different types of @xmath0 forces is very suitable for our purpose to study various aspects of the @xmath0 interactions .
the main purpose of this paper is to study the @xmath0 interactions , analyzing the double differences of binding energies , and to check the validity of our model examining various quantities such as the symmetry energy , the wigner energy , the odd - even mass difference and the separation energy in nuclei near @xmath5 .
the paper is organized as follows . in sec .
ii , we first review our model proposed in the previous paper . sec .
iii contains the analysis of the double differences of binding energies to probe the @xmath0 interactions .
the symmetry energy and the wigner energy are discussed in sec .
iv . in sec .
v , the odd - even mass differences are analyzed in detail , and the two - proton separation energies are calculated in sec .
vi . finally , sec .
vii gives conclusions .
we have proposed the following effective interaction extended from the @xmath19 force which is composed of four isospin - invariant forces ( see ref .
@xcite in detail ) : @xmath20 with @xmath21 where @xmath22=@xmath23 and @xmath24=@xmath25=@xmath26 .
we use the notation @xmath27 and @xmath28 for the spin and isospin of a nucleon pair , respectively .
the subscript @xmath29 denotes the @xmath30 components of isospin @xmath31 .
we also use the notation @xmath32 for a proton and @xmath33 for a neutron . here
@xmath34 is a single - particle hamiltonian and @xmath35 contains the four forces : @xmath36 stands for the isovector monopole pairing force , @xmath37 for the isoscalar quadrupole - quadrupole force , @xmath38 for the isovector quadrupole pairing force and @xmath39 for the @xmath4-independent isoscalar @xmath0 force .
the first two forces in the interaction ( [ eq:8 ] ) are an extension of the conventional @xmath19 force to the isospin - invariant one .
the p - n part of the monopole and quadrupole pairing forces as well as the @xmath0 component of the quadrupole - quadrupole force would be important for @xmath6 nuclei .
the last @xmath0 force is very important for reproducing the experimental binding energy .
it is important to note that @xmath39 can be expressed as a simple form @xmath40 where @xmath41 denotes the total number operator of valence nucleons ( @xmath42 ) and @xmath43 is the total isospin operator .
our @xmath0 interaction is composed of four different components and hence is useful in analyzing their respective contributions to various physical quantities .
we applied the above hamiltonian to examine the binding energies and energy spectra of nuclei with @xmath44 and @xmath45 @xcite .
we adopted only the @xmath2 shell for nuclei with @xmath46=80 - 100 regarding the @xmath47 core as inactive , and the @xmath3 shell for nuclei with @xmath46=40 - 50 regarding the @xmath48 core as inactive .
it may be necessary to extend these model spaces for quantitative discussion .
our calculation itself indicates the insufficiency of the model space @xmath49 about energy spectra .
we used an extended model space @xmath50 when comparing calculated energy levels with observed ones in nuclei with @xmath51 , while experimental data near @xmath52 seem to demand a further extension of the model space .
the previous paper , however , has shown that the single @xmath53 shell model is bearable for semiquantitative discussion about the nuclear binding energy .
this simple model makes it possible to clearly see the roles of respective @xmath0 interactions .
we therefore employ the same single @xmath53 shell model as that used in ref .
@xcite , where the following force strengths are used : @xmath54 @xmath55 our model with these sets of parameters is considered to be reliable for studying the p - n interactions in connection with physical quantities related to the binding energy in the @xmath2 and @xmath56 shell nuclei .
the mass dependence of the force parameter @xmath57 is taken into account in some cases , but it does not change qualitatively the result .
we define the @xmath58th double difference of binding energies as follows : @xmath59 where @xmath60 is the nuclear binding energy . here
the operator @xmath61 is defined as @xmath62 .
\label{eq:102}\end{aligned}\ ] ] the double difference of binding energies , @xmath63 , was introduced for investigating the semiempirical mass formula @xcite .
this quantity is expected to roughly represent the @xmath0 interactions between the last proton and neutron from the form of eq .
( [ eq:102 ] ) .
figure 1(a ) shows the plot of @xmath63 as a function of @xmath64 for nuclei in the mass region @xmath46=16 - 165 .
experimental data are taken from ref.@xcite .
we see two separate groups in fig .
1(a ) , namely , one is for the even-@xmath46 nuclei ( dots ) and the other is for the odd-@xmath46 nuclei ( crosses ) . in both cases , shell effects at @xmath65 or @xmath66 are present , while the patterns of dots and crosses are symmetric with respect to the average curve .
it is now convenient to divide @xmath63 into two parts : the average part of the even-@xmath46 and odd-@xmath46 nuclei and the deviation from it . as seen in fig .
1(a ) , the former is approximately written as @xmath67 and the latter has opposite signs for the even-@xmath46 and odd-@xmath46 nuclei as follows : @xmath68 this expression was originally given by de - shalit @xcite in the earliest investigations of the effective p - n interactions .
equation ( [ eq:103 ] ) describes the staggering with respect to the isotopes for even-@xmath46 and odd-@xmath46 nuclei .
large values of @xmath63 for even-@xmath46 nuclei ( dots ) near @xmath5 below @xmath52 are notable .
the data of @xmath69 are plotted in fig .
1(b ) as a function of @xmath64 .
( our definition of @xmath70 has a sign opposite to that of brenner s @xmath71 @xcite . )
the values of @xmath70 show a different behavior from @xmath72 .
it is interesting that the staggering of @xmath72 disappears in @xmath69 .
we see large scatters of dots and crosses for @xmath73 .
these correspond to @xmath70 of nuclei in @xmath6 , and the values of @xmath70 at @xmath5 are especially large . with decreasing mass @xmath46 , @xmath70 at @xmath5 increases .
if one neglects the dots and crosses in @xmath6 nuclei , @xmath70 varies rather smoothly .
this smooth trend is clear for @xmath74 and continues up to heavy nuclei .
this is due to the fact that there is no stable @xmath6 nuclei with @xmath75 .
figure 1(b ) clearly indicates the smooth systematic decrease of @xmath70 with increasing mass @xmath46 , which can be traced by the curve @xmath76 .
the deviations from the curve @xmath76 are small , and shell structure is not found .
this general trend of @xmath70 has long been known , and was discussed in several papers . in a recent paper @xcite , the dramatic spikes of @xmath70 at @xmath5 light nuclei were discussed in terms of both schematic and realistic shell model calculations , and the importance of @xmath7 @xmath0 interaction for the spikes was pointed out .
the following relationship is derived from eqs .
( [ eq:101 ] ) and ( [ eq:102 ] ) : @xmath77 . \label{eq:4}\end{aligned}\ ] ] substituting the empirical relationship ( [ eq:103 ] ) into eq .
( [ eq:4 ] ) , in the large @xmath46 limit we get @xmath78 thus , the systematic behavior of @xmath69 for @xmath75 can be explained from the relation ( [ eq:5 ] ) .
( strictly speaking , there is a deviation from @xmath76 due to the mass dependence of @xmath79 . )
furthermore , it is obtained that @xmath80 for @xmath81 have similar pattern to @xmath82 , and are also traced by the curve @xmath76 . to analyze the double differences of binding energies , we now express the ground - state energy as follows : @xmath83 where @xmath84 denotes the expectation value with respect to the ground state .
here , @xmath85 is the @xmath0 parts of the @xmath86 force , and @xmath87 is the proton - proton ( @xmath16 ) and neutron - neutron ( @xmath17 ) parts of the total interaction ( [ eq:8 ] ) .
figure 2 shows @xmath88 and @xmath89 as a function of valence - neutron number @xmath90 for nb and mo isotopes .
the @xmath0 part of the @xmath86 energy , @xmath88 , exhibits a characteristic odd - even staggering in nb isotopes , while @xmath89 gives a smooth line except for @xmath91 . on the other hand , for mo isotopes @xmath88
varies smoothly as @xmath90 increases , and indicates very different structure from that of nb isotopes .
this can be attributed to extra energy for the odd - odd nuclei , which mainly comes from the @xmath8 @xmath0 part of the @xmath86 force .
consider the double difference of ground - state energies , @xmath92 , using the operator @xmath93 defined by eq .
( [ eq:102 ] ) . since there are almost no contributions from the single particle energy @xmath94 and coulomb interaction to the double difference of binding energies as seen in the form of eq .
( [ eq:102 ] ) , one notices that @xmath92 is able to be compared directly with the experimental value @xmath63 . figure 3 shows the calculated and experimental double differences of binding energies , @xmath92 and @xmath63 , as a function of mass @xmath64 for the nb , mo , tc , and pd isotopes .
the nb and mo nuclei near @xmath5 at the beginning of @xmath2 shell region probably have the mixing of single particle levels , the @xmath95 , and @xmath96 .
as seen from fig .
3(a)-(d ) , however , the agreement with experiments is quite good .
our calculation reproduces the staggering , and also predicts the highest spikes at @xmath5 nuclei though no experimental data are present .
table i. the components of @xmath92 for the mo isotopes .
the first and second columns denote the @xmath8 components , and the third and fourth columns the @xmath7 components . [
cols="^,^,^,^,^,^",options="header " , ] according to refs .
@xcite , on the other hand , the oemd in light nuclei is strongly affected by deformation originated in the jahn - teller mechanism @xcite .
it is interesting to see what interactions contribute to @xmath97 being the main part of @xmath98 .
table iv presents the respective contributions of the @xmath99 , @xmath100 and @xmath101 forces to @xmath97 in the cr isotopes .
the dominant component is @xmath102 as expected , i.e. , about 2.0 mev for @xmath103 and about 1.2 mev for @xmath104 .
in addition , there are considerably large contributions of @xmath105 and @xmath106 .
since the @xmath100 correlation is intimately related to the nuclear deformation , the positive contribution of @xmath105 is consistent with the conclusion of ref .
the contribution of quadrupole pairing force @xmath101 , however , is negative , and is larger than that of the @xmath100 correlation for @xmath103 .
this is easily understood by the fact that the quadrupole pairing correlation breaks the @xmath12 cooper pairs of neutrons .
the present calculation tells us that the quadrupole pairing correlation probably cancels the effect of the @xmath100 correlation or the deformation on the oemd value @xcite .
figure 10 shows calculated and experimental values of @xmath107 as a function of @xmath108 for the ca , ti , cr , and fe isotopes . the agreement with experiments is quite good except for the ti isotopes .
the calculation reproduces the hill near @xmath5 and @xmath109 and also the gentle behavior of @xmath107 near the value 1.6 mev in the region of @xmath110 and @xmath111 .
the increase of @xmath107 at @xmath5 and @xmath112 is explained in terms of the same mechanism as that of @xmath113 at @xmath5 , which is caused by the coincident bends at @xmath5 of the two graphs in fig .
9 that illustrate the variations of the @xmath0 interaction energies .
the isovector p - n interactions of the @xmath114 force and the isoscalar @xmath0 force @xmath115 are important at @xmath5 also for @xmath107 .
we have investigated several quantities related to the nuclear binding energy in the previous sections .
the calculated data include binding energies of nuclei close to the proton drip line .
it is interesting to look at the two - proton separation energy , experimental data of which has been accumulated by the radioactive beam .
it provides the possibility for studying new decay modes such as diproton emission .
some nuclei around @xmath116ni are expected to possibly be two - proton emitters .
there is a large deviation between theory and experiment for the two - proton separation energy up to now .
all the predictions by the hartree - fock ( bogoliubov ) and relativistic hartree - fock ( bogoliubov ) treatments @xcite underestimate the two - proton separation energies at @xmath5 nuclei .
this discrepancy could be due to the lack of p - n interactions in these treatments . as seen in the oemd in sec .
v , the p - n correlations cooperate with the @xmath16 and @xmath17 correlations especially at @xmath5 nuclei .
we can expect that the @xmath0 interactions have a considerable influence on the two - proton separation energy .
we calculated the two - proton separation energies @xmath117 for the @xmath3 shell nuclei with @xmath118 and @xmath119 . in the calculation , the force strength @xmath120 is chosen as @xmath121 which was used in the previous paper @xcite , because the 1/a dependence of @xmath120 improves the binding energy and is also supported by the discussion in sec .
iii . calculated values of @xmath117
are compared with experimental data taken from ref .
@xcite in fig . 11 .
here we subtracted the coulomb energy following caurier @xmath71 @xcite .
the agreement is good .
the observed values of @xmath122 are reproduced well at @xmath5 nuclei .
again , the isoscalar @xmath0 force @xmath123 plays an important role in the two - proton separation energy at @xmath5 as well as in the other quantities discussed above .
the model space @xmath49 and the set of parameters used are not appropriate for nuclei with large @xmath65 and @xmath124 , strictly speaking . according to the experience in this paper , however , the results on the _ different quantities of binding energies _ might be still meaningful .
we note the prediction of our calculation that @xmath116ni , @xmath125co and @xmath126ni are possibly unstable , @xmath127ni is bound , and @xmath117 of @xmath128fe and @xmath116co are close to zero .
we have studied the @xmath0 interactions using the functional effective interaction with four force parameters which reproduces the energy levels and binding energies of @xmath6 nuclei considerably well .
first , we analyzed the double differences of binding energies @xmath72 and @xmath70 , because the two quantities are expected to directly represent the @xmath0 interactions .
our effective interaction reproduces fairly well the experimental values of @xmath72 and @xmath70 , and their characteristic behaviors , in the @xmath2 and @xmath3 shell nuclei .
the staggering of @xmath72 is due to the competition between @xmath129 and @xmath130 components .
the large spike of @xmath72 at @xmath5 is attributed to the @xmath129 @xmath0 interactions of the @xmath114 foce , and that of @xmath70 at @xmath5 contrarily represents the @xmath130 @xmath0 interactions .
the observed values of @xmath70 with respect to mass @xmath46 are approximated by the curve @xmath76 .
this curve may be explained by granting an @xmath46 dependence on the @xmath130 p - n force strength @xmath120 .
second , our effective interaction has also reproduced well the symmetry energy and the wigner energy for the @xmath3 shell nuclei . the strong @xmath130 @xmath0 force @xmath115 with assistance of the @xmath129
, @xmath12 pairing force is important to explain the magnitudes of the two quantities in this region .
it should be noted that the isospin parts of the two forces are proportional to @xmath9 and their sum is directly related to the symmetry energy in the mass .
the @xmath46 dependence of the symmetry energy coefficient seems to be determined mainly by that of @xmath120 .
third , our effective interaction has described well the observed values of the odd - even mass difference ( @xmath98 and @xmath131 ) for the @xmath3 shell nuclei .
the cooperation of the @xmath0 interactions with the like - nucleon interactions is remarkable at @xmath5 .
it causes the rise of @xmath98 and @xmath131 at @xmath5 .
the characteristic behaviors of the @xmath0 interaction energies at @xmath5 ( see fig . 9 ) have an important effect not only on the double differences of binding energies but in the odd - even mass difference .
we have briefly touched on the two - proton separation energy @xmath117 using the calculated binding energies .
the calculation indicates a considerably large effect of the @xmath0 force on @xmath117 at @xmath5 .
we noted the prediction of our calculation for @xmath117 near the @xmath3 proton drip line .
the present investigations have shown that the @xmath0 interactions cause the notable behaviors of the observed quantities related to the binding energy near @xmath5 in nuclei where valence protons and neutrons occupy the same shells .
furthermore , our calculations suggest that the @xmath0 interactions ( especially @xmath132 ) are important for describing these quantities over a wide range of @xmath133 nuclei including the neutron drip line .
all the results support the usefulness of the functional effective interaction composed of the four forces @xmath99 , @xmath100 , @xmath101 and @xmath115 , and clarify the essential roles of @xmath115 in the physical quantities related to the binding energy .
the present calculations , however , have been carried out in the single @xmath53 shell model .
calculations in more realistic model spaces are in progress .
99 a. l. goodman , adv .
11 , 263(1979 ) .
j. j@xmath134necke and h. behrens , phys .
c 9 , 1276(1974 ) .
a. s. jensen , p. g. hansen , and b. jonson , nucl .
phys . a 431 , 393(1984 ) .
j. j@xmath134necke , e. comay , nucl .
phys . a 436 , 108(1985 ) .
g. g. dussel , r. j. liotta , and r. p. j. perazzo , nucl .
a388 , 606(1982 ) .
zhang , r. f. casten , and d. s. brenner , phys .
b 227 , 1(1989 ) .
g. zaochun and y. s. chen , phys .
c 59 , 735(1999 ) .
d. s. brenner , c. wesselborg , r. f. casten , d. d. warner , and j. -y .
zhang , phys .
b 243 , 1(1990 ) .
k. t. hecht , phys .
b 139 ( 1965 ) 794 ; nucl . phys . a 102 , 11(1967 ) .
j. engel , k. langanke , and p. vogel , phys .
b 389 , 211(1996 ) .
k. langanke , d. j. dean , s. e. koonin , and p. b. radha , nucl .
phys . a 613 , 253(1997 ) .
w. satula and r. wyss , phys .
b 393 , 1(1997 ) .
w. satula , d. j. dean , j. gary , s. mizutori , and w. nazarewicz , phys .
b 407 , 103(1997 ) .
j. engel , s. pittel , m. stoitsov , p. vogel , and j. dukelsky , phys .
c 55 , 1781(1997 ) .
d. rudolph , et al .
76 , 376(1996 ) .
d. j. dean , s. e. koonin , k. langanke , and p. b. radha , phys .
b 399 , 1(1997 ) .
k. kaneko and jing - ye zhang , phys .
c 57 , 1732(1998 ) .
a. poves and g. mart@xmath135nez - pinedo , phys .
b 430 , 203(1998 ) .
p. vogel , nucl - th/9805015 .
k. kaneko , m. hasegawa , and jing - ye zhang , phys .
c 59 , 740(1999 ) .
m. hasegawa and k. kaneko , phys .
c 59 , 1449(1999 ) .
a. bohr , b. r. mottelson , and d. pines , phys .
110 , 936(1958 ) .
j. bardeen , l. n. cooper , and j. r. schrieffer , phys .
108 , 1175(1957 ) .
m. manninen , j. mansikka - aho , h. nishioka , and y. takahashi , z. phys .
d 31 , 259(1994 ) .
h. h@xmath136kkinen , j. kolehmainen , m. koskinen , p. o. lipas , and m. manninen , phys .
78 , 1034(1997 ) .
w. satula , j. dobaczewski , and w. nazarewicz , phys .
81 , 3599(1998 ) .
firestone and v. s. shirley _
table of isotopes _
( wiley - interscience new york , 1996 ) .
a. de - shalit , phys .
105 , 1528(1957 ) .
j. j@xmath134necke and e. comay , nucl .
a 436 , 108(1985 ) .
e. comay and j. j@xmath134necke , nucl .
phys . a 410 , 103(1983 ) .
e. caurier , p. zuker , a. poves and g. mart@xmath135nez - pinedo , phys .
c 50 , 225(1994 ) .
j. duflo and a. p. zucker , phys .
c 52 , r23(1995 ) . j. a. nolen and j. p. schiffer , annu .
19 , 471(1969 ) .
m. hasegawa , prog .
132 , 177(1998 ) .
w. nazarewicz , nucl .
phys . a 574 , 27c(1994 ) .
w. nazarewicz et al . , phys .
c 53 , 740(1996 ) .
g. a. lalazissis and s. ramman , phys
. rev c 58 , 1467(1998 ) .
fig . 1 plots of the double differences of binding energies derived from the experimental masses in the region @xmath137 : ( a ) @xmath63 as a function of @xmath64 ; ( b ) @xmath69 as a function of @xmath64 .
the dots stand for even-@xmath46 nuclei , and the crosses for odd-@xmath46 nuclei .
the curve @xmath76 is drawn both in ( a ) and ( b ) .
2 the calculated @xmath0 interaction energies as a function of the valence - neutron number @xmath138 for ( a ) the nb isotopes and ( b ) the mo isotopes .
the open circles stand for the @xmath0 part of the interaction energy @xmath139 , the open squares for the @xmath7 @xmath0 interaction energy @xmath89 , and the diamonds for the total @xmath0 energy .
5 the double differences of binding energies for the ti and cr isotopes , shown in the same manner as figs . 3 and 4 .
the solid squares denote the experimental values , and the open circles and crosses denote the calculated values with @xmath143 and @xmath144 , respectively .
6 the symmetry energy coefficients @xmath145 in the @xmath3 shell region .
the calculated and experimental values are denoted by the open circles and solid squares , respectively .
the diamonds represent @xmath146 in the @xmath12 isovector plus @xmath147odd isoscalar pairing force model , and the crosses @xmath148 in the @xmath12 isovector pairing force model .
11 the two - proton separation energy @xmath117 in the @xmath3 shell region .
the calculated and experimental values are denoted by the open circles and solid squares , respectively .
the force strength @xmath57 is taken as @xmath151 . | proton - neutron ( @xmath0 ) interactions and their various aspects in @xmath1 nuclei of @xmath2- and @xmath3 subshell are studied using a schematic model interaction with four force parameters proposed recently .
it is shown that the model interaction well reproduces observed physical quantities : the double differences of binding energies , symmetry energy , wigner energy , odd - even mass difference and separation energy , which testifies to the reliability of the model interaction and its @xmath0 interactions .
first of all , the double differences of binding energies are used for probing the p - n interactions .
the analysis reveals different contributions of the isoscalar and isovector @xmath0 pairing interactions to two types of double differences of binding energies , and also indicates the importance of a unique form of isoscalar @xmath0 pairing force with all @xmath4 components .
next , it is shown that this p - n force is closely related to the symmetry energy and the wigner energy .
other calculations demonstrate the significant roles of @xmath0 interactions in the odd - even mass difference and in the separation energy at @xmath5 .
pacs number(s ) : 21.10.dr;21.10.hw;21.60.-n;21.60.cs |
cern na62 experiment @xcite aims to measure the ultra - rare charged kaon decay @xmath4 ( branching fraction @xmath5 @xcite with a 10% accuracy .
the detector should efficiently suppress all the backgrounds with a signature similar to the one of the main decay channel .
one of the main backgrounds to be suppressed is coming from the abundant @xmath6 decay .
the main task of the na62 rich detector is to separate pions from muons in the momentum range 1535 gev / c with a @xmath3-misidentification factor better than @xmath0 .
in addition to that , it should provide the crossing time of the pion with a resolution better than 100 ps , and an input to level 0 and level 1 of the trigger .
the important information given by the rich is the number of the photomultiplier ( pm ) hits , the ring radius , the center of the ring , and the time of the charged particle .
the test resolution , the light collection technique and , even if biased by the small number of pms , the cherenkov angle and the track angular resolutions were checked in a test beam in 2007 @xcite .
the pion muon separation capability of the rich detector prototype equipped with about 400 pms ( rich-400 prototype ) was tested in a dedicated run in 2009 @xcite .
the prototype was built as a long tube ( about 18 m , and a diameter of about 60 cm ) filled with neon ( ne ) at atmospheric pressure and room temperature , equipped with a spherical mirror ( 17 m focal length ) at the downstream end and with 414 pms at the upstream end .
the detector response was tested using a positive hadron beam of variable momentum ( in the range 10 - 75 gev / c ) , with a spread of about 1% and a divergence of about 30 @xmath7 . at the prototype position it was composed of @xmath8
, @xmath9 , p , @xmath10 with fractions changing with the selected momentum , and a small amount of @xmath11 from @xmath9 decays .
two different mirrors were tested , both produced by the marcon company .
one of the mirrors was aluminized and coated with mgf2 by the supplier in 2008 , while the latter was aluminized and coated with sio2 and hfo2 at cern in 2009 in an attempt to improve the total mirror reflectivity .
a small difference of about 1% in favour of the new mirror in the number of hit pms was observed .
the rich operational range starts above a wavelength of 190 nm and above this value no visible absorption of the light is expected due to the contaminants .
however , the impurities change the refractive index of the radiator which could affect the cherenkov angle and its resolution . some tests with oxygen and @xmath2 fractions up to 1% were done to check the performance of the detector .
the data were taken at different momenta , with a beam centered on the mirror at a position ( x = 0 , y = 0 ) . during 2009 ,
several runs with polluted neon were taken as a part of the rich-400 test beam programme .
two of them were taken with an oxygen pollution of 380 ppm and 425 ppm . in these runs ,
the old mirror produced by marcon was used . after changing the mirror with a new one , with improved aluminized coating ,
new tests with polluted ne were performed .
the gas used for the pollution was @xmath2 .
four runs were taken for both 0.5% pollution and 1.0% pollution with @xmath2 .
the beam momentum of the charged particle for these special runs was fixed at 20.0 , 26.5 , 35.0 , and 46.3 gev / c .
the data were compared to a dedicated mc simulation based on geant 4 .
all the known efficiencies were taken into account : quartz window transparency , pm quantum efficiency , mirror reflectivity , winston cones efficiency .
a detailed geometry description is implemented . in the simulation ,
the rich vessel is put in air , and the pion or the electron beam at a given momentum is shot about 6 meters before its window .
the mirror reflectivity implemented in the simulation is the one of the old mirror .
this is an important remark , as the runs with @xmath2 pollution are taken with the new mirror .
the two data runs with oxygen pollution of 380 ppm and 425 ppm were taken at 46.3 gev / c pion momentum . as it can be seen in fig .
[ fig:1 ] , these two measurements are hardly distinguishable from the non - polluted data runs .
it can be concluded that the oxygen contamination did not affect the number of pm hits , and therefore there is no absorption effect for these amounts of oxygen .
the data and mc simulation distributions of the ring radius and of the hit multiplicity for non - polluted ne were compared , in order to validate the mc simulation . in fig .
[ fig:4 ] , the number of hit pms is plotted versus the ring radius for pions and electrons at 20.0 , 23.4 , 26.5 , 28.7,31.0 , 35.0 , 38.0 , 41.0 , 46.3 am 50.3 gev / c beam momentum .
the runs with ne contaminated with @xmath2 were taken at 20.0 , 26.5 , 38.0 and 46.3 gev / c .
the different momenta correspond to different ring radius for pions ( the ring radius increases with the momentum ) , but for electrons , as @xmath12 , the ring radius is always the same equals 1 , the ring radius for electrons is always the maximum possible one , and they are clearly distinguishable . ] .
the ring radius is @xmath13 , where @xmath14 is the focal length of the spherical mirror , and @xmath15 is the cherenkov angle ( the angle between the emitted cherenkov radiation and the particle path ) .
the cherenkov angle is related to the velocity by @xmath16 , where @xmath17 is the refractive index of the medium , and @xmath18 is the velocity of the charged particle . the data taken with and without pollution , together with the mc simulation are plotted in fig .
[ fig:4 ] . the number of hit pms increases with the ring radius ( and with the momentum ) . the data with the non - contaminated radiator is used as a reference of the quality of the overall data and mc simulation agreement
the effect of the pollution on the ring radius is clearly visible : for 1% of @xmath2 in the radiator gas the ring radius is 2.7% larger ( measured with electrons ) . for the mc simulation of the @xmath2 polluted data , a new radiator , with a new refractive index was introduced in the geant 4 simulation , according to the amount of @xmath2 used for the different runs .
the refractive index @xmath17 is calculated as @xmath19 for ne , @xmath20 and @xmath21 , while for @xmath2 these values are @xmath22 @xmath23 @xcite . in case of mixing ne with @xmath2 , a linear combination was used for the final refractive index @xmath24 where @xmath25 is the amount of @xmath2 used , and @xmath26 , where @xmath27 is the wavelength from 190 to 620 nm @xcite ] .
while for data and mc simulation the agreement on the ring radius for the positrons data with no pollution in ne , and with 1.0% @xmath2 pollution , in fig .
[ fig:5 ] is good , the disagreement on the ring radius for positrons data and mc taken with 0.5% @xmath2 pollution is clearly visible .
one of the reasons for such disagreement of the ring radii could be a beam with a slightly different momenta than the nominal one .
however , this would not give effect for positrons for which @xmath12 .
from the pion data ( fig . [ fig:4 ] ) , the biggest discrepancies for the ring radii of the pions is also found for the 0.5% @xmath2 pollution data . in order to explain this mismatch ,
a hypothesis was made , that the amount of the pollution is different from the measured one .
a test of the correlation of the ring radius and the amount of @xmath2 in the ne was made by using mc simulation .
the results are shown in fig .
[ fig:6 ] .
the correlation of the @xmath2 pollution and the ring radius was measured for positrons ( see fig .
[ fig:6 ] , right ) and for pions at 35 gev / c ( see fig .
[ fig:6 ] , left ) .
this analysis has shown that the discrepancy in the measured ring radius can be explained with a bigger amount of @xmath2 in the ne and that it is more likely that the ne gas is contaminated with 0.6% @xmath2 .
this is compatible with the fact that during the test beam , the amount of @xmath2 filled in the vessel was controlled by hand as the less significant bit on the apparatus scale was 0.1% . in fig .
[ fig:7 ] , mc data for 0.6% @xmath2 are plotted , and a better agreement of the test beam data with the simulation is visible . as a final step to improve the agreement ,
the data with the new mirror for the non polluted runs are added to the plot ( see fig .
[ fig:8 ] ) .
as mentioned before , the efficiency of the old mirror is used in the simulation . the comparison between data and mc simulation
is done for runs taken with the old mirror .
however , the @xmath2 polluted runs were taken with the new mirror . by adding the non - polluted runs with the new mirror
, it is shown that the new mirror has a slightly better reflectivity efficiency . therefore ,
if this efficiency is included in the simulation instead the efficiency of the old mirror , we can expect that the number of hit pms for the polluted data will be slightly increased leading to a better agreement of data and mc for these runs , especially for lower momenta .
the effect of the pollution on the ring radius resolution is shown in fig .
[ fig:9 ] .
the data for 0 , 0.5% , and 1.0% @xmath2 are plotted with black , red and green dots , correspondingly .
then , two sets of mc simulation were added : one generated with beam momentum resolution of 1% ( empty circles with the same color scheme for @xmath2 pollution as above ) , and one generated with beam momentum resolution of 1.7%(black , red and green triangles ) .
the runs used for this study are at 20.0 , 26.5 , 38.0 and 46.3 gev / c .
no absorption effect on the number of pm hits is observed in several runs taken with polluted ne gas radiator during the rich-400 testbeam .
however , an effect on the ring radius is clearly observed . for a radiator pollution with 1.0% of @xmath2 , an effect of 2.7% on the ring radius
this observation is confirmed by mc simulation .
the data which were assumed to be taken with 0.5% of @xmath2 are well reproduced by mc simulation with ne radiator with 0.6% contamination of @xmath2 .
the measurement of the amount of the @xmath2 could be affected by the limited apparatus precision .
the conclusion from these studies is that it is important to know the amount of @xmath2 in the final detector . in order to easily apply a correction to the ring radius
, @xmath28 data can be used for additional calibration . if the amount of @xmath2 is not known , the quality of the pion identification could be seriously compromised .
g. anelli et al . , _ proposal to measure the rare decay @xmath29 _ at the cern sps , cern - spsc-2005 - 013 , cern - spsc - p-326 ( 2005 ) .
j. brod and m. gorbahn , _ electroweak corrections to the charm quark contribution to @xmath29 _ , phys .
d 78 ( 2008 ) 034006 .
g. anzivino et al .
, _ construction and test of a rich prototype for the na62 experiment _ , nim a 593 ( 2008 ) 314 b. angelucci et al._pion - muon separation with a rich prototype for the na62 experiment . _ , nim.a621:205 - 211,2010 .
landolt bornstein , _ eigenschaften der materie in ihren aggregatzustanden 8 teil : optische konstanten _ , berlin 1962 na62 collaboration , _
na62 technical design _ , na62 - 10 - 07 , 2010 . | the na62 rich detector is used for the separation of pions and muons in the momentum range 15 35 gev / c and is expected to provide a muon suppression factor better than @xmath0 .
a prototype of the final detector equipped with about 400 pms ( rich-400 prototype ) was built and tested in a dedicated run in 2009 .
the @xmath1 separation was tested , as well as the effect of the pollution of the neon radiator with different amounts of oxygen and @xmath2 .
the @xmath3 misidentification probability is about 0.7% and the time resolution better than 100 ps in the whole momentum range .
we did not observe any absorption of the light due to the pollution of the radiator , however an effect on the ring radius is clearly observed due to the change of the change of the refractive index of the medium .
the conclusion of the studies is that the amount of @xmath2 in the final detector should be well known or the quality of the pion identification could be seriously compromised . |
we also use the infinite density matrix renormalization group ( idmrg ) @xcite to study this model . in the idmrg algorithm , we first start from a small system size
. then we insert one column in the center and optimize the energy by sweeping the inserted column .
after the optimization , we absorb the new column into the original existing system and get the new boundary hamiltonians .
we repeat the inserting , optimizing and absorbing procedure until the energy convergence is achieved . compared with the finite dmrg simulation
, idmrg grows the lattice by one column at each iteration and only sweeps the inserted column , thus the computation cost is significantly reduced .
idmrg is especially efficient to deal with the gapped topological order system , which allows us to obtain the ground states with well - defined anyonic flux as first proposed in ref .
@xcite . in our work , we have confirmed that the idmrg obtains the fixed - point ground state wavefunction in the center of cylinder that is exactly the same as that obtained from the finite dmrg simulations ( the same energy and the identical entanglement spectrum within the numerical error . ) .
there are two kinds of cylinder geometry on kagome lattice often being studied in dmrg , yc - geometry in sfig .
[ geo](a ) and xc - geometry in sfig .
[ geo](b ) . in the main text ,
the demonstrated results are all based on yc - geometry . here
we show that the entanglement spectrum and the spectrum flow shown on yc - geometry are robust and insensitive to the lattice geometry . in sfig .
[ flow_xc ] , we demonstrate the entanglement spectrum flow for @xmath115 on @xmath62 xc - geometry
. the features of the spectrum flow are consistent with the results on yc - geometry shown in fig .
3(a ) of the main text .
the eigenvalues in @xmath116 and @xmath72 sectors are degenerate at @xmath71 . by increasing flux @xmath28 , the eigenvalues in @xmath116 sector flow up while those in @xmath72 sector flow down continuously . at @xmath52 ,
the eigenvalues in the @xmath53 and @xmath72 sectors become degenerate , which results in a fractionally quantized chern number @xmath20 .
thus , the phase diagram shown in the main text is robust for different geometries . on @xmath78 yc cylinder , which is obtained by keeping @xmath117 states .
the red arrow denoted as @xmath82 indicates the two lowest values in @xmath118 and @xmath119 sectors that are used to calculate the chiral velocity .
( b ) near degenerating pattern for the low - lying entanglement spectra with different relative momentum @xmath38 and total spin @xmath39 for the same system in ( a ) .
@xmath38 is in unit of @xmath120 . ]
entanglement spectrum resembles the edge excitation spectrum that can be viewed as a fingerprint of topological order . in the main text
, we show the entanglement spectrum at @xmath121 in figs .
1(c ) and 1(d ) .
the characteristic chiral edge spectrum indicates the chiral spin liquid ( csl ) state .
here we show the spectrum at @xmath79 on yc - geometry , which is closer to the phase boundary @xmath122 . as shown in sfig .
[ es ] , the entanglement spectrum also exhibits the degeneracy pattern \{1,1,2,3,5 , ... } consistent with the csl state .
entanglement spectrum gives not only the characteristic degeneracy pattern of the edge excitation , but also the edge - mode velocity @xcite . in the conformal field theory ,
the edge - mode of the laughlin state is described by a single branch of chiral charged bosons .
the velocity of the charged bosons @xmath82 is not an universal quantity as it depends on the microscopic interaction and edge confinement . in the cylinder geometry
, we can define @xmath82 through the lowest values of entanglement spectrum with edge momentum @xmath118 and @xmath123 : @xmath124 , where @xmath125 is the lowest eigenvalue with edge momentum @xmath126 .
thus , from our results on yc - geometry , we have @xmath127 .
although the csl state is still robust at @xmath128 , the edge - mode velocity is reduced compared to @xmath129 in the deep csl phase . with further decreasing @xmath83
, we find that the difference of the lowest value between the momentum sectors @xmath118 and @xmath130 continuously decreases before the quantum phase transition takes place .
the reducing of chiral velocity obtained from entanglement spectrum is related with the drop of the bulk excitation gap with decreasing @xmath21 , which is consistent with a very weakly first order transition or a continuous transition driven by the collapsing of the bulk gap and the destruction of the gapless edge states at the same time . | the topological quantum spin liquids ( sl ) and the nature of quantum phase transitions between them have attracted intensive attentions for the past twenty years . the extended kagome spin-1/2 antiferromagnet emerges as the primary candidate for hosting both time reversal symmetry ( trs ) preserving and trs breaking sls based on density matrix renormalization group simulations . to uncover the nature of the novel quantum phase transition between the sl states , we study a minimum xy model with the nearest neighbor ( nn ) ( @xmath0 ) , the second and third nn couplings ( @xmath1 ) .
we identify the trs broken chiral sl ( csl ) with the turn on of a small perturbation @xmath2 , which is fully characterized by the fractionally quantized topological chern number and the conformal edge spectrum as the @xmath3 fractional quantum hall state . on the other hand , the nn xy model ( @xmath4 ) is shown to be a critical sl state adjacent to the csl , characterized by the gapless spin singlet excitations and also vanishing small spin triplet excitations .
the quantum phase transition from the csl to the gapless critical sl is driven by the collapsing of the neutral ( spin singlet ) excitation gap . by following the evolution of entanglement spectrum , we find that the transition takes place through the coupling of the edge states with opposite chiralities , which merge into the bulk and become gapless neutral excitations . the effect of the nn spin-@xmath5 coupling @xmath6 is also studied , which leads to a quantum phase diagram with an extended regime for the gapless sl .
quantum spin liquid ( sl ) is an exotic state of matter which escapes from forming the conventional orders even at zero temperature @xcite .
however , different from a featureless insulator , a sl develops a topological order @xcite with fractionalized quasiparticles encoded in the long - range entanglement of system @xcite .
the sl physics may play a fundamental role for understanding strongly correlated systems and unconventional superconductivity @xcite .
there have been intensive studies searching for sl in frustrated magnetic systems , however , the discovery of sl has been rare in the past 20 years .
a few frustrated square or honeycomb lattices spin systems with competing interactions have been proposed as the candidates for gapped sl @xcite .
however , further studies find that the competing plaquette valence - bond solid may dominate the magnetic disorder region @xcite .
interestingly , the nearest neighbor ( nn ) dominant spin-@xmath7 kagome heisenberg model has been identified to host a gapped sl based on the state of art density matrix renormalization group ( dmrg ) calculations @xcite , where a near quantized topological entanglement entropy @xcite has been found consistent with a @xmath8 sl @xcite .
the topological degeneracy as a signature evidence for such a gapped topological state @xcite has not been established , while different methods have been applied to tackle this problem @xcite .
meanwhile , the variational studies find that the dirac gapless sl has the lower variational energy among different states based on the projected fermionic parton wavefunctions @xcite .
the nature of the sl in the kagome heisenberg model remains not fully understood . by introducing the second- and third- nn couplings for the spin-@xmath7 kagome systems ,
dmrg studies @xcite discover the kalmeyer - laughlin csl theoretically predicted more than 20 years ago @xcite , which spontaneously breaks trs and is identified as the @xmath9 fractional quantum hall ( fqh ) state @xcite .
interestingly , the csl state is also found in the spin anisotropic kagome model involving the second and third nn @xmath10-plane couplings @xcite , or by introducing the trs breaking three - spin chiralities interactions @xcite .
however , the nature of the quantum phase transition , especially how the quantum state and entanglement spectrum ( es ) evolve near such a transition have not been addressed .
we do not know what a physical mechanism can drive the quantum phase transition in such a system , and if the emergence of the previously identified gapped sl for the nn kagome heisenberg model has close connection with the collapsing of the csl @xcite .
our work is motivated to address these open questions . along with theoretical developments
, experiments also discover different promising sl candidates in the triangular organic compounds @xcite and kagome antiferromagnets herbertsmithite and kapellasite cu@xmath11zn(oh)@xmath12cl@xmath13 in recent years @xcite .
these materials appear to have gapless excitations as revealed by the specific heat and neutron scattering measurements @xcite .
thus , it would be extremely interesting to also search for some minimum spin-@xmath7 kagome model which can host a gapless sl . in this letter
, we address the nature of the collapsing of csl and the related phase transition in kagome spin system based on dmrg and exact diagonalization ( ed ) calculations .
we study the spin-@xmath7 xxz kagome model with the spin xy interactions for the second and third neighbors as shown in the inset of fig .
[ fig1](a ) , whose hamiltonian is given as @xcite @xmath14 where the summations are taken over the nn @xmath15 , the second and third nn @xmath16 couplings .
we set @xmath17 as energy scale .
our main results are summarized as the phase diagram in fig .
[ fig1](a ) .
first of all , for the xy model with @xmath18 , we establish a csl for @xmath19 based on the topological features of the state : the conformal chiral edge spectrum in accordance with the @xmath3 fqh state and the topological quantized chern number @xmath20 .
we identify the physical driving force for the destruction of the csl as the collapsing of the singlet excitation gap with reducing @xmath21 . following the evolution of es , we find that the phase transition takes place through the coupling between the low - lying entanglement states with opposite chiralities , which naturally leads to a critical state with trs and gapless neutral excitations @xcite .
our results represent a significant progress in understanding the connection between different sls @xcite by identifying the mechanism of the phase transition and establishing the characteristic nature of the critical sl phase adjacent to the csl . after tuning on the nn spin-@xmath5 coupling @xmath6
, we identify the phase diagram for different @xmath21 , where the critical gapless sl is found for an extended regime with small @xmath22 .
the connection of the critical sl with the previously identified gapped sl in heisenberg model @xcite will also be discussed .
we use dmrg @xcite and ed to study cylinder and torus systems with the geometry shown in the inset of fig .
[ fig1](a ) .
the number of sites in cylinder ( open boundary condition in the @xmath23-direction ) or torus system is @xmath24 with @xmath25 and @xmath26 as the numbers of unit cell in the @xmath23 and @xmath27 directions @xcite .
we perform the flux insertion simulations on cylinder systems based on the newly developed adiabatic dmrg to detect the topological chern number @xcite . in this simulation
, we thread a flux @xmath28 in the cylinder which is equivalent to imposing the twist boundary conditions : @xmath29 for these bonds crossing the @xmath27 boundary . in dmrg studies we keep up to @xmath30 states for the simulations without flux and @xmath31 states in the flux insertion simulations for accurate results
. kagome model with hamiltonian eq .
( [ hamiltonian ] ) . for @xmath32 , @xmath33 , where both the csl and a critical sl with gapless singlet excitations are identified .
the color gradient in the critical sl denotes the growing spin gap on finite - size system with increasing @xmath6 .
the gapped sl @xcite may exist neighboring with the critical sl .
( b ) log - linear plot of chiral correlation @xmath34 versus the distance of triangles @xmath35 along the @xmath23 direction in the vacuum sector ( c ) and ( d ) are the es of the groundstates in the vacuum and spinon sectors , respectively .
@xmath36 is the eigenvalue of reduced density matrix .
the numbers @xmath37 label the near degenerating pattern for the low - lying es with different relative momentum @xmath38 and total spin @xmath39 .
, title="fig : " ] kagome model with hamiltonian eq .
( [ hamiltonian ] ) . for @xmath32 , @xmath33 , where both the csl and a critical sl with gapless singlet excitations are identified .
the color gradient in the critical sl denotes the growing spin gap on finite - size system with increasing @xmath6 .
the gapped sl @xcite may exist neighboring with the critical sl .
( b ) log - linear plot of chiral correlation @xmath34 versus the distance of triangles @xmath35 along the @xmath23 direction in the vacuum sector ( c ) and ( d ) are the es of the groundstates in the vacuum and spinon sectors , respectively .
@xmath36 is the eigenvalue of reduced density matrix .
the numbers @xmath37 label the near degenerating pattern for the low - lying es with different relative momentum @xmath38 and total spin @xmath39 .
, title="fig : " ] kagome model with hamiltonian eq .
( [ hamiltonian ] ) . for @xmath32 , @xmath33 , where both the csl and a critical sl with gapless singlet excitations are identified .
the color gradient in the critical sl denotes the growing spin gap on finite - size system with increasing @xmath6 .
the gapped sl @xcite may exist neighboring with the critical sl .
( b ) log - linear plot of chiral correlation @xmath34 versus the distance of triangles @xmath35 along the @xmath23 direction in the vacuum sector ( c ) and ( d ) are the es of the groundstates in the vacuum and spinon sectors , respectively .
@xmath36 is the eigenvalue of reduced density matrix .
the numbers @xmath37 label the near degenerating pattern for the low - lying es with different relative momentum @xmath38 and total spin @xmath39 .
, title="fig : " ] _ chiral spin liquid phase. _ the csl breaks trs but preserves lattice symmetries and spin rotational symmetry .
the trs broken is usually detected by the chiral order parameter @xmath40 ( @xmath41 triangle ) @xcite .
as shown in fig .
[ fig1](b ) , we demonstrate the chiral correlations @xmath42 between the up - triangles @xmath43 and @xmath44 as a function of distance @xmath35 for the xy model ( @xmath45 ) . at @xmath46 , the chiral correlations decay exponentially to vanish . with growing @xmath47 ,
the chiral correlations enhance gradually and appear to approach finite values for @xmath48 at large distance , which indicates the emerging long - range chiral order that characterizes the spontaneous trs breaking .
moreover , csl is a topological ordered state that hosts two - fold topological degenerate groundstates , which can be obtained by inserting flux with @xmath49 adibatically @xcite .
the es for these two states , labeled by the quantum number total @xmath39 of the half system , and their relative momentum quantum number along the @xmath27 direction @xmath38 @xcite , are shown in figs .
[ fig1](c ) and [ fig1](d ) .
the leading es has the robust degeneracy pattern @xmath37 with increasing @xmath38 in each @xmath39 sector , which follows the chiral @xmath50 wess - zumino - witten conformal field theory description of the @xmath9 fqh state @xcite as the fingerprint for the emergence of the csl @xcite .
the spectra of the vacuum ( @xmath51 ) and spinon ( @xmath52 ) sectors are symmetric about @xmath53 and @xmath54 respectively , indicating a spin-@xmath55 spinon at each end of cylinder in the spinon sector . at @xmath56 after adiabatically inserting a flux @xmath57 .
the area of the circle is proportional to the amplitude of @xmath58 .
the blue ( red ) color represents the positive ( negative ) @xmath58 .
( b ) accumulated spin magnetization @xmath59 in each column ( the summation is over all the @xmath60 sites for each column ) with increasing flux @xmath28 .
( c ) net spin transfer @xmath61 as a function of @xmath28 on @xmath62 cylinder .
the inset shows the results on @xmath63 cylinder .
( d ) phase diagram of the xy kagome model ( @xmath18 ) , which is determined from the results of chern number on @xmath62 system .
, title="fig : " ] + at @xmath56 after adiabatically inserting a flux @xmath57 .
the area of the circle is proportional to the amplitude of @xmath58 .
the blue ( red ) color represents the positive ( negative ) @xmath58 .
( b ) accumulated spin magnetization @xmath59 in each column ( the summation is over all the @xmath60 sites for each column ) with increasing flux @xmath28 .
( c ) net spin transfer @xmath61 as a function of @xmath28 on @xmath62 cylinder .
the inset shows the results on @xmath63 cylinder .
( d ) phase diagram of the xy kagome model ( @xmath18 ) , which is determined from the results of chern number on @xmath62 system .
, title="fig : " ] at @xmath56 after adiabatically inserting a flux @xmath57 .
the area of the circle is proportional to the amplitude of @xmath58 .
the blue ( red ) color represents the positive ( negative ) @xmath58 .
( b ) accumulated spin magnetization @xmath59 in each column ( the summation is over all the @xmath60 sites for each column ) with increasing flux @xmath28 .
( c ) net spin transfer @xmath61 as a function of @xmath28 on @xmath62 cylinder .
the inset shows the results on @xmath63 cylinder .
( d ) phase diagram of the xy kagome model ( @xmath18 ) , which is determined from the results of chern number on @xmath62 system .
, title="fig : " ] _ fractional quantization of topological chern number. _ to reveal the full topological nature of the csl phase , we perform the flux insertion simulation to obtain the topological chern number @xcite . by adiabatically inserting flux @xmath28 , we study the evolution of the spin-@xmath5 magnetization @xmath58 at each site @xmath56 .
one example with a small @xmath64 is shown in fig . [ chern](a ) .
we find that the nonzero magnetizations start to build up at both edges of cylinder after inserting flux .
the net magnetization near boundaries grows monotonically with increasing @xmath28 as shown in fig .
[ chern](b ) , which is equivalent to the spin transfer being pumped from the left edge to the right edge without accumulating in the bulk . as shown in fig .
[ chern](c ) , a spin pump linearly increases with @xmath28 on @xmath62 cylinder , which leads to a quantized net spin transfer @xmath65 at @xmath66 and a quantized chern number @xmath67 , fully characterizing the state as the @xmath9 fqh state @xcite . for the system with @xmath63 as shown in the inset of fig .
[ chern](c ) , we find some finite size effect as the spin pump initially is zero for small @xmath28 , which jumps to the expected values of the linear pumping behavior at a larger @xmath28 for @xmath68 .
the csl is protected by the finite bulk excitation gap ( shown later in fig .
[ spectrum ] ) and grows stronger with increasing system width .
based on the quantized chern number established on @xmath62 cylinder with different geometries @xcite and the conformal edge spectrum for the groundstates , we find a robust csl phase for @xmath19 as shown in the phase diagram fig .
[ chern](d ) for @xmath45 . in the critical sl region
, we observe strong magnetization fluctuations in the bulk during the process of inserting flux in consistent with the collapsing of the neutral excitation gap . _ entanglement spectrum flow. _ the csl and the critical state can be understood based on the response of es to inserted flux @xcite . for a csl at @xmath69 , as shown in fig . [ flow](a )
, the eigenvalues of the reduced density matrix are degenerating about the @xmath70 sectors at @xmath71 . by increasing @xmath28 ,
the spectrum lines in the positive @xmath39 sectors flow up continuously , while those in the negative @xmath39 sectors flow down ( this is selected by the sign of chern number due to spontaneous trs breaking ) . at @xmath52 ,
the eigenvalues in the @xmath53 and @xmath72 sectors become degenerate . as a result , after inserting a flux quantum , a net spin transfer @xmath73 is realized and the spectrum becomes symmetric about @xmath74 .
thus the es flow directly detects the gapless feature in the edge spectrum through inserting flux . by inserting @xmath75 flux
, the es continues to flow and , it becomes symmetric about @xmath72 at @xmath76 , indicating two spinons have been transferred from the left edge to the right edge while the bulk of the system goes back to the vacuum sector .
for @xmath77 on @xmath78 cylinder .
es for ( b ) @xmath79 , ( c ) @xmath80 , and ( d ) @xmath81 on @xmath62 cylinder .
the arrow denoted by @xmath82 in ( b ) indicates the two lowest eigenvalues in the @xmath53 sector which are used to calculate the chiral edge - mode velocity.,title="fig : " ] for @xmath77 on @xmath78 cylinder .
es for ( b ) @xmath79 , ( c ) @xmath80 , and ( d ) @xmath81 on @xmath62 cylinder .
the arrow denoted by @xmath82 in ( b ) indicates the two lowest eigenvalues in the @xmath53 sector which are used to calculate the chiral edge - mode velocity.,title="fig : " ] with decreasing @xmath83 , the robust chern number quantization and the spectrum flow persist to @xmath84 . by following the evolution of the es
, we find that the csl is becoming less strong at smaller @xmath83 , where the chiral velocity ( proportional to the gap between the lowest two spectrum levels in the @xmath53 sector as indicated in fig .
[ flow](b ) ) @xcite diminishes with decreasing @xmath21 . as illustrated in figs .
[ flow](b ) and [ flow](c ) , we observe that the es as a function of quantum number @xmath39 before and after the phase transition appear to be similar at @xmath69 and @xmath85 . however , they are significantly different in momentum space .
the spectrum for @xmath86 preserves the same robust conformal chiral edge spectrum as demonstrated in fig .
[ fig1](c ) with many entanglement eigenstates carrying nonzero @xmath87 @xcite . however , once the phase transition takes place , the groundstate wavefunction has trs , and the low - lying entanglement states shown in fig .
[ flow](c ) have the momentum quantum number @xmath87 either @xmath88 or @xmath89 if they are nondegenerate , which comes from the mixing between eigenstates with opposite chiralities .
furthermore , these low - lying eigenstates do not respond to the inserted flux .
the mixing of entanglement states with opposite chiralities illustrates what happens to the physical edge states @xcite .
these edge states with opposite chiralities also mix and merge into the bulk and become the low energy gapless excitations in the bulk .
these observations are consistent with the field theory description for the quantum phase transition between two states with different chern numbers @xcite .
interestingly , the es for the nn kagome heisenberg model in fig . [ flow](d ) is similar to the one of the nn xy model .
sector with @xmath47 for the xy model ( @xmath45 ) on the @xmath90 torus .
the singlet gap between the lowest two groundstates and higher - energy states is denoted as @xmath91 .
( b ) @xmath47 dependence of spin gap @xmath92 for the xy model on the @xmath90 and @xmath93 tori .
low - energy spectrum of the nn xxz model ( @xmath94 ) at @xmath95 and @xmath96 sectors for ( c ) @xmath97 , ( d ) @xmath98 , and ( e ) @xmath99 on the @xmath100 torus .
the label @xmath101 denote the quantum numbers related to spin inversion and lattice @xmath102rotation symmetries . ] _ low - energy excitations. _ we first study the evolution of the low - energy singlet excitations in the @xmath53 sector as a function of @xmath47 for the xy model ( @xmath45 ) . as shown in fig .
[ spectrum](a ) of the spectrum for @xmath103-site torus system in @xmath95 sector , we find two low - energy near degenerate groundstates separated by a finite singlet gap @xmath91 from higher energy excitations in the csl phase at larger @xmath104 side @xcite . with decreasing @xmath47 , the singlet gap @xmath91 reduces , which collapses to vanishing small at @xmath105 .
thus the quantum phase transition from the csl to the trs preserving state is driven by such a neutral excitation gap closing .
for comparison , we also obtain the triplet gap @xmath92 in dmrg calculations using torus systems with @xmath106 and @xmath93 as shown in fig .
[ spectrum](b ) @xcite .
similarly , @xmath92 drops with reducing @xmath47 and it becomes much smaller for @xmath107 system at @xmath4 .
thus , our results indicate that the critical state is centered near @xmath94 , where the singlet gap vanishes and the spin gap is very tiny or vanished ( @xmath92 reduces with @xmath108 and @xmath109 for @xmath110 ) . the appearance of low - energy singlet excitations below the finite - size spin gap can be understood as the gapless neutral mode for the topological quantum phase transition @xcite , which necessarily exists for such a transition .
we further study the whole phase diagram with varying @xmath6 and @xmath21 , where similar csl to critical phase transition is observed as illustrated in the phase diagram fig .
[ fig1](a ) .
furthermore , we examine the low energy spectra of the nn xxz model ( @xmath94 ) on @xmath106 torus system . as shown in figs .
[ spectrum](c)-[spectrum](e ) for @xmath111 and @xmath112 , we find near continuous low energy excitations @xcite collapsing together below the spin triplet gap , which implies the gapless singlet excitations in the whole critical sl region .
the structure of the energy spectra remains very similar , which indicates that the spin interaction @xmath6 term may only enhance the energy scale of excitations .
the gapped sl @xcite may exist neighboring with the critical sl close to the nn heisenberg model and we can not determine the precise phase boundary of the critical sl due to the limited system width we can access in dmrg simulations .
_ summary and discussions. _ we identify a trs broken csl phase with a small pertubation @xmath113 in the @xmath114 xy model , while the nn xy model is in a critical phase adjacent to the csl with vanishing singlet excitations and a small or vanishing spin triplet gap based on ed and dmrg studies .
furthermore , by studying the evolution of es crossing the quantum phase transition , we identify that the quantum phase transition takes place through the coupling and mixing of the chiral states with opposite chiralities , which naturally lead to a critical state with trs and gapless neutral excitations .
the quantum phase transition appears to be very smooth , which is driven by the continuous closing of the gap for spin singlet excitations .
however , it is important to mention that limited by the range of systems one can access using dmrg , one can not determine if there is a discontinuity in the singlet gap at the transition point in thermodynamic limit . thus the quantum phase transition can be a weakly first order or continuous transition , which demands study based on effective theory for these novel sl states . finally , we find that the nn @xmath6 coupling leads to a phase diagram with an extended regime for the critical sl possibly including or close to the nn heisenberg kagome model . while the neutral excitation has to be gapless in such a critical sl , we find that the spin gap is possibly finite , but very small , which grows bigger with the increase of @xmath6 .
our dmrg calculations for spin gap on larger tori and ed calculations for singlet gap indicate that these gaps decrease towards the nn models with reducing @xmath83 , which are more consistent with a critical sl .
however , the gapped @xmath8 sl @xcite may develop on larger systems in the nn models through opening the vison gap from the gapless neutral excitations outside the critical sl .
this open and challenging question is desired to be addressed in the future based on different numerical methods and effective field theory approaches .
we hope to thank l. balents , f. d. m. haldane , l. fu , p. a. lee , t. senthil , z. y. weng , and especially x. g. wen for stimulating discussions .
we also thank y. c. he and l. cincio for their insightful discussions about developing infinite dmrg algorithm .
this research is supported by the u.s .
department of energy , office of basic energy sciences under grant
no . de - fg02 - 06er46305 ( w.z . , d.n.s . ) , the national science foundation through grants dmr-1408560 ( s.s.g ) .
99 l. balents , nature * 464 * , 199 ( 2010 ) .
x. g. wen , phys .
rev .
b * 40 * , 7387 ( 1989 ) .
x. g. wen and q. niu , phys .
rev .
b * 41 * , 9377 ( 1990 ) .
x. g. wen , int .
j. mod .
phys .
b * 4 * , 239 ( 1990 ) .
x. chen , z. c. gu , and x. g. wen , phys . rev .
b * 82 * , 155138 ( 2010 ) .
p. w. anderson , science * 235 * , 1196 ( 1987 ) .
d. s. rokhsar and s. a. kivelson , phys .
rev .
lett . * 61 * , 2376 ( 1988 ) .
n. read and s. sachdev , phys .
rev .
lett . * 66 * , 1773 ( 1991 ) .
x. g. wen , phys .
rev .
b * 44 * , 2664 ( 1991 ) .
t. senthil and m. p. a. fisher , phys .
rev .
b * 62 * , 7850 ( 2000 ) ; phys . rev
. lett . * 86 * , 292 ( 2001 ) .
r. moessner and s. l. sondhi , phys .
rev .
lett . * 86 * , 1881 ( 2001 ) .
l. balents , m. p. a. fisher , and s. m. girvin , phys .
rev .
b * 65 * , 224412 ( 2002 ) .
t. senthil and o. i. motrunich , phys .
rev .
b * 66 * , 205104 ( 2002 ) . o. i. motrunich and t. senthil , phys . rev .
lett . * 89 * , 277004 ( 2002 ) .
d. n. sheng and l. balents , phys .
rev .
lett . * 94 * , 146805 ( 2005 ) .
p. a. lee , n. nagaosa , and x. g. wen , rev .
mod .
phys .
* 78 * , 17 ( 2006 ) .
a. kitaev , ann .
phys .
( n.y . )
* 321 * , 2 ( 2006 ) .
d. f. schroeter , e. kapit , r. thomale , and m. greiter , phys .
rev .
lett . * 99 * , 097202 ( 2007 ) .
a. lee , science * 321 * , 1306 ( 2008 ) .
s. v. isakov , m. b. hastings , r. g. melko , nat .
phy . * 7 * , 772 ( 2011 ) .
h. yao and s. a. kivelson , phys .
rev .
lett . * 108 * , 247206 ( 2012 ) .
f. wang , phys .
rev .
b * 82 * , 024419 ( 2010 ) .
y. m. lu and y. ran , phys .
rev .
b * 84 * , 024420 ( 2011 ) .
b. k. clark , d. a. abanin , and s. l. sondhi , phys .
rev .
lett . * 107 * , 087204 ( 2011 ) .
h. c. jiang , h. yao , and l. balents , phys .
rev .
b * 86 * , 024424 ( 2012 ) .
r. ganesh , j. v. d. brink , and s. nishimoto , phys .
rev .
lett . * 110 * , 127203 ( 2013 ) .
z. y. zhu , d. a. huse , and s. r. white , phys .
rev .
lett . * 110 * , 127205 ( 2013 ) .
s. s. gong , d. n. sheng , o. i. motrunich , and m. p. a. fisher , phys .
rev .
b * 88 * , 165138 ( 2013 ) . s. s. gong , w. zhu , d. n. sheng , o. i. motrunich , and m. p. a. fisher , phys .
rev .
lett . * 113 * , 027201 ( 2014 ) . h. c. jiang , z. y. weng , and d. n. sheng , phys .
rev .
lett .
* 101 * , 117203 ( 2008 ) .
s. yan , d. huse , and s. r. white , science * 332 * , 1173 ( 2011 ) .
s. depenbrock , i. p. mcculloch , and u. schollwck , phys .
rev .
lett . * 109 * , 067201 ( 2012 ) .
h. c. jiang , z. h. wang , and l. balents , nat .
phy .
* 8 * , 902 ( 2012 ) .
a. kitaev , j. preskill , phys .
rev .
lett . * 96 * , 110404 ( 2006 ) .
m. levin , x. g. wen , phys .
rev .
lett . * 96 * , 110405 ( 2006 ) .
l. balents , m. p. a. fisher , and c. nayak , phys .
rev .
b * 60 * , 1654 ( 1999 ) .
y. c. he , d. n. sheng , and y. chen , phys .
rev .
b * 89 * , 075110 ( 2014 ) .
y. ran , m. hermele , p. a. lee , and x. g. wen , phys .
rev .
lett . * 98 * , 117205 ( 2007 ) .
y. iqbal , f. becca , and d. poilblanc , phys .
rev .
b * 84 * , 020407(r ) ( 2011 ) . s. s. gong , w. zhu , and d. n. sheng , sci
. rep . * 4 * , 6317 ( 2014 ) .
y. c. he , d. n. sheng , and y. chen , phys .
rev .
lett . * 112 * , 137202 ( 2014 ) .
r. b. laughlin , phys .
rev .
lett .
* 50 * , 1395 ( 1983 ) .
v. kalmeyer and r. b. laughlin , phys .
rev .
lett . * 59 * , 2095 ( 1987 ) .
x. g. wen , f. wilczek , and a. zee , phys .
rev .
b * 39 * , 11413 ( 1989 ) .
k. yang , l. k. warman , and s. m. girvin , phys .
rev .
lett . * 70 * , 2641 ( 1993 ) .
f. d. m. haldane and d. p. arovas , phys .
rev .
b * 52 * , 4223 ( 1995 ) . j. w. mei and x. g. wen , arxiv:1407.0869 .
y. c. he and y. chen , arxiv:1407.2740 .
b. bauer , l. cincio , b. p. keller , m. dolfi , g. vidal , s. trebst , a. w. w. ludwig , arxiv:1401.3017 .
m. barkeshli , arxiv:1307.8194 .
y. shimizu , k. miyagawa , k. kanoda , m. maesato , and g. saito , phys .
rev .
lett . * 91 * , 107001 ( 2003 ) .
y. kurosaki , y. shimizu , k. miyagawa , k. kanoda , and g. saito , phys .
rev .
lett . * 95 * , 177001 ( 2005 ) .
t. itou , a. oyamada , s. maegawa , m. tamura , and r. kato , phys .
rev .
b * 77 * , 104413 ( 2008 ) .
p. mendels , f. bert , m. a. de vries , a. olariu , a. harrison , f. duc , j. c. trombe , j. s. lord , a. amato , and c. baines , phys .
rev .
lett . * 98 * , 077204 ( 2007 ) .
j. s. helton , k. matan , m. p. shores , e. a. nytko , b. m. bartlett , y. yoshida , y. takano , a. suslov , y. qiu , j .- h .
chung , d. g. nocera , and y. s. lee , phys .
rev .
lett . * 98 * , 107204 ( 2007 ) .
m. a. de vries , j. r. stewart , p. p. deen , j. o. piatek , g. j. nilsen , h. m. rnnow , and a. harrison , phys .
rev .
lett . * 103 * , 237201 ( 2009 ) . b. fk , e. kermarrec , l. messio , b. bernu , c. lhuillier , f. bert , p. mendels , b. koteswararao , f. bouquet , j. ollivier , a. d. hillier , a. amato , r. h. colman , and a. s. wills , phys .
rev .
lett . * 109 * , 037208 ( 2012 ) .
t. h. han , j. s. helton , s. chu , d. g. nocera , j. a. rodriguez - rivera , c. broholm , and y. s. lee , nature * 492 * , 7429 ( 2012 ) .
l. clark , j. c. orain , f. bert , m. a. de vries , f. h. aidoudi , r. e. morris , p. lightfoot , j. s. lord , m. t. f. telling , p. bonville , j. p. attfield , p. mendels , and a. harrison , phys .
rev .
lett . * 110 * , 207208 ( 2013 ) .
i. a. mcdonald and f. d. m. haldane , phys .
rev .
b * 53 * , 15845 ( 1996 ) .
x. g. wen , phys .
rev .
lett . * 84 * , 3950 ( 2000 ) .
s. r. white , phys .
rev .
lett . * 69 * , 2863 ( 1992 ) .
see more information in supplemental material .
l. cincio and g. vidal , phys .
rev .
lett . * 110 * , 067208 ( 2013 ) .
michael p. zaletel , roger s. k. mong , and frank pollmann , phys .
rev .
lett . * 110 * , 236801 ( 2013 ) .
p. di francesco , p. mathieu , and d. senechal , _ conformal field theory _
( springer , new york , 1997 ) , chap .
15.6 .
h. li and f. d. m. haldane , phys .
rev .
lett .
* 101 * , 010504 ( 2008 ) .
y. hatsugai , phys .
rev .
lett . * 71 * , 3697 ( 1993 ) .
w. zhu , s. s. gong , and d. n. sheng , j. stat . mech .
2014 ( 8) , p08012 .
a. g. grushin , j. motruk , m. p. zaletel , and f. pollmann , arxiv:1407.6985 . x. l. qi , h. katsura , a. w. w. ludwig , phys .
rev .
lett . * 108 * , 196402 ( 2012 ) . while one can not get a few low energy singlet excitations in dmrg accurately
, one can obtain the triplet gap @xmath92 by targeting the ground states in the different total @xmath39 sectors separately . c. waldtmann , h. u. everts , b. bernu , c. lhuillier , p. sindzingre , p. lecheminant , l. pierre , eur .
phys .
j. b * 2 * , 501 ( 1998 ) . |
quantum entanglement , first noted by einstein - podolsky - rosen(epr ) @xcite and schrdinger @xcite is a genuine property of quantum mechanics .
qualitative embodiment of this property was given by bell @xcite who showed that entangled states have stronger correlations than allowed by local hidden variable theories ( lhvt ) .
later clauser and coworkers@xcite suggested a more transparent inequality ( b - chsh ) , the violation of which was experimentally demonstrated for photons@xcite . during the last decade ,
several works have been achieved in the context of mesoscopic physics to explore two particle entanglement .
the initial proposals used superconducting sources of electrons connected to two normal metal leads , through which the two constituent electrons originating from a cooper pair were split to form a singlet state outside the superconductor@xcite .
other proposals for two electrons entanglement were subsequently made using ballistic electrons and point contacts placed in the integer quantum hall effect ( iqhe ) regime@xcite .
bell inequality tests based on stationary current noise cross correlations revealed that a maximal violation could be achieved@xcite . in the context of quantum optics , greenberger , horne and zeilinger introduced a maximally entangled tripartite state , commonly called the ghz - state@xcite .
the bell parameter in the tripartite case can take values up to 4 , while the corresponding parameter in the bipartite case is no more than @xmath0@xcite .
for both cases the lhvt limit remains 2 . in this sense
the entanglement is considered to be `` stronger '' than for the two particle state , and it is thus potentially easier for experimental detection . ref . pointed out that the measurement results with this state are in conflict with local realism when quantum mechanics makes definite ( non statistical ) predictions , in contrast with epr states .
the ghz state has a number of potential applications in quantum information .
it can be used for quantum error correction @xcite , and it was proved that tripartite states have advantages compared to bipartite ones in quantum teleportation @xcite and in dense coding @xcite .
there are two established methods to produce entangled states of electrons .
the first method exploits the indistinguishability of fermions and relies on postselection to get the desired entangled state@xcite ( see also ref . for the discussion of the role of projection in this case ) .
the second way is to use the interaction between particles ( implication of this method with flying electronic qubits is described in refs . ) .
the latter has the advantage that the evolution during the preparation of the state is unitary and thus can be deterministic with up to 100% efficiency for producing the desired state .
while there exist various ways to create and manipulate entangled states of qubits in optical systems@xcite and nmr ( nuclear magnetic resonance ) experiments@xcite , in mesoscopic systems the generation of an experimentally determined entangled state of few electrons and the subsequent proof that bell inequalities have successfully violated both represent a considerable challenge . a proposal which follows faithfully the quantum optics experimental protocol@xcite uses edge states in the iqhe and achieves the ghz state using postselection .
the purpose of the present work is to go beyond this existing protocol and thus to explore three particle orbital entanglement with ballistic electrons propagating in mesoscopic devices using the interaction between electrons .
this proposal is motivated by the recent progress in achieving single electron sources@xcite and for buiding effective electron mach - zehnder interferometers ( mzis)@xcite .
the outline of the paper is as follows . in sec .
[ description_section ] , we describe the 3 mzi setup and specify how the beam splitters ( bs ) and coulomb interaction operate . in sec .
[ ghz_section ] we justify that a ghz state is produced at the outcome . in sec .
[ bell_section ] we describe the bell inequality test which is used to show that maximal entanglement is achieved . in sec .
[ detection_section ] we implement this test for our 3 mzi device . in sec .
[ sec : iqhe ] we show a possible realization of our setup in the iqhe regime .
we conclude in sec .
[ conclusion ]
three electrons are injected simultaneously in the three mzis , and propagate from the bottom to the top of the figure .
the crossing of the lines indicate a beam splitter ( bs ) , where an electron can exit in one of the two output branches .
the dotted lines for each mzi show a third bs , used to explain theoretically the detection process .
the dashed horizontal lines specify different levels of the setup which are used in the theoretical explanations.,width=283 ] a schematic drawing of the device is depicted in fig .
[ setup(f ) ] .
it consists of three mzis , which are placed side by side ( shown with full lines in fig .
[ setup(f ) ] ) .
each mzi consists of two incoming channels , which meet at a first beam splitter ( bs ) .
the two outgoing channels from these bs propagate and meet at the second bs .
the theoretical setup also has a second loop ( shown with dotted lines in fig .
[ setup(f ) ] ) , with propagation from the output channels of the second bs , and recombination at a third bs .
we will use this second loop to explain theoretically the detection of the ghz states , and we will show in sec .
[ detection_section ] that it is possible in practice to get rid of this second loop and perform the ghz state production and detection with a single loop .
we distinguish the right ( r ) and left ( l ) side of each interferometer and label single electron wave functions accordingly .
for instance , at any stage of the wave packet evolution , @xmath1 denotes an electron wave packet on the @xmath2 side ( @xmath3 ) of the @xmath4-th mzi ( @xmath5 ) .
the most likely candidates for electron channels are edge states in the iqhe regime , which have the advantage that they are immune to backscattering effects by impurities and have long phase breaking time .
several experiments involving mzi in the iqhe have already been performed @xcite , some of them involving setups with two mzis .
a proposal for a detailed setup is given in sec .
[ sec : iqhe ] .
we assume that a single electron wavepacket is emitted in each mzi above the fermi sea .
such single electron emission was recently experimentally demonstrated with the on demand single electron source@xcite which uses the mesoscopic capacitor as the injector .
the bs are assumed to be reflectionless , i.e. incident particles can not be backscattered in the same channel , they can only be transmitted further `` up '' ( in fig . [
setup(f ) ] ) .
the bs can be parametrized by a transfer matrix @xmath6 which relates incoming states to outgoing states and @xmath7 is the transparency parameter ( see fig.[fig : bs ] ) .
for example if @xmath8 the beam - splitter is transparent meaning that the incident particle goes from r to l and from l to r without scattering , while for @xmath9 the incident particle may appear in each of output channels with equal probability @xmath10 .
there exists an additional freedom in choosing the phases of scattering matrix elements but it does not affect results in a crucial manner ( see appendix [ bs_appendix ] for details ) .
here we specify the bottom beam - splitters to have equal probability @xmath10 for transmission in the @xmath11 and @xmath12 side channels ( @xmath9 ) .
the transfer - matrix is then simply : @xmath13 this choice of the bottom beam - splitters as well as @xmath14 phase shift in coulomb interaction is analogous to the studies@xcite of two mach - zehnder interferometers . in each of them
maximally entangled bipartite state was achieved when the first row of beam - splitters was half - reflecting .
the main steps in the production of the ghz - state are as follows .
the first stage is the synchronized injection of electrons ( dots with arrows at the bottom of fig .
[ setup(f ) ] ) , which then pass through the first bs of each of the 3 mzis . beyond the first bs
, free propagation occurs for electrons wave packets , until they reach the `` interaction region '' : the @xmath12 channel of the first mz and the @xmath11 channel of the second mz , and separately the @xmath12 channel of the second mz and the @xmath11 channel of the third mzi are put in close proximity so that coulomb interaction effects between the two pairs of neighbouring channels become important .
the effective length of this interaction region is such that it generates an overall @xmath14 phase shift of the two particle wave function associated with the two neighbouring channels .
it may be unclear why coulomb interaction would produce a phase shift because coulomb interaction involves energy exchange process . this matter and conditions required for the phase shift to be produced
are discussed in appendix [ coulomb interaction . ] .
after the interaction region the three electrons propagate freely towards the second bs , but we allow to insert a phase difference @xmath15 in each mz loop .
the phase differences which are needed here could be achieved via the aharonov - bohm effect or via a scalar potential .
actually aharonov - bohm and coulomb interaction phase accumulation occur together but logically the process can be decoupled into sequence of independent processes . from this point of view
the aharonov - bohm phase effectively changes the transfer matrix of the following bs ( see appendix [ bs_appendix ] ) which is convenient for further calculations .
the final stage of our setup is the measurement of the state of the three particle in the output arms ( top of fig .
[ setup(f ) ] ) . to understand the role of the interaction regions , let us consider first a given mzi with an electron injected into the right lead ( see fig .
[ one_mzi(f ) ] ) .
the electron in this interferometer could be treated as a `` flying qubit '' , i.e. a two - state system , where states are specified by the side chosen in the interferometer , where the electron travels .
the initial state therefore is @xmath16 .
after the bottom beam - splitter , at the level `` lv1 '' ( see fig . [
one_mzi(f ) ] ) , the state of the electron is @xmath17 next , we add a wire coupled to the interferometer with an electron injected simultaneously .
so the initial state is @xmath18 .
when two electrons pass the interaction zone they accumulate mutual phase @xmath14 . after interaction
the two - particle state remains separable , and each particle has a wave - function . if the interaction occurs we get at the level `` lv2 '' for the left electron @xmath19 if we do not inject an electron into the right lead then at the level `` lv2 '' we have for the left electron @xmath20 these two outcomes are orthogonal to each other @xmath21 thus we can distinguish cases when 0 or 1 electrons travel in the neighbouring lead .
moreover the relation @xmath21 is universal in the sense that it is preserved when we change the lead were electrons are injected in the mzi or when we place the wire to the left of the mzi ( as long as the bottom beam - splitter remains half - reflecting ) .
we may put into correspondence to the wave function @xmath22 a qubit state @xmath23 and to @xmath24 a state @xmath25 .
this makes our three mzi setup very similar to the situation displayed in fig .
[ scheme_setup(fig ) ] .
there we have a spin - based electron counter with spin flipping after one electron passes through a wire ( like it is described in ref . ) .
this analogy between mzi and spin - based counter described here is helpful for understanding the nature of the ghz - state in the proposed setup ( fig .
[ setup(f ) ] ) . up to the level `` lv2 '' on fig .
[ setup(f ) ] , the evolution can be represented in terms of spin - counters ( fig .
[ scheme_setup(fig ) ] ) act like spin - counters ( large arrows on the figure ) coupled to left and right wire of middle mzi.,width=151 ] after passing through the beam - splitter the electron state in the middle mzi is @xmath26 .
depending on the arm from which electron came the corresponding spin flips .
this produces the final state @xmath27 which is obviously a ghz - type state .
we give here the detailed explanation about the ghz - state production .
we find an exact expression for the electron state at the level `` lv2 '' on fig .
[ setup(f ) ] , and we identify the local unitary transform making it a standard ghz state .
`` local '' means that the unitary transform is a direct product of three unitary operators each acting over a hilbert space of a corresponding electron @xmath28 .
specifically the initial three particle state is chosen to be ( see bottom of fig .
[ setup(f ) ] ) : @xmath29 note that we choose a simple product state , instead of choosing a slater determinant of the three single particle wave functions .
this is due to the fact that the electron wave function do not have mutual parts of trajectories , so that exchange effects do not play any considerable role . at the level `` lv1 '' after passing the first row of beam - splitters : @xmath30 in order to find the state after the coulomb interaction has acted , we rewrite the previous equation as @xmath31 the state after interaction ( at the lv2 ) : @xmath32 or in a more symmetrical way @xmath33 now it is easy to see the unitary transformation which produces the canonical ghz - state @xmath34 from ( [ initial state ] ) .
it should transform @xmath35 into @xmath36 , @xmath37 into @xmath38 , @xmath39 into @xmath40 , @xmath41 into @xmath41 , @xmath42 into @xmath43 , and @xmath44 into @xmath45 .
this transform is @xmath46 where @xmath47 so , at the level `` lv2 '' of our setup ( fig .
[ setup(f ) ] ) we have the ghz - state up to a change of basis achieved by the transform eq .
[ transform ] .
this transform could be performed by the second bs row between level `` lv2 '' and level `` ( lv2 ) '' .
as one may notice , transformation @xmath48 does not correspond to any @xmath7 parameter in the formula for transfer matrix of beam - splitter ( [ eq : tmatrix ] ) .
we may avoid this problem by choosing @xmath49 and shifting the ab phase in the next mz loop by @xmath50 .
in reality we do not need this action because we intend to use `` condensed '' scheme where preparation and measurement are performed with one row of the beam - splitters ( see section [ section : condensed ] ) .
it is interesting to note that from the quantum computation point of view we have an algorithm implemented on 3 flying qubits which is essentually similar to the coding scheme depicted on fig.10.2 . in nielsen - chuang@xcite . in this work
we stay focused on the creation of a ghz - state in its original form @xmath51 .
however in some cases it might be useful to obtain more general ghz - type state such as @xmath52 with arbitrary @xmath53 and @xmath54 .
such a generalized state could be useful e.g. for quantum error correction.@xcite in our setup it is possible to get such a ghz - type state by a slight modification of the above presented method .
we simply need to replace the half - reflecting bs in the second mzi with another one .
as we pointed earlier in the general case the transfer - matrix of the bs can be represented as follows @xmath55 with this beam splitter in the second mzi the electronic wavefunction is : @xmath56 which is a generalized ghz state in a rotated " basis . using the same unitary transformation @xmath28 , where @xmath57 are defined in eq .
( [ transformation ] ) , produces the generalized ghz - state in its canonical form .
to detect the ghz - state experimentally we suggest to use a bell - type inequality violation test for three particles@xcite .
let us discuss some basic facts about bell - type inequalities for 3 particles which constitute one of the possible generalizations of bell inequality on a tripartite case .
we start from an algebraic inequality @xmath58 which is satisfied when @xmath59 are real variables with absolute values less or equal to @xmath60 ( this is a three particle generalization of the algebraic inequality used for bell inequality tests on two particles ) .
consider a three particle entangled state , written in pseudospin notation ( @xmath61 , @xmath62 , each particle can be detected in one of two leads l and r ) .
the projection of the pseudospin of @xmath4 on some vector @xmath63 corresponds to @xmath64 while the projection on @xmath65 corresponds to @xmath66 . from the point of view of local hidden variable theories , after the creation of the three particle state , between the two measurements the first electron has projections @xmath67 , @xmath68 , the second one @xmath69 , @xmath70 , and the third one @xmath71 , @xmath72 ( @xmath73 if the spin is parallel to @xmath74 , and @xmath75 if it has opposite direction ) .
different outcomes of measurements are due to a hidden variable @xmath76 which varies from measurement to measurement .
we have identified the real numbers @xmath77 with @xmath78 . the average value of b is then given by : @xmath79 where @xmath80 is the distribution function of the hidden variable .
experimentally this value can be measured by the following procedure .
let us define the correlator @xmath81 then @xmath82 and from eq .
( [ algebraic ] ) the local hidden variable average is such that : @xmath83 from a quantum mechanical point of view , the correlator of eq .
( [ correlator ] ) is an average over the state of an operator describing the measurement of three spins in specific directions : @xmath84 and @xmath85 where the analog of the `` bell operator '' for three particles reads @xmath86 in appendix [ ghz_appendix ] we recall under which conditions this bell operator gives a maximal violation of the bell type inequality .
the demonstration is based on the fact that first , in order for @xmath87 to have an eigenvalue equal to @xmath88 , the state must be a linear superposition of @xmath89 and @xmath90 .
second , the bell operator @xmath91 has eigenvalues @xmath92 only if the state is related to the ghz state : @xmath93 by local unitary transformations . finally , in this appendix , we specify which angles of spin measurement give the maximal value for @xmath94 .
the resulting angles are : @xmath95 where @xmath96 ( different signs identify two classes of angles corresponding to the upper and lower signs in the formulae ) .
the origin of these classes lies in the symmetry of the ghz - state with respect to reflection in the x - y plane .
these angles will further be transformed into the corresponding beam - splitter parameters @xmath7 and @xmath15 .
we are now in a position to describe all the steps for production and detection , which are achieved in a rather complicated three mzi setup with each mzi containing double loops ( fig .
[ setup(f ) ] ) . the incoming state of eq .
( [ incoming ] ) is injected into the first bs and becomes that of eq .
( [ after_first ] ) at level `` lv1 '' .
the @xmath14 shift is applied to neighbouring channels and it results in the rotated ghz state of eq .
( [ initial state ] ) at level `` lv2 '' .
the true ghz state of eq .
( [ nominal ] ) is achieved at level `` ( lv2 ) '' by passing through the second beam splitter row ( with phase shift ) .
the next mzi loop is associated with the bell measurements process . to produce the bell measurement correlator we need three values @xmath97 corresponding to separate measurements with results within the band @xmath98 $ ] . in the spin case
, the spin projection measurements serve this purpose . in our setup
we can detect a particle in the left or in the right arm .
so one of the possible assignments is @xmath99 for the particle detected in left arm and @xmath100 for particle detected in right arm .
then the bell correlator is an average of a product @xmath101 which by definition is @xmath102 where @xmath103 is the probability for @xmath101 to be equal to @xmath60 and @xmath104 to -1 correspondingly .
@xmath105 in four cases .
first one is @xmath106 , @xmath107 , @xmath108 , which corresponds to the case where the three electrons were observed in right arms .
we define the probability of this event as @xmath109 .
other three possible cases have corresponding probabilities @xmath110 , @xmath111 , @xmath110 .
so @xmath112 analogously @xmath113 finally , @xmath114 this correlator measured on level `` lv3 '' of the setup of fig .[setup(f ) ] corresponds to the @xmath115 correlator of the pseudospin .
_ measurement of pseudospin in arbitrary direction .
_ as we have seen in sec .
[ ghz_section ] , in order to make bell measurements we should be able to measure pseudospin projections in arbitrary directions , say , @xmath116 .
nevertheless it can not be achieved directly because we only can measure the presence of the particles in the right or the left lead ( spin up - spin down " ) .
instead one can do an equivalent measurement which gives the same expectation values .
one should transform the state in such a way that the direction @xmath116 of the pseudospin converts to @xmath117 and then perform the standard measure procedure eq .
( [ measproc ] ) .
this is done by the upper beam - splitters , and we obtain : @xmath118 for the state on level `` ( lv2 ) '' . the relation between rotation parameters and beam - splitter characteristics
is described in appendix [ bs_appendix ] . for each bell
correlator the parameters of this loop are different .
as soon as we have defined the correlators @xmath119 we are ready to calculate the value of the bell observable @xmath120 which is the result of the experiment .
granted , the fact that we have to include an additional loop in each mzi is quite cumbersome and renders the implementation of our setup with mesoscopic devices more difficult . in the next section
we will show that it is possible to perform the same task without the additional loop . here , we insist that the use of an additional loop in each mz can be circumvented , provided that we exploit the composition of two bs in series
. we will call the setup where the additional loop has been removed : `` condensed detection scheme '' .
consider one of three interferometers ( fig .
[ setup(f ) ] ) . after level `` lv2 '' it has two beam - splitters each performing its own unitary transformation .
let the transformation matrices of the bs located between level `` lv2 '' and level `` ( lv2 ) '' and the bs between `` ( lv2 ) '' and `` lv3 '' be @xmath121 and @xmath122 correspondingly .
then total transformation @xmath123 is also unitary .
the idea is to perform this total transformation with a single bs , and thus to keep the setup as simple as possible ( i.e. remove the dotted line part of fig .
[ setup(f ) ] ) . to be precise ,
note that a single beam - splitter can not reproduce fully all unitary transformations , and there is the possibility that the transform @xmath123 can not be reproduced in full generality with a single beam - splitter .
but it can be shown that , from the * point of view of measurement * , it is always possible with a single bs to perform a transformation which gives the same measure as the total transformation @xmath123 .
we leave the details of the proof of this statement to the appendices ( in particular appendix [ settings_appendix ] ) .
it is interesting to note that in the condensed detection setup , there exists a measurement scheme where the desired violation @xmath124 is achieved using only changes of the phase difference @xmath125 , @xmath126 of the mzi loops , while the transparencies of the bs remain constant : @xmath127 here we stress that the @xmath15 phases presented in the last formula are effective " , meaning that they depend on the practical implementation of the beam - splitters ( values in the formula are for @xmath128 given by eq .
( [ eq : tmatrix ] ) ) . in appendix
[ bs_appendix ] we show how to adjust the phases in the case of real experiments .
this constitutes the justification for removing the additional loop in each mz .
the whole set of measurement schemes with @xmath124 is presented in appendix [ settings_appendix ] .
while the setups which we have presented and discussed in this work were schematic , we think that a practical implementation is in reach with current experimental techniques . using edge states in the integer quantum hall effect regime ( iqhe )
, one can obtain chiral channels where backscattering is impossible .
beam splitters can be obtained using quantum point contact ( qpc ) .
several experiments realizing one or two electronic mzi have already been achieved @xcite . the actual geometry for a working device , reproducing the setup of fig .
[ setup(f ) ] is shown in fig .
[ practical ] .
( @xmath5 ) are electron sources and @xmath129 are drains .
guiding centers of electron edge states meet at quantum point contacts , whose transparency can be controlled with gate voltages .
the rounded rectangles indicate the interaction regions between neighbouring mzis .
space available for the fermi sea is within the hatched regions , width=283 ] this picture represents a 3d potential landscape .
lines going near the potential walls represent chiral edge eigenstates in iqhe .
dotted lines represent unoccupied edge states .
the fermi sea is expulsed from the setup region to prevent screening effects and parasitic entanglement .
the edge states meet at qpc s , whose transparency can be controlled by gate voltages .
each mzi has two electronic sources ( noted @xmath130 and @xmath131 , @xmath132 , on fig .
[ practical ] ) , and two drains collecting the electronic current(@xmath133 and @xmath134 ) .
note that by changing gate voltages , it is possible to modulate the length of the different paths , and thus to control the aharonov - bohm phase acquired by the electrons during propagation .
the two circled rectangles on the pictures show the interaction regions , where two branches of two different mzis are put close to each other , in such a way that a @xmath14 phase shift is produced when electrons are present in the two branches . in order to have time - correlated propagation of electrons over interaction regions , we need on - demand electron sources similar to the ones which have already been experimentally realized in the iqhe @xcite . in experiments on single - electron injection time of injection varied from @xmath135 to @xmath136 ( in this paragraph we use numerical values from ref . ) .
the wave packets should be shorter than the length of interaction region .
the mall size of wave - packets ensures that the interaction produces a simple phase shift of the whole wave packet ( see appendix [ coulomb interaction . ] ) .
the suitable interaction time @xmath137 then should be at least 10 times larger ( more detailed calculations in the case of two mzi are presented in ref . ) and starts with @xmath138 .
the phase accumulated due to the interaction must be equal to @xmath14 . for this to happen
the interaction strength must exceed @xmath139 . the unscreened coulomb interaction @xmath140 ( @xmath141 is dielectric constant of gaas ) provides such a strength at the distances starting at @xmath142 which is about @xmath143 of magnetic length @xmath144 for magnetic field @xmath145 .
if the fermi sea regions and various gates are at the larger distance than the distance between the edge states in the interaction region , using the unscreened form of coulomb interaction looks reasonable . therefore needed phase accumulation is accessible in realistic structures . if one uses short electron wave - packets another demand appears . as short packets have wide energy distribution we must ensure that beam - splitter scattering is energy - independent within that range .
the energy width of a wave - packet with duration @xmath135 is approximately @xmath146 .
this requires a short enough scattering structure ( e.g. qpc ) in the beam - splitter .
next issue is the size of setup .
if the interaction time equals to @xmath138 then for a drift velocity @xmath147 ( taken from ref . ) interaction region length would be @xmath148 .
this length exceeds the size of electronic mach - zehnder interferometers where ab - oscillations were observed . a satisfactory length for the interaction region should be of order @xmath149 .
to achieve this length one may either produce shorter wave - packets ( keeping in mind that the beam - splitters should remain energy - independent ) or reduce the drift velocity via making the potential walls smoother .
another issue is the accuracy of the interaction phases .
first , one can tune the interaction strength with extra gates .
if there is no such possibility , the setup should be fabricated with an interaction producing phase shift exceeding @xmath14 .
electrons then should be sent with a delay between them to reduce the interaction time and strength . the procedure for phase control
may be developed in the same style as the ab - phase adjustment described in appendix [ bs_appendix ] .
one should tune the bottom beam - splitter in the second mzi to transfer incoming electrons to the adjacent edge state and set the ab - phase in the first mzi to @xmath150 .
the interaction strength then should be tuned to the point where the electron sent to the first mzi comes out from the different lead .
after this for better tuning one applies an ab - phase @xmath151 . here
if the coulomb interaction phase is equal to @xmath14 the sensitivity of one particle probabilities in the first mzi reaches its maximum .
then the same procedure should be repeated for the second interaction zone .
although it looks possible to meet the requirements listed above , the whole realization of the experiment remains very challenging .
it might be therefore very interesting to realize less ambitious project with constant voltage and detection of low frequency third order current - current correlator .
the latter rids of single particle detection while the former substitutes single electron sources .
one can also express orbital correlators ( [ measproc ] ) in terms of current - current correlators ( see refs . ) .
third order current correlation can be measured along the lines realized in ref . or theoretically suggested in ref . .
still , in this case there is small hope for observation of bell inequality violation as it requires time correlated propagation of electrons for accumulation of definite phase ( @xmath14 ) with coulomb interaction , which is a very rare event in the constant voltage regime - the most probable event is no phase at all since electrons usually pass interaction region separately .
therefore , in the constant voltage regime bell inequlity violation paradigm becomes impractical and one has to come back to the usual physics and compare the experiment with some specific theory .
this approach does not have the main advantage of bell inequality violation which indicates entanglement without addressing to the details of the setup construction .
instead , one has to measure irreducible third order current cross - correlators as a function of interaction strength and other parameters ( e.g. ab phases and bs transparencies )
. from that function it should be possible to find a term originating from coulomb interaction which would indirectly indicate entanglement .
this kind of observation would also be a significant step towards the realization and detection of ghz - state and electronic orbital states .
in this article we have demonstrated , how to produce a generalized ghz states for 3 flying qubits , formed as orbital electron states in the mzi - type geometries .
our setup relies on two state of the art devices of nanophysics : a ) single electron sources which have been demonstrated@xcite and characterized@xcite experimentally , as well as theoretically@xcite ; b ) mach zehnder interferometers , which have equally been the object of thorough experimental@xcite and theoretical@xcite investigations . in our proposal
, both devices should be integrated together , and moreover three mzi placed in parallel are needed to achieve the ghz state . while it is plausible to think that it will be challenging to build a prototype in the near future , we judge that it is useful for the mesoscopic physics community to be aware that advanced quantum information protocols here the production of a ghz state can be achieved with electrons . at the beginning of the paper we justified that our ideal devices generate the ghz state .
subsequently , we provided a detailed explanation of the type of bell inequalities which need to be used to prove unambiguously that we have generated the proper state .
the problem which we encountered with the actual mzi setup which is needed to implement the bell test is that it requires 3 more mzi loops than `` condensed scheme '' , thus making the integration of the device even more challenging .
fortunately , we provided a `` condensed detector scheme '' were this complication can be circumvented , arguing that the operations achieved by two bs in series can be combined with a single bs .
this allowed to actually draw a `` realistic '' device inspired from the so called `` air bridge '' technique which is used in the experiments of ref . .
this work thus belongs to the ongoing effort called `` electron quantum optics '' where experiments and paradigms of quantum optics are reproduced with mesoscopic physics devices .
the great advantage of the present way for producing ghz states relies in the fact that no postselection procedure is needed here .
the steps described in this paper rely on a unitary evolution of the initial state , and moreover , we exploit the coulomb interaction between electron wave packets in order to generate the desired phase shift on the electron wave function .
this particular feature departs strongly from the photon protocols , as photons interact weakly when travelling in vacuum .
we thus state that the use of electron - electron interactions in single electron devices schemes may open up new possibilities for quantum information schemes which are not envisionable for photons .
granted the present work has addressed the case of an ideal device which is free from dephasing effects .
such phenomena are likely to affect the operation of our device as they are already known to be present in single mzi electronic setups.@xcite nevertheless our top priority is to clarify how to implement an abstract quantum information schemes in an actual mesoscopic device
. our analysis could be refined by taking dephasing effects into account .
the authors plan to consider this in the future . in case of similar scheme with two mzi
it is done in ref .
also we pointed out some factors that need to be accounted for in experiment and mentioned simplified scheme with constant voltage instead of on - demand electron sources .
we acknowledge financial support by the cnrs lia agreements with landau institute and the rfbr grant # 11 - 02 - 00744-a ( aav and gbl ) .
t.j . and t.m .
acknowledge support from grant anr 2010 blanc 0412 02 .
here we will prove that in the tripartite case @xmath124 can be achieved only for the ghz - state , or the states that can be reduced to ghz with some local unitary transformations .
next we will find all possible series of measurement angles which give @xmath124 for the ghz - state .
we use the spin - based approach in this section for convenience .
first we compute the square the bell operator ( [ bell operator def ] ) and find conditions necessary for @xmath152 .
@xmath153\otimes[\hat{\sigma}_\mathbf{a_2 } , \hat{\sigma}_\mathbf{a_2'}]\otimes\mathbf{1 } \nonumber\\ & ~&~~- \mathbf{1}\otimes[\hat{\sigma}_\mathbf{a_2},\hat{\sigma}_\mathbf{a_2'}]\otimes[\hat{\sigma}_\mathbf{a_3},\hat{\sigma}_\mathbf{a_3 ' } ] \nonumber\\ & ~&~~- [ \hat{\sigma}_\mathbf{a_1},\hat{\sigma}_\mathbf{a_1'}]\otimes\mathbf{1}\otimes[\hat{\sigma}_\mathbf{a_3},\hat{\sigma}_\mathbf{a_3'}].\end{aligned}\ ] ] pauli matrices ( @xmath154 ) obey angular momentum commutation relations : @xmath155=2i\epsilon^{ijk}\sigma_k,\ ] ] from which we obtain the commutator for a spin pointing along arbitrary vectors @xmath156 and @xmath157 @xmath158=2i|e|\sigma_\mathbf{e},\ ] ] where @xmath159 . for our task we will need vectors @xmath160 , @xmath161 , @xmath162 .
they allow to represent @xmath87 in the following form : @xmath163 this operator has a maximum eigenvalue @xmath164 . for the maximum value to be achieved
we require that @xmath165 , which implies @xmath166 , @xmath167 , @xmath168 .
the eigenstates @xmath169 of this value obey the relations @xmath170 it is convenient to choose a basis in each particle s space in relation with the corresponding @xmath171-vector .
for example , @xmath172 for the first spin means that @xmath173 . in this basis
the only states which give @xmath174 are @xmath175 and @xmath176 , so in full generality we choose a superposition : @xmath177 we can now find which state of the form of eq .
( [ superposition ] ) is an eigenstate of the operator @xmath91 with eigenvalue @xmath124 . choosing the basis we associate the @xmath178 axis of the first spin space with vector @xmath179 , and similarly for the other two spin states .
this implies that @xmath74 and @xmath180 are orthogonal vectors in the @xmath181 plane , and similarly for @xmath182 , @xmath183 and for @xmath184 , @xmath185 . for specificity
we choose @xmath186 and @xmath187 .
analogously for the other spin components . note that @xmath188 thus implying that @xmath189 from this we easily confirm that when @xmath190 we have the desired eigenstate with @xmath124 @xmath191 also , note that the orthogonal state @xmath192 has an eigenvalue @xmath193 .
both states are equivalent to the ghz state up to local unitary transformations . in conclusion
we have shown that each state which can violate the bell - type inequality maximally is a ghz - state up to trivial transformations .
now we proceed with the reverse problem . starting from the ghz - state ,
let us find all sets of vectors @xmath74 , @xmath182 , @xmath184 , @xmath180 , @xmath183 , @xmath185 , which yield @xmath194 .
we consider a spin correlator @xmath118 for arbitrary unitary vectors @xmath195 ,
@xmath196 , @xmath197
with corresponding spherical coordinates @xmath198 . the operator for the spin projection on direction @xmath116 is then @xmath199 this allows us to find @xmath200 @xmath201 and the average of the spin correlator in the ghz state is then : @xmath202 the simple formula which we have obtained is a justification of the spin - based approach to bell type inequalities . the tripartite case is simpler than the bipartite state when the question comes to finding the specific angles giving tsirelson bound , since here all the correlators should have unit absolute value . this implies @xmath203 and thus we should consider only vectors of the @xmath181 plane as candidates for @xmath124 .
the polar angles of these vectors @xmath74 , @xmath182 , @xmath184,@xmath180 , @xmath183 , @xmath185 are defined by the conditions : @xmath204 which in terms of polar angles reads : @xmath205 this system of equations has the solution : @xmath206 with @xmath207 , where @xmath208 .
all the vectors lye in the @xmath181 plane .
this concludes the definition of the angles giving @xmath124 for the ghz - state .
one of essential parts of the device is the bs .
the bs is a 4-arm scatterer with a special type of scattering matrix which allows to split the beam into two parts without any reflection .
generally the scattering matrix of 4-arm splitter looks like @xmath209 in this matrix the @xmath210 parameters refer to the transmission amplitudes from arm @xmath4 to arm @xmath211 , and @xmath212 is the reflection amplitude in arm @xmath211 .
this matrix obeys only a unitarity condition .
if time - reversal invariance is added a new condition appears @xmath213 .
the beam - splitter scattering matrix is @xmath214 for these parameters the following system of equations is valid @xmath215 where @xmath216 for such an equation set it is rather easy to find a simple parametrization .
let us assume that @xmath217 then we substitute all the t - values into last equation and get : @xmath218 for this equation to be valid must be @xmath219 in the present work we focus on symmetrical beam - splitters for simplicity .
for a symmetrical beam - splitter @xmath220 so the phases are @xmath221 and eq .
( [ phase cond ] ) simplifies to : @xmath222 one of the possible solutions is @xmath223 this parametrization will be used in further calculations .
we are interested in the properties of electron transport from bottom to top , so we will need only the values of @xmath224 .
now we can introduce the transfer matrix : @xmath225 note that we have changed indexes because of the reflectionless nature of the bs : @xmath226 . in order to find the state of particles after the bs one must perform the following substitution with wave - functions before the beam - splitter : @xmath227
since we have chosen a specific parametrisation of the bs , we must confirm that if the bs used in the experiment has a transfer matrix which is different from the one in eq .
( [ transfer - matr ] ) it will not spoil the whole experiment .
we consider again the phase relation in the general case : @xmath228 each phase can be presented as : @xmath229 where the phases @xmath230 represent the deviation from the chosen parametrisation .
these obey the relation : @xmath231 all solutions of this equation are taken into account by the following representation : @xmath232 where @xmath233 are some real values .
this answer has an explicit physical meaning .
all beam - splitters have the same transfer matrix up to the phase accumulation in each channel .
this phase accumulation can affect the phase difference in the left and the right paths of the mzi . in order to restore the original interference pattern we need to adjust the applied phase difference . during the experiment it means choosing a reference point with @xmath234 for given mzi .
this could be achieved using the fact that at @xmath235 an electron injected in the right arm of the mzi has zero probability to be detected in the right arm ( fig .
[ 1mzi ] ) .
yet better tuning can be achieved near @xmath151 . here
the probabilities for each arm are equal and the difference between the probabilities has a maximum derivative with respect to the phase .
electron goes to left arm.,width=94 ] _ beam - splitter with phase_. in our mzi the electrons have two possible paths to travel .
if there is a difference in the accumulated phase between two paths , it will affect the interference between them in a subsequent beam - splitter .
it is convenient to account for this phase difference ( which can be generated e.g. by applying a magnetic flux through the mzi loop ) in the transfer matrix of the bs .
if the phase lengths of the right and left the paths are @xmath236 and @xmath237 , then the transfer matrix of the beam - splitter takes the form : @xmath238 actually the final results depend only on the phase difference of the two paths @xmath239 so the transfer matrix can be multiplied by any phase factor without affecting the final results
. we can represent the transfer matrix in a form similar to the spin rotation operator @xmath240 since the measurement procedure in the condensed scheme includes changing @xmath241 and measuring @xmath242 at level `` ( lv2 ) '' ( fig .
[ setup(f ) ] ) it is convenient to define a measurement operator with respect to level `` lv2 '' of the scheme , where the electron state is invariant during the measurement procedure . as we see
the operator is a sum of two rotations , first a rotation of angle @xmath15 around the @xmath117 axis , second another rotation of angle @xmath243 around the @xmath244 axis .
the rotation around x - axis corresponds to the action of the beam - splitter itself , while the rotation around @xmath117 corresponds to an aharonov - bohm phase accumulation .
this whole rotation transforms a vector @xmath245 unto the @xmath117 axis . as the vector thus defined is in one - to - one correspondence with the pair of bs parameters
@xmath241 we call it a characteristic vector of the bs .
rotation means that @xmath246 the projection of the rotated pseudospin onto @xmath117 can be shown to be equal to the projection of the initial pseudospin onto @xmath116 .
@xmath247 for three - particle spin - correlators the same relation holds @xmath248 so now we have found that the measured correlator with respect to level `` lv2 '' is @xmath249 .
previously in appendix [ ghz_appendix ] we have defined angles yielding @xmath124 for three - spin ghz - state . to obtain the parameters of the actual setup ( fig .
[ setup(f ) ] ) we need to `` translate '' the spin angles into bs parameters .
this chapter contains the derivation of the `` translation '' rules .
we start with the logic of the ghz - experiment described in sec .
[ ghz_section ] .
we need three values @xmath97 corresponding to separate measurements with results within the interval @xmath98 $ ] . in the spin case ,
the spin projection measurements serve this purpose . in our setup
we can detect the particle in the left or the right arm .
so one of possible assignments is @xmath99 for the particle detected in left arm and @xmath100 for particle detected in right arm
. then the bell correlator is @xmath250 this correlator measured on level `` lv3 '' of the setup ( fig .[setup(f ) ] ) corresponds to the @xmath115 correlator of pseudospin . due to the rotation between `` (
lv2 ) '' and level `` lv3 '' the correlator of eq .
( [ corel ] ) corresponds to @xmath118 measured on the initial pseudospin state at `` ( lv2 ) '' which is ghz .
the vectors @xmath251 are characteristic vectors of the corresponding beam - splitters ( [ character vector ] ) .
now we calculate bs matrices corresponding to the angles giving maximum violation ( [ angles ] ) .
all the vectors are in the @xmath181 plane , thus @xmath252 for all of them . for the vector with the polar angle @xmath253 corresponding @xmath254 and
the bs transfer matrix is @xmath255 where equivalence means that the overall phase does not change the observable outcomes .
@xmath256 parameters then should obey the restrictions : @xmath257 for @xmath132 .
in appendix [ pseudospin analogy ] we have defined the properties of bs between levels `` ( lv2 ) '' and `` lv3 '' of fig . [ setup(f ) ] .
the next step lies in defining the parameters of bs in the condensed scheme . for this purpose
we use the properties of unitary transformation made between levels `` lv2 '' and `` ( lv2 ) '' of the initial setup ( [ transformation ] ) .
the problem is simplified by the fact that all bs between levels `` ( lv2 ) '' and `` lv3 '' are of the same type @xmath258 _ first and third mzi . _ since @xmath259 all the results are valid for the third mzi as well .
the first mzi chain makes two consequent unitary transformations , the first one is @xmath260 next is @xmath261 .
the product of these is unitary @xmath262 shifting the whole matrix phase by @xmath263 we get following expression @xmath264 the next step is to find from the matrix the beam - splitter the parameters @xmath7 , @xmath15 .
we encounter the problem that the matrix of eq .
( [ matrix1 ] ) can not represent the transfer - matrix of any bs .
the relation @xmath265 which is universal for any representation of the transfer matrix ( see eq .
( [ transfer matr ] ) ) is not valid for eq .
( [ matrix1 ] ) .
the bs transfer matrix should not necessarily be equal to @xmath266 but it should lead to the same three - particle probabilities .
this happens if transfer matrix of beam - splitter equals to : @xmath267 where @xmath268 and @xmath269 are some phase shifts .
they do not contribute to tripartite probabilities while this matrix with properly chosen phase shifts can be transfer matrix of bs . if @xmath270 then @xmath271 , @xmath272 and @xmath273 if @xmath274 then @xmath271 , @xmath275 and @xmath276 both formulas can be unified as : @xmath277 if @xmath278 $ ] last formula is simplified @xmath279 _ second mzi . _ for second mzi
everything is simpler since @xmath48 is a diagonal matrix .
@xmath280 this matrix corresponds to a bs with parameters @xmath9 and @xmath281 .
finally we can introduce the algorithm for the generation of the measurement settings resulting in @xmath124 . 1 .
choose angles @xmath282 obeying relations @xmath283 where @xmath132 .
2 . with necessary @xmath284 shifts
make @xmath285 be in @xmath286 $ ] ; 3 .
calculate bs parameters @xmath287 from the last equation we can extract following common features of these measurement sets .
* in the first and the third mzi the ab phase can take values @xmath288 ; * in the second mzi the transparency of the bs is not changing during the experiment ; among measurement schemes there exist ones where transparencies of all bs do not change during experiment .
for example , choosing @xmath289 we get @xmath290 in conclusion , we found the general parametrization for measurement settings giving @xmath124 in a regular mz setup , and described a special case where we get a maximum violation only by adjusting ab phases .
here we solve the two - particle problem for the propagation of two electrons in neighbouring channels between the two beam splitters .
we consider two chiral edge - channels of the iqhe .
the two particles propagating in different channels interact electrostatically within a specified interaction region .
the potential of interaction @xmath291 depends on the position of both particles .
the energy spectrum is assumed to be linear and the drift velocity of the electrons in each channel is the same @xmath292 .
we consider the hilbert space of two electron wave functions such that @xmath293 if @xmath294 or @xmath295 is negative .
the shrdinger equation for this two particle problem reads : @xmath296 let s make a substitution @xmath297 . then @xmath298 . here
@xmath299 and this allows to rewrite the equation @xmath300
the solution is @xmath301 in terms of the initial variables @xmath302 the phase accumulated during the interaction is position dependent and describes the energy exchange and the deformation of the wave - packets . here the condition for simple phase accumulation is that the size of the wave - packets is small . in this case
the position dependence of the phase should not manifest itself .
the solution agrees well with the derivation of ref .
except for the fact the interaction kernel differs ( we used a kernel of general form @xmath291 )
. 99 a. einstein , b.podolsky and n. rosen , phys . rev . * 47 * , 777 ( 1935 ) .
e. schrdinger , die naturwissenschaften * 23 * , 807(1935 ) .
j. s. bell , phys . * 1 * , 195 ( 1964 ) .
j. f. clauser , m. a. horne , a. shimony and r. a. holt , phys .
* 23 * , 880 ( 1969 ) a. aspect , p. grangier and g. roger , phys .
lett . * 47 * , 460 ( 1981 ) .
g. lesovik , t. martin and g. blatter , european physical j. b * 24 * , 287 ( 2001 ) .
patrik recher , eugene v. sukhorukov and daniel loss phys .
b * 63 * , 165314 ( 2001 ) . c. w. j. beenakker , c. emary , m. kindermann , and j. l. van velsen , phys . rev .
lett . * 91 * , 147901 ( 2003 ) .
p. samuelsson , e. v. sukhorukov , and m. bttiker , phys .
lett . * 91 * , 157002 ( 2003 ) ; _ ibid . _ * 92 * , 026805 ( 2004 ) .
n. m. chtchelkatchev , g. blatter , g. b. lesovik and t. martin , phys .
b * 66 * , 161320 ( 2002 ) .
a. v. lebedev , g. b. lesovik and g. blatter , phys .
b * 71 * , 045306 ( 2005 ) .
o. sauret , t. martin and d. feinberg , phys .
b * 72 * , 024544 ( 2005 ) .
k. v. bayandin , g. b. lesovik , and t. martin , phys .
b * 74 * , 085326 ( 2006 ) .
d. m. greenberger , m. a. horne , a. shimony and a. zeilinger , american journal of physics * 58 * , 1131 ( 1990 ) .
b. s. cirelson , lett .
* 4 * , 93 ( 1980 ) . m. a. nielsen and i. l. chuang , _ quantum computation and quantum information _
( cambridge university press , 2000 ) .
a. karlsson and m. bourennane , phys . rev .
a * 58 * , 4394 ( 1998 ) .
j. c. hao , c. f. li and g. c. guo , phys .
a * 63 * , 054301 ( 2001 ) .
a. zeilinger , m.a .
horne , h. weinfurter , and m. zukowski , phys .
* 78 * , 3031 ( 1997 ) .
a. v. lebedev , g. blatter , c. w. j. beenakker and g. b. lesovik , phys .
b * 69 * , 235312 ( 2004 ) kicheon kang and kahng ho lee , physica e * 40 * , 1395 ( 2008 ) .
a. a. vyshnevyy , a. v. lebedev , g. b. lesovik and g. blatter ( unpublished ) .
j. dressel , y. choi , and a. n. jordan , phys . rev .
b * 85 * , 045320 ( 2012 ) .
dik bouwmeester , jian - wei pan , matthew daniell , harald weinfurter and anton zeilinger , phys .
82 * , 1345 ( 1999 ) .
chao - yang lu , xiao - qi zhou , otfried ghne , wei - bo gao , jin zhang , zhen - sheng yuan , alexander goebel , tao yang and jian - wei pan , nature physics * 3 * , 91 ( 2007 ) .
lieven m. k. vandersypen , matthias steffen , gregory breyta , costantino s. yannoni , mark h. sherwood and isaac l. chuang , nature * 414 * , 883 ( 2001 ) .
et al . _ ,
science * 316 * , 1169 ( 2007 ) .
yang ji , yunchul chung , d. sprinzak , m. heiblum , d. mahalu and hadas shtrikman nature * 422 * , 415 ( 2003 ) ; i. neder , n. ofek , y. chung , m. heiblum , d. mahalu and v. umansky , nature * 448 * , 333 ( 2007 ) ; p. roulleau , f. portier , d. c. glattli , p. roche , a. cavanna , g. faini , u. gennser and d. mailly , phys .
b * 76 * , 161309 ( 2007 ) .
a. mah , f. d. parmentier , e. bocquillon , j .-
m . berroir , d. c. glattli , t. kontos , b. placais , g. fve , a. cavanna and y. jin , phys . rev .
b * 82 * , 201309(r ) ( 2010 ) .
g. b. lesovik , m. v. suslov and g. blatter , phys .
a * 82 * , 012316 ( 2010 ) a. v. belinskii and d. n. klyshko , phys .
usp . * 36 * , 653 - 693 ( 1993 ) m. albert , c. flindt and m. bttiker , phys .
b * 82 * , 041407(r ) ( 2010 ) ; m. albert , c. flindt and m. bttiker , phys . rev . lett . *
107 * , 086805 ( 2011 ) ; g. haack , m. moskalets , j. splettstoesser and m. bttiker , phys .
b * 84 * , 081303(r ) ( 2011 ) ; t. jonckheere , t. stoll , j. rech and t. martin , phys .
b * 85 * , 045321 ( 2012 ) .
h. forster , s. pilgram and m. bttiker , phys .
b * 72 * , 075301 ( 2005 ) ; f. marquardt , europhys . lett . *
72 * , 788 ( 2005 ) ; f. marquardt and c. bruder , phys .
* 92 * , 056805 ( 2004 ) .
youn , h .- w . lee and h .- s .
sim , phys .
* 100 * , 196807 ( 2008 ) d. l. kovrizhin and j. t. chalker , phys .
b * 81 * , 155318 ( 2010 ) .
p. roulleau , f. portier , p. roche , a. cavanna , g. faini , u. gennser and d. mailly , phys .
* 101 * , 186803 ( 2008 ) .
beenakker , c. emary and m. kindermann , phys .
b * 69 * , 115320 ( 2004 ) g. gershon , yu .
bomze , e. v. sukhorukov and m. reznikov , phys .
lett . * 101 * , 016803 ( 2008 ) k. v. bayandin , a.v . lebedev and g. b. lesovik ,
jetp * 106 * , 117 ( 2008 ) | we propose a new single - step scheme for the generation of a ghz entangled state of three single - electron excitations ( flying qubits ) .
we also present a method to get a generalized ghz - state .
our idea relies upon the most recent progress in the field of on - demand electron sources and mesoscopic mach - zehnder interferometry .
we also provide the recipe for the unambiguous detection of this ghz state via correlations measurements at the output , which imply the violation of a bell - type inequality which is generalized to the case of three particles .
we explain how such measurements can be achieved in the context of mach - zehnder interferometry , and draw an actual prototype device which could be achieved with point contacts and metallic gates placed on a gaas sample , in the integer quantum hall effect regime . |
the brain has a highly developed and complex self - generated dynamical neural activity , and this fact raises a series of interesting issues . does this self - sustained neural dynamics , its eigendynamics , have a central functional role , organizing overall cognitive computational activities ? or does this ongoing autonomous activity just serve as a kind of background with secondary computational task , like non - linear signal amplification or time encoding of neural codes ?
the answer to this question is important not only to system neurobiology , but also for research in the field of cognitive computation in general .
we will review here approaches based on the notion that the autonomous neural dynamics has a central regulating role for cognitive information processing .
we will then argue , that this line of research constitutes an emerging field in both computational neuroscience and cognitive system research .
some preliminaries , before we start .
this is a mostly non - technical review with emphasis on content , an exhaustive and complete discussion of the published work on the subject is not the objective here . centrally important equations will be given and explained , but for the numerical values of the parameters involved , and for the details of the simulation set - ups , we will refer to the literature .
the discussion will be given generally from the perspective of cognitive system theory , _ viz _ bearing in mind the overall requirements of prospective complete cognitive systems , akin to ones of real - world living animals @xcite .
on the experimental side , the study of self - induced or autonomous neural activity in the brain has seen several developments in recent years , especially by fmri studies @xcite , and we will start by discussing some key issues arising in this respect . the vast majority of experiments in cognitive neuroscience study the evoked neural response to certain artificial or natural sensory stimuli , often involving a given task which has been trained previously . it has been realized early on , that the neural response shows strong trial - to - trial variation , which is often as large as the response itself .
this variability in the response to identical stimuli is a consequence of the ongoing internal neural activities ( for a discussion see @xcite ) .
experimentally one has typically no control over the details of the internal neural state and it is custom to consider it as a source of noise , averaging it out by performing identical experiments many times over .
it is on the other side well known that the majority of energy consumption of the brain is spent on internal processes @xcite , indicating that the ongoing and self - sustained brain dynamics has an important functional role .
two possibilities are currently discussed : _ ( a ) _ the internal neural activity could be in essence a random process with secondary functional roles , such as non - linear signal amplification or reservoir computing for the spatiotemporal encoding of neural signals ( for a theory review see @xcite ) .
_ ( b ) _ the internal neural activity could represent the core of the cognitive information processing , being modulated by sensory stimuli , but not directly and forcefully driven by the input signals .
indications for this scenario arise , e.g. , from studies of the visual information processing @xcite and of the attention system @xcite .
the overall brain dynamics is still poorly understood and both possibilities ( a ) and ( b ) are likely to be relevant functionally in different areas . in this review
we will focus on the ramifications resulting from the second hypothesis .
there are indications , in this regard , that distinct classes of internal states generated autonomously correspond to dynamical switching cortical states , and that the time series of the spontaneous neural activity patterns is not random but determined by the degree of mutual relations @xcite .
additionally , these spontaneous cortical state may be semantic in nature , having a close relation to states evoked by sensory stimuli @xcite and to neural activity patterns induced via thalamic stimulation @xcite .
a second characteristics recurrently found in experimental studies is the organization of the spontaneously active states into spatially anticorrelated networks @xcite , being transiently stable in time , in terms of firing rates , with rapid switching between subsequent states @xcite .
these results indicate that certain aspects of the time evolution of the self - sustained neural activity in the brain have the form of transient state dynamics , which we will discuss in detail in sect .
[ sect_transient_state_dynamics ] , together with a high associative relation between subsequent states of mind .
this form of spontaneous cognitive process has been termed ` associative thought process ' @xcite .
it is currently under debate which aspects of the intrinsic brain dynamics is related to consciousness .
the global organization of neural activity in anticorrelated and transiently stable states has been suggested , on one side , to be of relevance also for the neural foundations of consciousness @xcite , _ viz _ the ` observing self ' @xcite .
the persistent default - mode network ( for a critical perspective see @xcite ) , _ viz _ the network of brain areas active in the absence of explicit stimuli processing and task performance , has been found , on the other side , to be active also under anesthetization @xcite and light sedation @xcite . it is interesting to note , in this context , that certain aspects of the default resting mode can be influenced by meditational practices @xcite . the term ` neural transients ' characterizes evoked periods of neural activities , remaining transiently stable after the disappearance of the primary stimulating signal . in the prolonged absence of stimuli , neural architectures based on neural transients
relax back to the quiescent default state .
network setups based on neural transients therefore occupy a role functionally in between pure stimulus - response architectures and systems exhibiting continuously ongoing autonomous neural activity . an important class of neural architectures based on neural transients are neural reservoirs , which we discuss now briefly . a recurrent neural net is termed a reservoir , if it is not involved in the primary cognitive information processing , having a supporting role . a typical architecture is illustrated in fig .
[ figure_reservoir_dynamics ] .
the reservoir is a randomly connected network of artificial neurons which generally has only a transiently stable activity in the absence of inputs , _
the reservoir has a short - term memory . in the standard mode of operation an input signal
stimulates the network , giving raise to complex spatiotemporal reservoir activities .
normally , there is no internal learning inside the reservoir , the intra - reservoir synaptic strengths are considered fixed .
time prediction is the standard application range for reservoir computing . for this purpose the reservoir
is connected to an output layer and the activities of the output neurons are compared to a teaching signal . with supervised learning , either online or off - line , the links leading from the reservoir to the output then acquire a suitable synaptic plasticity .
there are two basic formulations of reservoir computing .
the ` echo - state ' approach using discrete - time rate - encoding neurons @xcite , and the ` liquid state machine ' using continuous - time spiking neurons @xcite . in both cases
the dimensionality of the input signal , consisting normally of just a single line , is small relative to the size of the reservoir , which may contain up to a few hundred neurons .
many nonlinear signal transformations are therefore performed by the reservoir in parallel and the subsequent perceptron - like output neurons may solve complex tasks via efficient linear learning rules .
neural reservoirs are possible candidates for local cortical networks like microcolumns .
the bare - bones reservoir network is not self - active , but feedback links from the output to the reservoir may stabilize ongoing dynamical activity @xcite . in any case ,
reservoir nets are examples of network architectures of type ( a ) , as defined in the previous section .
the task of the reservoir , non - linear signal transformation , is performed automatically and has no semantic content .
all information is stored in the efferent synaptic links .
there is an interesting similarity , on a functional level , of reservoir computing with the notion of a ` global workspace ' @xcite .
the global workspace has been proposed as a global distributed computational cortical reservoir , interacting with a multitude of peripheral local networks involving tasks like sensory preprocessing or motor output .
the global workspace has also been postulated to have a central mediating role for conscious processes @xcite , representing the dominating hub nodes of a large - scale , small - world cortical network @xcite .
a central question in neuroscience regards the neural code , that is the way information is transmitted and encoded ( see @xcite for reviews ) . keeping in mind that there is probably no pure information transmission in the brain , as this would be a waste of resources , that information is also processed when transmitted
, one may then distinguish two issues regarding the encoding problem . on one hand there is the question on how sensory signals are reflected , on relative short timescales , in subsequent neural activities .
available neural degrees of freedom for this type of short - time encoding are the average firing rates ( rate encoding ) , transient bursts of spikes and the temporal sequence of spikes ( temporal encoding ) .
in addition , the response of either individual neurons may be important , or the response of local ensembles @xcite .
the subsequent sensory signal processing , on timescales typically exceeding 25 - 100ms , may , on the other hand , involve neural dynamics in terms of transiently stable activity patterns , as discussed earlier in sect .
[ subsect_autonomous_brain_dynamics ] . in fig .
[ figure_transient_states ] two types of model transient state activities are illustrated .
alternating subsets of neurons are either active , to various degrees , or essentially silent , resulting in well characterized transient states having a certain degree of discreteness .
this discreteness should be reflected , on a higher level , on the properties of the corresponding cognitive processes . of interest in this context
is therefore the ongoing discussion , whether visual perception is continuous or discrete in the time domain @xcite , on timescales of the order of about 100ms , with the discrete component of perception possibly related to object recognition @xcite .
transient state dynamics in the brain may therefore be related to semantic recognition , a connection also found in models for transient state dynamics based on competitive neural dynamics . in the following
we will examine the occurrence and the semantic content of autonomous transient state dynamics in several proposed cognitive architectures .
the concept of saddle point networks is based on the premises , ( a ) that the internal ongoing autonomous dynamics organizes the cognitive computation and ( b ) that the cognitive behavior is reproducible and deterministic in identical environments @xcite .
as we will discuss in the next section , the first assumption is shared with attractor relic networks , while the second is not .
technically , one considers a dynamical system , _ viz _ a set of @xmath0 first - order differential equations and the set of the respective saddle points , compare fig . [ figure_scenarios ] .
the precondition is now that every saddle point has only a single unstable direction and @xmath1 stable directions .
any trajectory approaching the saddle point will then leave it with high probability close to the unique unstable separatrix and the system therefore has a unique limiting cycle attractor .
this limiting cycle does not need to be a global attractor , but normally has a large basin of attraction . during one passage most ,
if not all , saddle points are visited one after the other , giving raise to a transient state dynamics illustrated in fig .
[ figure_transient_states ] , with the trajectory slowing down close to a saddle point .
another condition for this concept to function is the formation of a heteroclinic cycle , which is of a set in phase space invariant under time evolution @xcite .
implying , as illustrated in fig .
[ figure_scenarios ] , that the unstable separatrix of a given saddle point needs to end up as a stable separatrix of another saddle point .
such a behavior occurs usually only when the underlying differential equations are invariant under certain symmetry operations , like the exchange of variables @xcite . for any practical application
, these symmetries need to be broken and the limiting cycle will vanish together with the heteroclinic sequence .
it can however be restored in form of a heteroclinic channel , if the strength of the symmetry - breaking is not too strong , by adding a stochastic component to the dynamics . with noise ,
a trajectory loitering around a saddle point can explore a finite region of phase space close to the saddle point until it finds the unstable direction .
once the trajectory has found stochastically the unstable direction , it will leave the saddle point quickly along this direction in phase space and a heteroclinic channel is restored functionally . cognitive computation on the backbone of saddle point networks
is therefore essentially based on an appropriate noise level .
cognitive computation with saddle point networks has been termed ` winnerless competition ' in the context of time encoding of natural stimuli @xcite and applied to the decision making problem . in the later case interaction with the environment
may generate a second unstable direction at the saddle points and decision taking corresponds to the choice of unstable separatrix taken by the trajectory @xcite .
a trivial form of self - sustained neural activity occurs in attractor networks @xcite .
starting with any given initial state the network state will move to the next attractor and stay there , with all neurons having a varying degree of constant firing rates , the very reason attractor nets have been widely discussed as prototypes for the neural memory @xcite . as such , an attractor network is useless for a cognitive system , as it needs outside help , or stimuli from other parts of the system , to leave the current attractor .
there is a general strategy which transforms an attractor network into one exhibiting transient state dynamics , with the transient neural states corresponding to the fixpoints of the original attractor network .
this procedure is applicable to a wide range of attractor networks and consists in expanding the phase space by introducing additional local variables akin to local activity reservoirs @xcite . to be concrete ,
let us denote with @xmath2 the set of dynamical variables of the attractor network , as illustrated in fig .
[ figure_scenarios ] , and by @xmath3 the additional reservoir variables .
we assume that the reservoirs are depleted / filled when the neuron is active / inactive , @xmath4 together with a suitable coupling of the reservoir variables @xmath3 to the neural activities @xmath2 one can easily achieve that the fixpoints of the attractor networks become unstable , _ viz _ that they are destroyed , turning into attractor ruins or attractor relics .
this situation is illustrated in fig .
[ figure_scenarios ] . in the expanded phase space @xmath5
there are no fixpoints left .
it is not the case that the attractors would just acquire additional unstable directions , upon enlargement of the phase space , turning them into saddle points .
instead , the enlargement of the phase space destroys the original attractors completely .
the trajectories will however still slow down considerably close to the attractor ruins , as illustrated in fig .
[ figure_transient_states ] , if the reservoirs are slow variables , changing only relatively slowly with respect to the typical time constants of the original attractor network . in this case the time constant @xmath6 entering the time evolution of the reservoir , eq .
( [ eq_general_dot_phi ] ) , is large . in the limit @xmath7
the reservoir becomes static and the dynamics is reduced to the one of the original attractor network .
the dynamics exhibited by attractor relic networks is related to the notion of chaotic itinerancy @xcite , which is characterized by trajectories wandering around chaotically in phase space , having intermittent transient periods of stability close to attractor ruins .
here we consider the case of attractor relics arising from destroyed point attractors . in the general case one
may also consider , e.g. , limit cycles or strange attractors .
the coupling to slow variables outlined here is a standard procedure for controlling dynamical systems @xcite , and has been employed in various fashions for the generation and stabilization of transient state dynamics .
one possibility is the use of dynamical thresholds for discrete - time rate - encoding neural nets @xcite . in this case one
considers as a slow variable the sliding - time averaged activity of a neuron and the threshold of a neuron is
increased / decreased whenever the neuron is active / inactive for a prolonged period .
another approach is to add slow components to all synaptic weights for the generation of an externally provided temporal sequence of neural patterns @xcite . in the following
we will outline in some detail an approach for the generation of transient state dynamics which takes an unbiased clique encoding neural net as its starting point @xcite , with the clique encoding network being a dense and homogeneous associative network ( dhan ) .
transient state dynamics is intrinsically competitive in nature .
when the current transient attractor becomes unstable , the subsequent transient state is selected via a competitive process .
transient - state dynamics is a form of ` multi - winners - take - all ' process , with the winning coalition of dynamical variables suppressing all other competing activities @xcite .
competitive processes resulting in quasi - stationary states with intermittent burst of changes are widespread , occurring in many spheres of the natural or the social sciences . in the context of darwinian evolution , to give an example , this type of dynamics has been termed ` punctuated equilibrium ' @xcite . in the context of research on the neural correlates of consciousness
, these transiently stable states in form of winning coalitions of competing neural ensembles have been proposed as essential building blocks for human states of the mind @xcite . the competitive nature of transient state dynamics
is illustrated in fig .
[ figure_competition ] , where a representative result of a simulation for a dhan net is presented . during the transition from one winning coalition to the subsequent
, many neurons try to become members of the next winning coalition , which in the end is determined by the network geometry , the synaptic strengths and the current reservoir levels of the participating neurons .
the transition periods from one transient state to the next are periods of increased dynamical sensibility . when coupling the network to sensory inputs , the input signal may tilt the balance in this competition for the next winning coalition , modulating in this way the ongoing internal dynamical activity .
transient state dynamics therefore opens a natural pathway for implementing neural architectures for which , as discussed in the introduction , the eigendynamics is modulated , but not driven , by the sensory data input stream .
a concrete example of how to implement this procedure will be discussed in sect .
[ sect_influence_stimuli ] .
only a small fraction of all neurons are active at any time in the brain in general , and in areas important for the memory consolidation in particular @xcite . for various reasons , like the optimization of energy consumption and the maximization of computational capabilities @xcite , sparse coding is an ubiquitous and powerful coding strategy @xcite .
sparse coding may be realized in two ways , either by small non - overlapping neural ensembles , as in the single - winner - take - all architecture , or by overlapping neural ensembles .
the latter pathway draws support from both theory considerations @xcite , and from experimental findings .
experimentally , several studies of the hippocampus indicate that overlapping neural ensembles constitute important building blocks for the real - time encoding of episodic experiences and representations @xcite .
these overlapping representations are not random superpositions but associatively connected .
a hippocampal neuron could response , e.g. , to various pictures of female faces , but these pictures would tend to be semantically connected , e.g. they could be the pictures of actresses from the same tv series @xcite .
it is therefore likely that the memory encoding overlapping representations form an associative network , a conjecture that is also consistent with studies of free associations @xcite .
there are various ways to implement overlapping neural encoding with neural nets . here
we discuss the case of clique encoding .
the term clique stems from graph theory and denotes , just as a clique of friends , a subgraph where ( a ) every member of the clique is connected with all other members of the clique and where ( b ) all other vertices of the graph are not connected to each member of the clique . in fig .
[ figure_clique_encoding ] a small graph is given together with all of its cliques .
also shown in fig .
[ figure_clique_encoding ] are the associative interconnections between the cliques .
one may view the resulting graph , with the cliques as vertices and with the inter - clique associative connections as edges , as a higher - level representation of an implicit hierarchical object definition @xcite . the clique ( 4,5,9 ) in the original graph in fig .
[ figure_clique_encoding ] corresponds to a primary object and the meta - clique [ ( 4,5,9)-(2,4,6,7)-(4,5,6,8 ) ] in the graph of the cliques would in this interpretation encode a meta object , composed of the primary objects ( 4,5,9 ) , ( 2,4,6,7 ) and ( 4,5,6,8 ) .
this intrinsic possibility of hierarchical object definitions when using clique encoding has however not yet be explored in simulations and may be of interest for future studies .
cliques can be highly overlapping and there can be a very large number of cliques in any given graph @xcite .
we will construct now a neural net where the cliques of the network are the attractors .
it is a homogeneously random and dense associative network ( dhan ) , where the associative relations between cliques are given by the number of common vertices . starting from this attractor network we will introduce slow variables , as discussed in sect . [ subsec_attractor_relic_networks ] , in terms of local reservoirs .
the network will then show spontaneously generated transient state dynamics , with the neural cliques as the attractor ruins . in a second step
we will couple the dhan net to sensory stimuli and study the interplay between the internal autonomous dynamical activity and the data input stream .
we will find that the cliques acquire semantic content in this way , being mapped autonomously to the statistically independent patterns of the data input stream .
the starting point of our considerations is the underlying attractor network , for which we employ a continuous time formulation , with rate encoding neurons , characterized by normalized activity levels @xmath8 $ ] .
the growth rates @xmath9 govern the respective time developments , @xmath10 when @xmath11 , the respective neural activity @xmath12 increases , approaching rapidly the upper bound ; when @xmath13 , it decays to zero .
we split the rates into three contributions : @xmath14 an internal positive contribution @xmath15 , an internal negative contribution @xmath16 , and the influence of external stimuli , @xmath17 .
we will discuss the influence of non - trivial external stimuli in sect .
[ sect_influence_stimuli ] , for the moment we consider @xmath18 . the division into an exciting and a depressing contribution in eq .
( [ eq_r_pos_neg_contribution ] ) reflects on one side the well known asymmetry between excitatory neurons and inhibitory interneurons in the brain @xcite and is on the other side essential for clique encoding .
the @xmath19 are determined via @xmath20 by the influence of the excitatory synaptic weights , @xmath21 .
the function @xmath22 entering eq .
( [ eq_r_i_plus ] ) couples the dynamics of the neurons locally to the slow variables @xmath23 .
we will examine the reservoir function @xmath22 in the next section .
for the time being we set @xmath24 , the primary neural dynamics is then decoupled from the reservoir dynamics and we will retain the starting attractor network . the @xmath25 are given by @xmath26 where the @xmath27 are the inhibitory synaptic weights and where @xmath28 sets the scale for the inhibition .
( [ eq_r_i_neg ] ) leads to a normalization @xmath29 .
we postpone the discussion of the reservoir function @xmath30 and consider for the time being @xmath31 .
clique encoding corresponds to a multi - winners - take - all formulation .
an inhibitory background is therefore necessary .
the dhan architecture contains hence an inhibitory link @xmath32 whenever there is no excitatory link @xmath33 , @xmath34 _ viz _ the excitatory links are shunting the inhibitory synapses .
this inhibitory background is implicitly present for the 9-site network shown in fig . [ figure_clique_encoding ] .
the edges of the network shown in fig .
[ figure_clique_encoding ] .
correspond to excitatory links @xmath35 .
all pairs of sites not connected by an edge in fig . [ figure_clique_encoding ] inhibit each other via @xmath36 .
the formulation of the attractor network with clique encoding is such complete @xcite .
all members of a given clique excite each other via intra - clique @xmath35 .
neurons which are not members of the current active clique are suppressed by at least one inhibitory link @xmath36 .
this suppression @xmath37 , compare eq .
( [ eq_r_i_neg ] ) , dominates the residual positive signal the out - of - clique neuron may receive , whenever @xmath38 is large enough .
an interesting feature of the dhan architecture is the absence of a bias in eq .
( [ eq_x_dot ] ) .
there is no self excitation or suppression , @xmath39 .
the dynamics of an individual neuron is exclusively driven by the influence of the network , it has no preferred firing state .
this feature would correspond for real - world neurons to the existence of a background of afferent activities with a level close to the firing threshold .
next we note , that the separation of scales @xmath40 implies that hebbian - type modification of the inhibitory links @xmath32 would be meaningless , small changes of a relatively large quantity will not lead to a substantial effect .
hebbian learning in the dhan architecture is therefore operational only for the excitatory links @xmath33 , in accordance to the general assumption that most learning taking place in the brain involves synapses interconnecting neurons and not interneurons @xcite . in sect .
[ sect_influence_stimuli ] we will consider the synaptic plasticity of links afferent to the dhan layer .
unsupervised and local hebbian - style learning can however be implemented easily for the intra - dhan excitatory links @xmath33 for unsupervised and homeostatic calibration of the excitatory links @xcite .
it is however not essential for the occurrence and for the stabilization of transient state dynamics , our focus here .
we consider normalized slow variables @xmath41 $ ] , with the time evolution @xmath42 where a neuron is active / inactive whenever its activity level @xmath12 is close to unity / zero .
the @xmath23 behave functionally as reservoirs , being depleted / refilled for active / inactive neurons .
the term @xmath43 on the rhs of eq .
( [ eq_dot_phi ] ) is not essential for the establishment of transient state dynamics , but opens an interesting alternative interpretation for the slow variables .
@xmath43 vanishes for inactive neurons and takes the value @xmath44 for active neurons .
the reservoir levels @xmath45 of all active neurons are drawn together consequently .
all members of the currently active winning coalition have then similar reservoir levels after a short time , on the order of @xmath46 .
this is a behavior similar to what one would expect for groups of spiking neurons forming winning coalitions via synchronization of their spiking times . for each neuron of the winning coalitions
one could define a degree of synchronization , given by the extent this neuron contributes to the overall synchronization .
initially , this degree of synchronization would have a different value for each participating neuron . on a certain timescale , denoted here by @xmath46 , the spiking times
would then get drawn together , synchronized , and all members of the winning coalition of active neurons would then participate to a similar degree in the synchronized firing .
the firing of the winning coalition would however not remain coherent forever .
internal noise and external influences would lead to a desynchronization on a somewhat longer time scale @xmath47 .
when desynchronized , the winning coalition would loose stability , giving way to a new winning coalition .
in this interpretation the reservoirs allow for a `` poor man s '' implementation of self organized dynamical synchronization of neural ensembles , a prerequisite for the temporal coding hypothesis of neural object definition @xcite .
finally we need to specify the reservoir coupling functions @xmath22 and @xmath30 entering eqs .
( [ eq_r_i_plus ] ) and ( [ eq_r_i_neg ] ) .
they have sigmoidal form with @xmath48 and a straightforward interpretation : it is harder to excite a neuron with depleted reservoir , compare eq .
( [ eq_r_i_plus ] ) , and a neuron with a low reservoir level has less power to suppress other neurons , see eq .
( [ eq_r_i_neg ] ) .
reservoir functions obeying the relation ( [ eq_res_functions ] ) therefore lead in a quite natural way to transient state dynamics . on a short time scale the system relaxes towards the next attractor ruin in the form of a neural clique .
their reservoirs then slowly decrease and when depleted they can neither continue to mutually excite each other , nor can they suppress the activity of out - of - clique neurons anymore . at this point , the winning coalition becomes unstable and a new winning coalition is selected via a competitive process , as illustrated in fig .
[ figure_competition ] .
any finite @xmath49 leads to the destruction of the fixpoints of the original attractor network , which is thus turned into an attractor relic network .
the sequence of winning coalitions , given by the cliques of the network , is however not random .
subsequent active cliques are associatively connected . the clique ( 1,9 ) of the 9-site network shown in fig . [ figure_clique_encoding ] , to give an example ,
could be followed by either ( 4,5,9 ) or by ( 1,2,3 ) , since they share common sites .
the competition between these two cliques will be decided by the strengths of the excitatory links and by the history of previous winning coalitions .
if one of the two cliques had been activated recently , the constituent sites will still have a depressed reservoir and resist a renewed reactivation . and cliques ( 0,1,2 ) ,
... receives sensory signals via the input layer ( middle ) in the form of certain input patterns ( bottom ) .
, scaledwidth=48.0% ] the finite state dynamics of the dhan architecture is robust . for the isolated network
, we will discuss the coupling to sensory input in the next section , the dynamics is relaxational and dissipative @xcite .
the system relaxes to the next attractor relic and the reservoirs are relaxing either to zero or to unity , depending on the respective neural activity levels . for a network with a finite number of sites
, the long - time state will be a long limiting cycle of transient states . the simulation results shown in fig .
[ figure_competition ] are for a set of parameters resulting in quite narrow transitions and long plateaus @xcite .
the formulation presented here allows for the modelling of the shape of the plateaus and of other characteristics of the transient state dynamics .
a smaller @xmath6 would result in shorter plateaus , a longer @xmath46 in longer transition times .
one can , in addition , adjust the shape of the reservoir functions and details of eqs .
( [ eq_r_i_plus ] ) and ( [ eq_r_i_neg ] ) in order to tune the overall competition for the next winning coalition .
the dhan architecture providing therefore a robust framework for the generation of transient state dynamics , offering at the same time ample flexibility and room for fine tuning , paving the way for a range of different applications .
the transient state dynamics generated by the dhan architecture is dynamically robust .
the dhan dynamics has at the same time windows of increased sensibility to outside influences during the transition periods from one transient state to the subsequent , as shown in fig .
[ figure_competition ] .
these transition periods are phases of active inter - neural competition , reacting sensibly to the influence of afferent signals .
we couple the input signals via an appropriate input layer , as illustrated in fig .
[ figure_dhan_input ] , denoting by @xmath50 $ ] the time dependent input signals , which we will take as black - and - white or grey - scaled patterns .
we denote by @xmath51 the afferent links to the dhan layer , with the external contribution to the dhan - layer growth rates , compare eq .
( [ eq_r_pos_neg_contribution ] ) , given by @xmath52 the rationale behind this formulation is the following .
the role of the input signal is not to destabilize the current winning coalition , the afferent signal is therefore shunted off in this case , eq .
( [ eq_r_i_ext ] ) .
the input signal should influence the competition for the next winning coalition , modulating but not driving directly the dhan dynamics .
this rational is realized by the above formulation .
inactive neurons will receive a bias @xmath53 from the input layer which increases / decreases its chance of joining the next winning coalition for @xmath54 / @xmath55 .
a cognitive system with a non - trivial and self - sustained internal neural activity has to decide how and when correlations with the sensory data input stream are generated via correlations encoded in the respective synaptic plasticities .
this is clearly a central issue , since the input data stream constitutes the only source for semantic content for a cognitive system .
it makes clearly no sense if the afferent links to the dhan layer , _ viz _ the links leading from the input to the internal network supporting a self - sustained dynamical activity , would be modified continuously via hebbian - type rules , since the two processes , the internal and the environmental dynamics , are per se unrelated .
it makes however sense to build up correlation whenever the input has an influence on the internal activity , modulating the ongoing associative thought process .
from the perspective of the cognitive system such a modulation of the internal dynamics by environmental stimuli corresponds to something novel and unexpected happening .
novelty detection is therefore vital for neural networks with a non - trival eigendynamics processing sensory data .
the importance of novelty detection for human cognition has been acknowledged indeed since long @xcite , and a possible role of dopamine , traditionally associated with reinforcement reward transmission @xcite , for the signalling of novelty has been suggested recently @xcite .
the influence of modulating and of not modulating sensory signals is illustrated in fig .
[ figure_input_yes_no ] , where simulation results for a dhan layer containing seven neurons coupled to an intermittent input signal are presented .
the signal is not able to deactivate a currently stable winning coalition , compare eq .
( [ eq_delta_r_i ] ) , but makes an impact when active during a transition period .
the system has the possibility to figure out whenever the later has happened .
when the input signal @xmath17 is relevant then @xmath56 in this case the internal contribution @xmath57 to the growth rate is negative and the input makes a qualitative difference .
we may therefore define a global novelty signal @xmath58 obeying @xmath59 where we have used eq .
( [ eq_r_pos_neg_contribution ] ) , @xmath60 , and where a @xmath61 is implicit on the rhs of the equation .
the novelty signal needs to be activated quickly , with @xmath62 .
learning then takes place whenever the novelty signal @xmath63 exceeds a certain threshold . of a dhan layer containing seven neurons ,
compare fig .
[ figure_dhan_input ] , the growth rates @xmath9 and the contributions from the input - layer @xmath53 , see eq .
( [ eq_delta_r_i ] ) .
the first input stimulus does not lead to a deviation of the transient state dynamics of the dhan layer .
the second stimulus modulates the ongoing transient state dynamics , influencing the neural competition during the sensitive phase .
, scaledwidth=48.0% ] bars problem .
, scaledwidth=30.0% ] bars problem .
, scaledwidth=48.0% ] having determined when learning takes place , we have now to formulate the rules governing how learning modifies the links afferent to the dhan layer .
for this purpose we will use the hebbian principle , that positive interneural correlations are enforced and negative correlations weakened .
our system is however continuously active , at no point are activities or synaptic strengths reset .
the hebbian principle therefore needs to be implemented as an optimization process @xcite , and not as a maximization process , which would lead to a potentially hazardous runaway growth of synaptic strengths .
there are four quadrants in the @xmath64 hebbian learning matrix , corresponding to active / inactive pre- and post - synaptic neurons , out of which we use the following three optimization rules : \(a ) the sum over active afferent links leading to active dhan neurons is optimized to a large but finite value @xmath65 , @xmath66 \(b ) the sum over inactive afferent links leading to active dhan neurons is optimized to a small value @xmath67 , @xmath68 \(c ) the sum over active afferent links leading to inactive dhan neurons is optimized to a small value @xmath69 , @xmath70 the @xmath65 , @xmath69 and @xmath67 are the target values for the respective optimization processes , where the superscripts stand for ` active ' , ` inactive ' and ` orthogonal ' @xcite .
these three optimization rules correspond to fan - in normalizations of the afferent synapses .
positive correlations are build up whenever @xmath65 dominates in magnitude , and orthogonalization of the receptive fields to other stimuli is supported by @xmath67 .
a small but non - vanishing value for @xmath69 helps to generate a certain , effective , fan - out normalization , avoiding the uncontrolled downscaling of temporarily not needed synapses .
knowledge about the environment lies at the basis of all cognition , before any meaningful action can be taken by a cognitive system . for simple organisms this knowledge
is implicitly encoded in the genes , but in general a cognitive system needs to extract this information autonomously from the sensory data input stream , via unsupervised online learning .
this task includes signal separation and features extraction , the identification of recurrently appearing patterns , i.e. of objects , in the background of fluctuation and of combinations of distinct and noisy patterns . for the case of linear signal superposition
this problem is addressed by the independent component analysis @xcite and blind source separation @xcite , which seeks to find distinct representations of statistically independent input patterns . in order to examine how our system of an input layer coupled to a dhan layer , as illustrated in fig .
[ figure_dhan_input ] , analyzes the incoming environmental signals , we have selected the bars problem @xcite .
the bars problem constitutes a standard non - linear reference task for feature extraction via a non - linear independent component analysis for an @xmath71 input layer .
the basic patterns are the @xmath72 vertical and @xmath72 horizontal bars and the individual input patterns are made up of a non - linear superposition of the @xmath73 basic bars , containing any of them with a certain probability @xmath74 , typically @xmath75 , as illustrated in fig .
[ figure_star_pattern ] .
our full system then consist of the dhan layer , which is continuously active , and an input layer coding the input patterns consisting of randomly superimposed black / white bars .
for the dhan network we have taken a regular 20-site ring , containing a total of 10 cliques @xmath76 , each clique having @xmath77 sites , as illustrated in fig .
[ figure_star_pattern ] .
the self - sustained transient - state process is continuously active in the dhan layer , modulated by the contributions @xmath78 it receives via the links @xmath51 from the input layer . for the simulation
a few thousands of input patterns were presented to the system @xcite . in fig .
[ figure_crp_graph ] we present for the @xmath79 bars problem the simulation results for the susceptibility @xmath80 of the 10 cliques @xmath81 in the dhan layer to the 10 basic patterns @xmath82 , the 10 individual horizontal and vertical bars , with @xmath83 , @xmath84 , and so on .
all cliques have the size @xmath85 and the notation @xmath86 denotes the set of all sites defining the clique @xmath81 . at the start
all @xmath51 are drawn randomly .
the result is quite remarkable . at the beginning of the simulation the system undergoes an associative thought process without semantic
content . during the course of the simulation , via the competitive novelty learning scheme , the individual attractor relics of the transient state dynamics , the cliques of the dhan layer , acquire a semantic connotation , having developed pronounced susceptibilities to statistically distinct objects in the sensory data input stream .
this can be seen directly inspecting the clique receptive fields @xmath87 of the @xmath88 cliques in the dhan layer with respect to the @xmath89 input neurons , which are also presented in fig .
[ figure_crp_graph ] .
the clique receptive fields @xmath90 correspond to the averaged receptive fields of their constituent neurons .
the data presented in fig . [ figure_crp_graph ]
are for the @xmath79 bars problem .
we note that simulation for larger systems can be performed as well , with similar results @xcite .
the learning scheme employed here is based on optimization and not on maximization , as stressed in sect .
[ subsec_afferent_link_plasticity ] .
the clique receptive fields , shown in fig
. [ figure_crp_graph ] , are therefore not of black / white type , but differentiated .
synaptic modifications are turned progressively off when sufficient signal separation has been achieved .
this behavior is consistent with the ` learning by error ' paradigm @xcite , which states that a cognitive system learns mostly when making errors and not when performing well .
we may take a look at the results presented in fig .
[ figure_crp_graph ] from a somewhat larger perspective .
the neural activity of newborn animals consists of instinct - like reflexes and homeostatic regulation of bodily functions .
the processing of the sensory signals has not yet any semantic content and internal neural activity states do not correspond yet to environmental features like shapes , colors and objects .
the neural activity can acquire semantic content , philosophical niceties apart , only through interaction with the environment .
this is a demanding task , since the optical or acoustical sensory signals are normally overloaded with a multitude of overlapping primary objects .
the animal therefore needs to separate these non - linearly superposed signals for the acquisition of primary knowledge about the environment and to map the independent signals , the environmental object to distinct neural activity patters .
this very basic requirement is performed by the dhan architecture .
the internal transient states have , at the start of the simulation , no relation to environmental objects and are therfore void of semantic content . in the simulation presented here
, there are 10 primary environmental objects , the 5 horizontal and vertical bars of the @xmath79 bars problem . in the setting
used these 10 objects are independent and statistically uncorrelated . during the course of the unsupervised and online learning process
, the receptive fields of the transiently stable neural states , the cliques in the dhan layer , acquire distinct susceptibilities not to arbitrary superpositions of the primary objects but to the individual primary bars themselves . a sensory signal consisting of the non - linear superposition of two or more bars
will therefore lead , in general , to the activation of one of the corresponding cliques . to be concrete , comparing fig .
[ figure_crp_graph ] , an input signal containing both the top - most and the bottom - most horizontal bar would activate either the clique @xmath91 or the clique @xmath92 .
these two cliques will enter the competition for the next winning coalition whenever the input is not too weak and when it overlapps with a sensitive period .
the present state together with its dynamical attention field @xcite will then determine the outcome of this competitions and one of the two objects present in this input signal is then recognized .
the vast majority of neural nets considered to date for either research purposes , or for applications , are generalized stimulus - response networks @xcite .
one has typically an input signal and an output result , as , e.g. , in speech recognition . in most settings the network is reset to a predefined default state after a given task
is completed , and before the next input signal is provided .
this approach is highly successful , in many instances , but it is clearly not the way the brain works on higher levels . it is therefore important to examine a range of paradigmal formulations for the non - trivial eigendynamics of cognitive systems , evaluating their characteristics and computational capabilities . as an example for a concept situated somewhere in between a pure stimulus response net and systems with a fully developed eigendynamics , we have discussed in sect . [ subsec_reservoir_computing ] the notion of reservoir computing . for reservoir networks
the dynamics is , in general , still induced by the input signal and decays slowly in the absence of any input .
any given stimulus encounters however an already active reservoir net , with the current reservoir activity caused by the preceding stimuli .
the response of the network therefore depends on the full history of input signals and time prediction tasks constitute consequently the standard applications scenaria for reservoir computing .
a somewhat traditional view , often presumed implicitly , is that the eigendynamics of the brain results from the recurrent interlinking of specialized individual cognitive modules .
this viewpoint would imply , that attempts to model the autonomous brain dynamics can be considered only after a thorough understanding of the individual constituent modules has been achieved .
here we have examined an alternative route , considering it to be important to examine the mutual benefits and computational capabilities of a range of theory proposals for the overall organization of the eigendynamics . in sect .
[ subsec_saddle_point_networks ] we have examined a first proposal for the organization of the eigendynamics in terms of saddle point networks . in this framework
the internal neural dynamics is guided by heteroclines in a process denoted winnerless competition .
this neural architecture aims to model reproducible cognitive behavior and a single robust attractor in terms of a heteroclinic channel constitutes the eigendynamics in the absence of sensory inputs . in sect .
[ subsec_attractor_relic_networks ] we have examined the viewpoint that a non - trivial associative thought process constitutes the autonomous dynamics in the absence of sensory input . for any finite ( and isolated ) network
these thought processes turn eventually into limiting cycles of transient states . in this architecture
there is however not a unique limiting cycle , but many possible and overlapping thought processes , every one having its respective basin of attractions .
the transient state dynamics required for this approach is obtained by coupling an attractor network to slow variables , with the neural time evolution slowing down near the such obtained attractor relics .
this is a quite general procedure and a wide range of concrete implementations are feasible for this concept .
the coupling of neural nets having a non - trivial eigendynamics to the sensory input is clearly a central issue , which we have discussed in depth in sect .
[ sect_influence_stimuli ] , for the case of networks with transient state dynamics based on attractor ruins , emphasizing two functional principles in this context : \(a ) the internal transient state dynamics is based intrinsically on the notion of competitive neural dynamics .
it is therefore consistent to assume that the sensory input contributes to this neural competition , modulating the already ongoing internal neural competition .
the sensory input would therefore have a modulating and not a forcing influence .
the sensory signals would in particular not deactivate a currently stable winning coalition , influencing however the transition from one transiently stable state to the subsequent winning coalition .
\(b ) the eigendynamics of the cognitive system and of the sensory signals resulting from environmental activities are , a priori , unrelated dynamically
. correlations between these two dynamically independent processes should therefore be built up only when a modulation of the internal neural activity through the sensory signal has actually occurred .
this modulation of the eigendynamics by the input data stream should then generate an internal reinforcement signal , which corresponds to a novelty signal , as the deviation of the internal thought process by the input is equivalent , from the perspective of the cognitive system , to something unexpected happening .
we have shown , that these two principles can be implemented in a straightforward manner , resulting in what one could call an ` emergent cognitive capability ' .
the system performs , under the influence of the above two general operating guidelines , autonomously a non - linear independent component analysis .
statistically independent object in the sensory data input stream are mapped during the life time of the cognitive system to the attractor relics of the transient state network .
the internal associative thought process acquires thus semantic content , with the time series of transient states , the attractor ruins , now corresponding to objects in the environment .
we believe that these results are encouraging and that the field of cognitive computation with autonomously active neural nets is an emerging field of growing importance
. it will be important to study alternative guiding principles for the neural eigendynamics , for the coupling of the internal autonomous dynamics to sensory signals and for the decision making process leading to motor output .
architectures built up of interconnected modules of autonomously active neural nets may in the end open a pathway towards the development of evolving cognitive systems .
gros c. emotions , diffusive emotional control and the motivational problem for autonomous cognitive systems . in : vallverdu j , casacuberta d , editors .
handbook of research on synthetic emotions and sociable robotics : new applications in affective computing and artificial intelligence .
igi - global ; 2009 , in press .
fox md , corbetta m , snyder az , vincent jl , raichle me .
spontaneous neuronal activity distinguishes human dorsal and ventral attention systems .
proceedings of the national academy of sciences 2003;103:10046 - 10051 .
fox md , snyder az , vincent jl , corbetta m , van essen dc , raichle me . the human brain is intrinsically organized into dynamic , anticorrelated functional networks .
proceedings of the national academy of sciences 2005;102:96739678 .
lin l , osan r , shoham s , jin w , zuo w , tsien jz .
identification of network - level coding units for real - time representation of episodic experiences in the hippocampus .
proceedings of the national academy of sciences 2005;102:6125 - 613 . | the human brain is autonomously active . to understand the functional role of this self - sustained neural activity , and its interplay with the sensory data input stream , is an important question in cognitive system research and we review here the present state of theoretical modelling
. this review will start with a brief overview of the experimental efforts , together with a discussion of transient vs. self - sustained neural activity in the framework of reservoir computing
. the main emphasis will be then on two paradigmal neural network architectures showing continuously ongoing transient - state dynamics : saddle point networks and networks of attractor relics .
self - active neural networks are confronted with two seemingly contrasting demands : a stable internal dynamical state and sensitivity to incoming stimuli .
we show , that this dilemma can be solved by networks of attractor relics based on competitive neural dynamics , where the attractor relics compete on one side with each other for transient dominance , and on the other side with the dynamical influence of the input signals .
unsupervised and local hebbian - style online learning then allows the system to build up correlations between the internal dynamical transient states and the sensory input stream .
an emergent cognitive capability results from this set - up .
the system performs online , and on its own , a non - linear independent component analysis of the sensory data stream , all the time being continuously and autonomously active .
this process maps the independent components of the sensory input onto the attractor relics , which acquire in this way a semantic meaning . |
let @xmath0 be the the toric variety of a fan @xmath1 .
when @xmath0 is smooth and projective , batyrev @xcite defines a collection @xmath2 of 1-dimensional cones of @xmath1 to be a _ primitive collection _ provided it does not span a cone of @xmath1 but every proper subset does .
each primitive collection gives a _ primitive inequality _ , and one of the nice results of @xcite states that the nef cone of @xmath0 is defined by the primitive inequalities . for a proof ,
batyrev cited the work of oda and park @xcite and reid @xcite , without giving details .
the survey article @xcite by the first author notes that batyrev s theorem applies to simplicial projective toric varieties .
casagrande @xcite and sato @xcite explain how primitive collections relate to reid s paper @xcite , and kresch @xcite gives a proof in the smooth case .
however , a complete proof of batyrev s result in the simplicial case has never appeared in print . in this paper , we give two proofs of batyrev s theorem , one based on @xcite and the other on @xcite .
we also extend the definition of primitive collection to the non - simplicial case and show that primitive collections still have the required properties .
our results apply to all quasi - projective toric varieties whose fans have convex support of maximal dimension .
we use standard notation and terminology for toric varieties .
let @xmath3 and @xmath4 be dual lattices of rank @xmath5 with associated real vector spaces @xmath6 and @xmath7 .
let @xmath8 be a toric variety of a fan @xmath1 in @xmath9 .
we always assume that the support @xmath10 of @xmath1 is convex of dimension @xmath5 .
hence all maximal cones have dimension @xmath5 . given @xmath1 , @xmath11 denotes the set of @xmath12-dimensional cones of @xmath1 , and @xmath13 is the subset of @xmath11 consisting of @xmath12-dimensional cones not lying on the boundary of @xmath10 .
an _ interior wall _ is an element of @xmath14 .
we use the convention that @xmath15 will denote both an element of @xmath16 and its primitive generator in @xmath3 . the torus - invariant divisor associated to @xmath17
is denoted @xmath18
. also recall that a piecewise - linear function @xmath19 can be represented by giving @xmath20 for each @xmath21 , i.e. , @xmath22 if @xmath23 .
we define @xmath24 as the vector space of all piecewise - linear functions on @xmath1 .
the function @xmath19 is well - defined in @xmath24 if and only if the following statement holds : if @xmath25 is an interior wall and @xmath26 , @xmath27 are the @xmath5-dimensional cones on each side of @xmath25 , then @xmath28 .
the support function @xmath19 of a torus - invariant cartier divisor @xmath29 satisfies @xmath30 .
note the absence of minus signs . for us , @xmath19 is _ convex _ if and only @xmath31 for all @xmath32 .
we also define @xmath33 to be the cone consisting of all convex piecewise - linear functions on @xmath1 .
a function @xmath34 is _ strictly convex _ when @xmath35 for all @xmath32 not lying in the same cone of @xmath1 .
the toric variety @xmath0 is quasi - projective if and only if there exists a strictly convex @xmath34 . when this happens , the interior of @xmath36 is nonempty and consists of all strictly convex piecewise - linear functions in @xmath24 . in section 1
we give a new definition of primitive collection and state our main theorem .
we also recall the nef and mori cones and review the description of the mori cone in terms of the wall relations coming from interior walls . in section 2
we prove batyrev s theorem in the simplicial case , and then in section 3 we treat the non - simplicial case .
this section also studies how primitive collections for @xmath1 relate to primitive collections for a simplicial subdivision @xmath37 of @xmath1 .
the final section of the paper explores an open problem dealing with the quasi - projective hypothesis .
the nef cone @xmath38 of @xmath0 is the quotient of the cone @xmath39 by all linear functions on @xmath1 .
thus @xmath40 here is the central definition of our paper .
[ pcdef ] a subset @xmath41 is called a _ primitive collection _ for @xmath1 if @xmath42 is not contained in a single cone of @xmath1 but every proper subset is . in the smooth projective case ,
batyrev defined @xmath43 to be a primitive collection when @xmath44 does not generate a cone of @xmath1 but every proper subset does .
when @xmath1 is smooth or more generally simplicial , this is clearly equivalent to definition [ pcdef ] .
let @xmath45 be a primitive collection .
we say that @xmath34 satisfies the _ primitive inequality for _
@xmath46 if @xmath47 if @xmath34 is convex , i.e. , @xmath48 , then @xmath19 clearly satisfies the primitive inequality for every primitive collection . in other words ,
@xmath49 the main result of our paper is that the inclusion is in fact an equality , i.e. , the nef cone is defined by the primitive inequalities .
here is the precise statement .
[ mainthm ] let @xmath0 be a quasi - projective toric variety coming from the fan @xmath1 in @xmath50 .
if @xmath10 is convex of dimension @xmath5 , then @xmath51 section 2 will prove theorem [ mainthm ] when @xmath0 is simplicial and section 3 will treat the non - simplicial case .
the proof of the main theorem will use extremal rays .
hence we need recall the mori cone of a toric variety .
although this material is well - known to experts , we include many details since the results we need do not appear explicitly in the literature .
we begin with the exact sequence @xmath52 which dualizes to @xmath53 where @xmath54 denotes the dual of @xmath24 and @xmath55 is the dual of @xmath56 .
the cone @xmath57 contains the image of @xmath58 and hence dualizes to a cone @xmath59 we call @xmath60 the _ mori cone _ of @xmath0 . when @xmath0 is quasi - projective , @xmath38 has maximal dimension in @xmath56 , so that the mori cone @xmath61 is strongly convex .
the unique minimal generators of the mori cone are called _ extremal rays_. we now review the combinatorial description of @xmath60 in terms of the interior walls of @xmath1 .
the basic observation is that relations among elements of @xmath16 give elements of @xmath55 .
the map @xmath62 gives a commutative diagram @xmath63 & m_{\mathbb{r}}\ar[r ] & { \mathbb{r}}^{\fan(1 ) } \ar[r ] & a_{n-1}(x)_{\mathbb{r}}\ar[r ] & 0\\ 0 \ar[r ] & m_{\mathbb{r}}\ar[r ] \ar[u ] & { \mathrm{pl}}(\fan ) \ar[r ] \ar[u ] & { \mathrm{pic}}(x)_{\mathbb{r}}\ar[r ] \ar[u ] & 0 } \ ] ] where @xmath64 is the chow group of @xmath65-cycles modulo rational equivalence .
this dualizes to @xmath66 & a_{n-1}(x)_{\mathbb{r}}^ * \ar[r ] \ar[d ] & { \mathbb{r}}^{\fan(1 ) * } \ar[r ] \ar[d ] & n_{\mathbb{r}}\ar[r ] \ar[d ] & 0\\ 0 \ar[r ] & n_1(x ) \ar[r ] & { \mathrm{pl}}(\fan)^ * \ar[r ] & n_{\mathbb{r}}\ar[r ] \ar[u ] & 0 } \ ] ] in the top row , the map @xmath67 sends the standard basis element @xmath68 to @xmath69 .
thus @xmath70 can be interpreted as all linear relations among the @xmath71 , and the surjective map @xmath72 shows that all elements of @xmath55 come from linear relations among the @xmath73 .
interior walls of @xmath1 give the following linear relations .
given an interior wall @xmath25 , let @xmath26 and @xmath27 be the @xmath5-dimensional cones on each side of @xmath25 , i.e. , @xmath74 . pick @xmath75 linearly independent vectors @xmath76 in @xmath77 and pick vectors @xmath78 , @xmath79 .
then there is a nontrivial relation @xmath80 where the final condition holds since @xmath81 and @xmath82 lie on opposite sides of the wall . hence the coefficients @xmath83 are unique up to multiplication by a positive constant .
let @xmath84 have components @xmath85 for @xmath86 and @xmath87 otherwise . using the above diagram
, we see that @xmath88 .
[ ltau_def ] depending on the context , we use the term _ wall relation _ to refer to the equation , the vector @xmath89 , or its image @xmath90 .
notice that in the non - simplicial case , a given wall can have many choices for the @xmath91 in the wall relation , while in the simplicial case , @xmath91 are uniquely determined by the wall .
[ thm_walls_mori ] let @xmath1 be a fan in @xmath50 with convex support of dimension @xmath5 . 1 . for @xmath92 , the different choices of the wall relation all give the same @xmath93 up to a positive constant .
the mori cone in @xmath55 is given by @xmath94 this is the kleiman - nakai criterion from oda and park ( * ? ? ?
* thm.2.3 ) . we give the details since their definition of @xmath95 differs from ours .
let @xmath96 and pick a wall relation @xmath97 , @xmath98 as in .
rescaling by a positive constant , we may assume @xmath99 , so that @xmath100 then @xmath101 is the linear functional on @xmath24 given by @xmath102 since @xmath103 for @xmath104 and @xmath105 since @xmath106 .
however , implies that @xmath107 hence the above formula for @xmath108 simplifies to @xmath109 note that @xmath110 .
since @xmath111 and @xmath112 is 1-dimensional , it follows that up to a positive constant , @xmath113 for _ any _ @xmath114 .
this proves the first part of the theorem and also shows that our @xmath95 agrees with the @xmath95 appearing in the statement of ( * ? ? ?
then the second part follows immediately from ( * ? ? ? * thm.2.3 ) .
let @xmath115 be a primitive collection for @xmath1 .
then @xmath116 lies in some unique minimal cone @xmath26 of @xmath117 .
pick a subset @xmath118 satisfying @xmath119 the equation gives the vector @xmath120 defined by @xmath121 from , it follows that @xmath122 .
depending on the context , we use the term _ primitive relation _ to refer to the equation , the vector @xmath123 , or its image @xmath124 under the map @xmath72 .
when @xmath0 is smooth , the case @xmath125 can not occur ( * ? ? ?
3.1 ) , but happens frequently in the simplicial case . in this case , we bound @xmath126 as follows .
[ poscoeffs ] the coefficients in the primitive relation satisfy @xmath127 when @xmath125 .
hence @xmath128 is determined by the positive entries of @xmath129 .
suppose @xmath130 and @xmath131 .
subtracting @xmath132 from each side of gives @xmath133 since @xmath134 lie in a cone of @xmath117 , this equation implies that @xmath134 and @xmath135 lie in the same cone of @xmath117 .
adding in @xmath132 shows that @xmath128 lies in a cone of @xmath117 , which is impossible .
the minimal cone @xmath26 containing @xmath136 need not be simplicial , so there may be many subsets @xmath137 satisfying .
but when there are many choices for @xmath129 , they all give the same element @xmath124 , as shown by the following proposition .
[ primrel_unique ] let @xmath138 be a primitive collection for @xmath1 and let @xmath139 be defined as above .
then : 1 . when regarded as an element of @xmath54 , @xmath140 is the linear functional on @xmath24 defined by @xmath141 2 .
@xmath142 .
let @xmath143 be the smallest cone containing @xmath116 .
since @xmath19 is linear on @xmath26 and @xmath118 , we obtain @xmath144 this proves the first part of the proposition , and then the second part follows immediately from the first part and since @xmath145
. we can formulate theorem [ mainthm ] in terms of primitive relations as follows .
[ equivalent ] let @xmath0 be the toric variety of the fan @xmath1 in @xmath50
. if @xmath10 is convex of dimension @xmath5 , then the following are equivalent : 1 .
@xmath146 for all primitive collections @xmath147 .
@xmath148 , where the sum is over all primitive collections for @xmath1 .
this follows easily from proposition [ primrel_unique ] and @xmath145 .
the strategy for proving theorem [ mainthm ] in the simplicial case is the observation , implicit in @xcite , that every minimial generator of @xmath60 is a primitive relation @xmath140 for some primitive collection @xmath128
. then theorem [ mainthm ] for simplicial fans follows immediately from proposition [ equivalent ] .
we give the details of this argument in section 2 . an interior wall
@xmath25 gives a complete torus - invariant curve @xmath149 in @xmath0 .
let @xmath150 denote the linear functional that sends an @xmath151-cartier divisor @xmath29 to the intersection product @xmath152 .
thus @xmath153 up to a positive multiple , this gives the same class as the wall relation @xmath90 from definition [ ltau_def ] .
although this result is well - known to experts , we include a proof for completeness .
[ ctaultau ] let @xmath1 be a fan in @xmath50 with convex support of dimension @xmath5 .
for each @xmath92 , @xmath154 is a positive multiple of the wall relation @xmath95 appearing in theorem [ thm_walls_mori ] .
when @xmath1 is simplicial , we have @xmath155 and as in the proof of theorem [ thm_walls_mori ] , we have the wall relation @xmath156 since @xmath1 is simplicial , the divisors @xmath157 corresponding to @xmath73 are @xmath158-cartier , so that @xmath159 is defined . by (
* ( 2.7 ) ) , we have @xmath160 the proof in @xcite assumes @xmath1 is simplicial and complete and @xmath25 is any wall ; the argument applies without change when @xmath1 is simplicial and @xmath25 is an interior wall . for the general case
, we use the well - known fact that @xmath1 has a simplicial refinement @xmath37 such that @xmath161 ( see corollary [ cor_simpl_ref_general ] and remark [ rem_simpl_ref_general ] ) .
if @xmath162 is the toric variety of @xmath37 , then we have a proper map @xmath163 let @xmath164 be an interior wall of @xmath37 contained in @xmath25 , and let @xmath165 and @xmath166 be the corresponding curves in @xmath162 and @xmath0 .
the induced map @xmath167 has degree @xmath168 $ ] , which implies that @xmath169 .
let @xmath29 be a cartier divisor on @xmath0 . by the projection formula ,
@xmath170 if we write @xmath171 on @xmath0 , then @xmath172 on @xmath162 since @xmath161 .
if @xmath173 is the wall relation of @xmath164 coming from , then up to a positive constant , @xmath174 since @xmath37 is simplicial .
however , the wall relation for @xmath164 is one of the ( possibly many ) wall relations for @xmath25 , i.e. , @xmath175 is one of the possible choices for @xmath176 . then the formula @xmath177 shows ( again up to a positive constant ) that the class of @xmath166 in @xmath55 is the image of @xmath178 in @xmath55 . in other words
, @xmath179 equals @xmath95 up to a positive constant , as claimed .
we conclude our discussion of the mori cone explaining how our definition of @xmath60 relates to the standard geometric approach . since @xmath0 need not be complete , we work in the relative context .
let @xmath180 be the affine toric variety of the strongly convex cone @xmath181 .
this gives a proper toric morphism @xmath182 .
following @xcite or @xcite , the mori cone of @xmath182 is defined as follows .
let @xmath183 be the free group generated by irreducible curves in @xmath0 that map to a point in @xmath180 .
then we have a natural pairing @xmath184 by restricting to torus - invariant curves coming from interior walls , one sees easily that this pairing is nondegenerate with respect to @xmath185 , i.e. , if a cartier divisor satisfies @xmath186 for all torus - invariant curves @xmath187 coming from interior walls , then @xmath188 = 0 $ ] in @xmath185 .
it follows that the above pairing induces a perfect pairing @xmath189 thus @xmath190 is what we call @xmath55 .
dropping the @xmath180 from the notation is reasonable since in our situation @xmath180 is determined functorially by @xmath0 . finally , @xmath191 is the cone generated by irreducible curves in @xmath0 that map to a point in @xmath180 , and the _ mori cone _ is its closure @xmath192 in @xmath55
. then theorem [ thm_walls_mori ] and proposition [ ctaultau ] easily imply that @xmath193 where @xmath194 is the class of the torus - invariant curve @xmath166 associated to @xmath25 .
this is the _ relative toric cone theorem_. the relative toric cone theorem is stated for toric morphisms @xmath195 by matsuki ( * ? ? ?
. 14 - 1 - 4 ) or reid ( * ? ? ?
* ( 1.7 ) ) . as pointed out by fujino and sato in (
4.3 ) , this fails when the torus action on @xmath137 has no fixed points .
they give the easy example of the projection map @xmath196 .
the fibers of this map are never torus - invariant , so that torus - invariant curves can not generate @xmath197 .
fortunately , the relative toric cone theorem holds for our map @xmath182 because @xmath198 is convex of dimension @xmath5 .
a nice feature of the simplicial case is that @xmath199 .
hence an interior wall @xmath25 gives @xmath200 , and a primitive collection @xmath128 gives @xmath201 .
let @xmath25 be an interior wall of a simplicial fan @xmath1 with @xmath155 , and let @xmath81 and @xmath82 be the generators that are needed to span the cones on each side of the wall .
the uniquely determined wall relation is @xmath202 by the discussion following .
[ reid_primcoll ] let @xmath1 be a quasi - projective simplicial fan with convex support of dimension @xmath5 .
let @xmath25 be an extremal wall , meaning that the wall relation generates an extremal ray @xmath203 .
then : 1 .
@xmath204 is a primitive collection for @xmath1 .
2 . in @xmath205 ,
the primitive relation @xmath129 of @xmath128 and wall relation @xmath178 of @xmath25 are equal up to a positive constant .
we give two proofs .
the first is based on ( * ? ? ?
2.4 ) , which assumes that @xmath0 is smooth and complete .
we adapt the argument to the simplicial case as follows .
we first make a useful observation about convex support functions .
if @xmath19 is convex and @xmath206 , we can change @xmath19 by a linear function so that @xmath207 now take an extremal wall @xmath25 with wall relation .
consider the set @xmath208 we will prove that @xmath128 is a primitive collection whose primitive relation @xmath140 equals @xmath95 up to a positive constant . recall that @xmath201 and @xmath209 since @xmath117 is simplicial .
we first prove by contradiction that @xmath210 for all @xmath143 .
suppose @xmath211 and take a strictly convex support function @xmath19 .
we may assume that @xmath19 is of the form .
since @xmath212 for @xmath213 , we have @xmath214 however , @xmath215 by , and @xmath216 implies @xmath217 for @xmath218 .
it follows that @xmath219 , which is impossible since @xmath19 is strictly convex and @xmath220 .
thus no cone of @xmath117 contains all rays in @xmath128 .
it follows that some subset @xmath221 is a primitive collection .
this gives the primitive relation @xmath222 , and we also have @xmath223 .
let @xmath224 where @xmath225 .
we claim that if @xmath226 is sufficiently small , then @xmath227 to prove this , suppose that @xmath228 and @xmath229
. then the definition of @xmath230 forces @xmath231 , so that @xmath232 by lemma [ poscoeffs ] . combined with @xmath221
, we see that @xmath233 by the definition of @xmath128 .
but we can clearly choose @xmath226 sufficiently small so that @xmath234 this inequality and the definition of @xmath230 imply @xmath235 , a contradiction .
we next claim that @xmath236 . by , we have @xmath237 where the second inclusion follows from and the definition of @xmath178 .
now let @xmath19 be convex . by with @xmath238
, we may assume @xmath239 then @xmath240 where the final inequality follows since @xmath215 and @xmath228 can only happen only when @xmath241 .
this proves that @xmath236 . since @xmath242 by proposition [ primrel_unique ] ,
the equation @xmath243 expresses @xmath209 as a sum of elements of @xmath60 . but
@xmath95 is extremal .
this forces @xmath244 and @xmath230 to lie in the ray generated by @xmath178 .
since @xmath244 is nonzero , @xmath178 is a positive multiple of @xmath244 .
in particular , they have the same positive entries .
then @xmath245 follows from the definition of @xmath128 and lemma [ poscoeffs ] .
this completes the first proof .
the second proof begins with the extremal wall relation .
for @xmath246 set @xmath247 in @xcite , reid proves that @xmath248 ( see the lemma on @xcite ) and @xmath249 ( see and ( * ? ? ?
reid assumes that @xmath1 is simplicial and complete .
his proofs generalize to our situation without change
see @xcite .
let @xmath250 , so that @xmath251 . in order to prove that @xmath128 is a primitive collection , we first show that @xmath252 is not a cone in @xmath1 .
so assume @xmath253 and consider the relation @xmath254 where the coefficients on the left are positive .
then @xmath255 lies in the relative interior of the cone @xmath256 , but @xmath257 lies in the wall @xmath258 since @xmath259 and @xmath260 for @xmath261 .
it follows that @xmath262 , which is a contradiction since @xmath81 and @xmath82 do not lie in the wall .
now we show that every proper subset of @xmath128 generates a cone of @xmath1 .
let @xmath263 be any proper subset of @xmath264 .
then @xmath265 is a face of @xmath266 for any @xmath267 . but
@xmath268 by .
hence @xmath269 is a primitive collection .
finally , we consider the primitive relation of @xmath128 , which can be written @xmath270 where @xmath26 is the minimal cone of @xmath1 containing @xmath271 . since @xmath272 and @xmath273 by and , it follows that is a relation among @xmath91 .
this relation is unique up to a nonzero constant since @xmath117 is simplicial .
thus @xmath129 is a nonzero multiple of @xmath178 , necessarily positive by lemma [ poscoeffs ] and the definition of @xmath128 .
batyrev clearly knew this proposition , though it is not stated explicitly in @xcite .
proposition [ reid_primcoll ] is closely related to theorem 1.5 in casagrande s paper @xcite and appears implicitly in the remarks preceeding proposition 2.2 in sato s paper @xcite .
we can now prove the simplicial case of our main theorem .
[ theorem_mori_primitive ] let @xmath1 be a simplicial quasi - projective fan in @xmath50 with convex support of dimension @xmath5 .
then the cone @xmath36 is defined by the primitive inequalities , i.e. , @xmath274 by proposition [ equivalent ] it suffices to show that the primitive relations @xmath140 generate the mori cone .
we already know that @xmath275 ( proposition [ primrel_unique ] ) and that @xmath60 is generated by the extremal wall relations @xmath95 ( theorem [ thm_walls_mori ] ) .
furthermore , @xmath60 is generated by extremal wall relations since @xmath0 is quasi - projective .
hence it suffices to show that every extremal wall relation is a primitive relation .
this is what we proved in corollary [ reid_primcoll ] , and the theorem follows . here is an example of theorem [ theorem_mori_primitive ] .
[ ex_simplicial ] figure [ figure : counterex_def_h ] shows the complete simplicial fan @xmath1 in @xmath276 with five minimal cone generators : @xmath277 and six maximal cones : @xmath278 ( 0,0),title="fig : " ] # 1#2#3#4#5 @font ( 945,1527)(1801,-2098 ) ( 2746,-691)(0,0)[lb ] ( 2476,-2041)(0,0)[lb ] ( 2656,-1006)(0,0)[lb ] ( 1801,-1006)(0,0)[lb ] ( 1936,-736)(0,0)[lb ] the primitive collections for this fan are : @xmath279 so that @xmath34 is convex if and only if @xmath280 to get a more concrete characterization , we use the associated primitive relations : @xmath281 let @xmath282 be the torus - invariant divisor associated to @xmath283
. then the divisor @xmath284 corresponds to the support function @xmath19 satisfying @xmath285 .
it follows that @xmath29 is nef if and only if @xmath286 in contrast , @xmath1 has 9 walls , so theorem [ thm_walls_mori ] describes @xmath60 using 9 generators , corresponding to 9 wall inequalities defining @xmath57 .
fortunately , these can be simplified considerably .
we denote by @xmath287 the wall that is spanned by @xmath283 and @xmath288 . by abuse of notation
we will also call @xmath287 the corresponding class in @xmath60 .
one can compute that the 9 walls fall into three groups : @xmath289 hence @xmath290 and @xmath291 give the extremal rays of the mori cone , while @xmath292 do not give an extremal ray .
one can check that the primitive collection @xmath293 generates the same ray as @xmath290 and @xmath294 generates the same ray as @xmath295 .
example [ ex_simplicial ] is nice because there were few primitive collections .
however , there are examples where the primitive collections vastly outnumber the interior walls .
[ polygon_example ] let @xmath1 be a complete fan in @xmath296 with @xmath297 minimal generators , say @xmath298 , arranged counterclockwise around the origin
. then there are @xmath299 walls , all interior .
one easily checks that the primitive collections are given by @xmath300 for @xmath301 and @xmath302 not adjacent .
hence the fan @xmath1 has @xmath303 primitive collections .
this is greater than the number of walls provided @xmath304 .
in order to prove our main theorem in the non - simplicial case , we need to consider simplicial refinements . here
we present a general theorem about the existence of simplicial refinements with special properties .
[ thrm_simpl_ref_general ] let @xmath1 be a fan in @xmath50 with convex support of dimension @xmath5 and fix @xmath305 such that @xmath306 is linearly independent for all @xmath143 .
then : 1 . there exists a simplicial refinement @xmath37 of @xmath1 satisfying @xmath161 such that @xmath307 generates a cone in @xmath37 for all @xmath143 .
if in addition @xmath1 is quasi - projective , then the refinement @xmath37 in part ( 1 ) can be chosen to be quasi - projective . to create @xmath37 , assign a weight @xmath308 to each @xmath73 as follows : * for @xmath309 , set @xmath310 . * for @xmath311 ,
pick @xmath312 generic .
the exact meaning of generic will be explained in the course of the proof . for each cone
@xmath21 , fix @xmath313 in the interior of the dual cone @xmath314 .
the affine hyperplane @xmath315 defined by @xmath316 intersects @xmath26 in a convex polytope @xmath317 with vertices @xmath318 , where @xmath319 for @xmath320 .
the idea is to triangulate @xmath321 using a variant of the method used in example 1.1 of @xcite , p. 215 .
consider @xmath322 since the vectors @xmath323 lie on the 1-dimensional rays of @xmath26 , it is easy to see that the vertices of @xmath324 consist of the origin and the points @xmath325 for @xmath326 .
furthermore , the faces of @xmath324 not containing the origin project to a polyhedral subdivision of @xmath321 . projecting from the origin in @xmath327
, we get a refinement @xmath328 of @xmath26 that satisfies @xmath329 .
figure [ figure : simplicial_proof_weights ] shows two 3-dimensional cones @xmath26 , each with a set @xmath330 and a choice of weights giving the polytope @xmath324 inside @xmath26 . ( 0,0 ) , @xmath330 , and a choice of weights.,title="fig : " ] # 1#2#3#4#5 @font ( 3914,3619)(1059,-3008 ) ( 1756 , 29)(0,0)[lb ] ( 1351,-1006)(0,0)[lb ] ( 1296,479)(0,0)[lb ] ( 3736 , 29)(0,0)[lb ]
( 3421,-1006)(0,0)[lb ] ( 4231,-1096)(0,0)[lb ] ( 3296,479)(0,0)[lb ] we claim that the fans @xmath328 have the following three properties : * @xmath331 generates a cone of @xmath328 for all @xmath143 . *
if a face @xmath25 lies in maximal cones @xmath26 and @xmath27 , then @xmath328 and @xmath332 induce the same refinement of @xmath25 . *
if the @xmath308 are sufficiently generic for @xmath333 , then @xmath328 is simplicial for all @xmath143 . assuming a c , the set @xmath334 is a fan that refines @xmath1 by b and satisfies @xmath335
furthermore , @xmath37 is simplicial by c. finally , given @xmath143 , @xmath331 generates a cone of @xmath37 by a. hence the proof of part ( 1 ) of the theorem will be complete once we prove a c .
_ proof of a_. consider the hyperplane @xmath315 . since @xmath336 for all @xmath15 , the polytope @xmath324 lies on the side of @xmath337 containing the origin , and
the intersection @xmath338 is clearly the convex hull of the points @xmath318 for @xmath339 by the choice of the weights @xmath308 .
it follows that @xmath330 generates a cone of @xmath328 .
_ proof of b_. suppose @xmath25 is a face of @xmath26 .
set @xmath340 and observe that @xmath341 this tells us that the refinement of @xmath25 induced by @xmath342 is determined entirely by the vectors @xmath323 for @xmath343 . since we are working with fixed weights in this part of the proof ,
the only choice involved is this refinement of @xmath25 comes from the vectors @xmath318 for @xmath344 .
recall that @xmath345 , where @xmath337 is defined by @xmath346 and @xmath347 .
now suppose that @xmath25 is a face of @xmath27 for a maximal cone @xmath348 .
the refinement of @xmath25 induced by @xmath332 is determined by @xmath349 where @xmath350 determines the affine hyperplane @xmath351 whose intersection with @xmath343 determines @xmath352 .
now consider the map @xmath353 defined as follows : * if @xmath354 , then @xmath355 .
* if @xmath354 and @xmath356 , then @xmath357 .
the map @xmath358 sends line segments to line segments and is homogeneous of degree @xmath359 for non - negative scalars .
furthermore , @xmath360 , and thus @xmath361 for all @xmath343 .
it follows that @xmath362 this map preserves faces , so that @xmath363 and @xmath364 induce the same refinement of @xmath25 .
this completes the proof of b. _ proof of c_. we may assume @xmath365 . for @xmath366 , write @xmath367 when we have @xmath368 , , we write instead @xmath369 and we set @xmath370 .
now suppose that @xmath342 is non - simplicial for some choice of weights @xmath308 .
this implies that @xmath324 has a face @xmath371 of dimension @xmath75 not containing the origin that is not an @xmath65-simplex .
it follows that @xmath371 has at least @xmath372 vertices .
pick @xmath372 vertices of @xmath371 as follows .
we first pick those vertices of @xmath371 of the form @xmath373 for @xmath374 .
there are at most @xmath5 such vertices since @xmath307 is linearly independent .
since @xmath375 , we can extend them to affinely independent vertices @xmath376 . since @xmath371 is non - simplicial
, we can pick one more , @xmath377 . now consider the @xmath378 matrix @xmath379 by construction , the vectors @xmath380 are affinely independent , but once we add the last vector , @xmath381 are affinely dependent since they lie in a @xmath65-dimensional face .
it follows that this matrix has rank exactly @xmath5 .
since the weights @xmath382 are nonzero , the same is true for the matrix @xmath383 the determinant of @xmath384 must vanish .
the resulting linear equation in the @xmath385 gives a necessary condition for @xmath342 to be non - simplicial .
furthermore , our careful choice of @xmath386 guarantees that @xmath387 actually appears in the determinant and that @xmath388 for some @xmath389 . since there are only finitely many maximal cones in @xmath1 and for each @xmath26 we have a fixed choice of the @xmath390 ( determined by the fixed choice of @xmath347 ) , we get a finite system of non - trivial linear equations in the @xmath391 for @xmath389 that give necessary conditions for @xmath342 to be non - simplicial . if we pick the weights @xmath308 to avoid these finitely many subspaces , the resulting subdivisions @xmath342 will be all simplicial .
this completes the proof of c , and part ( 1 ) follows .
turning to part ( 2 ) , we assume that @xmath1 is quasi - projective .
it suffices to find a simplicial refinement @xmath37 of @xmath1 satisfying @xmath161 such that the induced map @xmath392 is projective .
the latter happens when @xmath37 has a piecewise - linear function @xmath393 which is strictly convex relative to @xmath1 , meaning that for all @xmath143 , @xmath394 is strictly convex with respect to the subfan @xmath395 ( see ( * ) on page 27 and theorem 10 on pages 3132 of @xcite )
. since @xmath1 is quasi - projective , we can find @xmath48 which is strictly convex .
we first modify @xmath19 so that it takes positive values on @xmath16 . to see why this is possible ,
consider the cone @xmath396 since @xmath19 is strictly convex for @xmath1 , it follows that @xmath397 is a strongly convex cone with minimal generators given by @xmath398 for @xmath71 .
hence we can find @xmath399 such that @xmath400 replacing @xmath19 with @xmath401 , we may assume @xmath402 for all @xmath71 , as claimed .
the proof of part ( 1 ) used a hyperplane @xmath315 for each @xmath21 .
more precisely , we picked @xmath403 such that @xmath404 , and then for @xmath326 , @xmath318 was the unique vector in @xmath405 satisfying @xmath406 . using @xmath19 , we get a consistent set of hyperplanes since @xmath407 is linear , i.e. , @xmath408 for some @xmath409 .
our hypothesis that @xmath402 for all @xmath71 guarantees that @xmath313 is in the interior of @xmath410 .
hence we can use these @xmath313 s to give the hyperplanes @xmath337 .
then the point @xmath318 is the unique vector in @xmath405 satisfying @xmath411 .
now pick generic weights @xmath308 for @xmath311 .
then the properties a c are satisfied ( note b is now trivial because of our consistent choice of the @xmath337 ) .
thus we get a simiplicial refinement @xmath37 of @xmath1 that satisifes part ( 1 ) of the theorem .
now define @xmath412 by setting @xmath413 and extending linearly on each cone @xmath414 .
this gives a well - defined function in @xmath415 since @xmath37 is simplicial .
assuming @xmath19 is rational , we can also assume that the @xmath308 are rational .
hence we can assume that @xmath358 is rational as well .
we claim that @xmath358 is strictly convex with respect to @xmath416 for each @xmath143 . to see this , first observe that @xmath417 since @xmath411 .
it follows that inside @xmath26 , the inequality @xmath418 defines @xmath322 then the convexity of @xmath324 implies that if @xmath419 , then @xmath420 with equality if and only if @xmath421 lie in the same cone of @xmath342 . to prove this
, we may assume @xmath422 , so that @xmath423 then @xmath424 , so that @xmath425 it is equally easy to show that equality occurs exactly when @xmath421 lie in the same cone of @xmath342 . hence @xmath358 has the required properties , which completes the proof of part ( 2 ) of the theorem .
[ cor_simpl_ref_general ] if @xmath1 is a fan in @xmath50 with convex support of dimension @xmath5 , then there exists a simplicial refinement @xmath37 with the same 1-dimensional generators .
furthermore , if @xmath1 is quasi - projective , then we can assume that @xmath37 is also quasi - projective .
apply theorem [ thrm_simpl_ref_general ] with @xmath426 .
[ rem_simpl_ref_general ] this corollary guarantees the existence of simplicial refinements that introduce no new generators and preserve quasi - projectivity .
this result has other proofs , including fujino @xcite ( via the toric mori program ) and thompson @xcite ( via stellar subdivision ) .
we can now prove the non - simplicial case of our main theorem .
[ theorem_mori_nonsimp ] let @xmath1 be a non - simplicial quasi - projective fan in @xmath9 with convex support of dimension @xmath5 .
then the cone @xmath36 is defined by the primitive inequalities , i.e. , @xmath427 by corollary [ cor_simpl_ref_general ] , @xmath1 has a quasi - projective simplicial refinement @xmath37 satisfying @xmath335 . then observe that @xmath428 and that @xmath429 since @xmath37 is simplicial and quasi - projective .
hence @xmath430 we divide primitive collections @xmath431 for @xmath37 into two types : @xmath432 note that if @xmath34 , then @xmath433 when @xmath138 is a type a primitive collection for @xmath37 .
hence these can be omitted in , so that @xmath434 however , a type b primitive collection @xmath128 for @xmath37 is a primitive collection for @xmath1 .
this is easy to prove .
first , @xmath128 is not contained in any cone of @xmath1 by the definition of type b , and second , every proper subset of @xmath128 is contained in a cone of @xmath37 and hence lies in a cone of @xmath1 since @xmath37 refines @xmath1 .
it follows that @xmath435 is a primitive inequality for @xmath1 whenver @xmath436 is a type b primitive collection for @xmath37 .
hence shows that a subset of the primitive inequalities for @xmath1 define @xmath36 inside @xmath24 . using the inclusion , the theorem now follows immediately .
here is an example to illustrate theorem [ theorem_mori_nonsimp ] and its proof .
[ nonsimpl_ex ] figure [ figure : counterex_def ] shows the complete non - simplicial fan @xmath1 in @xmath276 with five minimal generators : @xmath437 and five maximal cones : @xmath438 ( 0,0),title="fig : " ] # 1#2#3#4#5 @font ( 945,1527)(1801,-2098 ) ( 2746,-691)(0,0)[lb ] ( 2476,-2041)(0,0)[lb ] ( 2656,-1006)(0,0)[lb ] ( 1801,-1006)(0,0)[lb ] ( 1936,-736)(0,0)[lb ] the primitive collections for this fan are : @xmath439 a first observation is that if we used batyrev s definition of primitive collection in this case , we would want every proper subset of @xmath293 and @xmath294 to generate a cone of @xmath1 .
this clearly is nt true , and in fact this example has _ no _ primitive collections if we use batyrev s definition .
this explains why definition [ pcdef ] is the correct definition in the non - simplicial case .
theorem [ theorem_mori_nonsimp ] states that @xmath440 is defined by the primitive inequalities coming from the primitive collections @xmath293 and @xmath294 .
however , the proof of the theorem shows that we need only one . to see why ,
consider the simplicial refinement @xmath37 of @xmath1 given by subdividing non - simplicial cone @xmath441 along @xmath442 .
this gives the fan pictured in example [ ex_simplicial ] .
the fan @xmath37 has the same generators @xmath443 as @xmath1 , and the primitive collections for @xmath37 are @xmath444 one easily checks that @xmath445 is of type a and @xmath294 is of type b and hence is a primitive collection for @xmath1 . by , @xmath36 is defined by @xmath294 , so that @xmath34 is convex if and only if @xmath446 it is interesting to note that the type a primitive collection @xmath447 also plays an important role .
the primitive relation of @xmath445 is @xmath448 now take @xmath34 . as noted in the proof of theorem [ theorem_mori_nonsimp ] , this type a primitive collection gives the _ equality _ @xmath449 which by the above primitive relation implies @xmath450 it is easy to see that this equality defines @xmath24 inside of @xmath415 . in other words
, @xmath451 lies in @xmath24 if and only if it satisfies coming from the type a primitive collection for @xmath37 .
if we turn our attention to the other primitive collection @xmath452 for @xmath1 , then one easily sees that @xmath453 is convex if and only if @xmath454 this follows by considering the other simplicial refinement of @xmath1 obtained by subdividing @xmath441 along @xmath455 .
example [ nonsimpl_ex ] has some interesting features : * every primitive collection for @xmath1 comes from a type b primitive collection for a simplicial refinement @xmath37 of @xmath1 satisfying @xmath161 .
* for each such refinement @xmath37 of @xmath1 , the type a primitive collections for @xmath37 define @xmath456 .
we will see below that these properties hold in general .
we begin with the following useful property of primitive collections .
[ thm_strict ] let @xmath1 be a fan in @xmath50 such that @xmath1 has convex support of dimension @xmath5 . if @xmath128 is a primitive collection for @xmath1 , then every proper subset @xmath457 of @xmath128 is linearly independent .
we use induction on @xmath458 . if @xmath459 there is nothing to show .
now assume that @xmath460 , @xmath461 , and that every @xmath12-element subset of @xmath457 is linearly independent .
we show that @xmath457 is linearly independent by contradiction .
hence suppose @xmath457 is linearly dependent .
then our induction hypothesis implies that the subspace @xmath462 has dimension @xmath12 .
define @xmath463 .
we omit the straightforward proof that @xmath464 is a fan in @xmath462 . now fix @xmath465 and let @xmath466 be the minimal cone of @xmath1 containing @xmath467 . notice that @xmath466 does not contain @xmath15 since @xmath128 is a primitive collection .
also let @xmath468 be the minimal cone of @xmath1 containing @xmath457 .
the cones @xmath469 and @xmath470 are in the fan @xmath464 and @xmath471 since @xmath15 is contained in @xmath472 but not in @xmath473 .
therefore , their intersection is at most @xmath474-dimensional . on the other hand ,
the intersection contains @xmath12 linearly independent vectors @xmath475 which is a contradiction .
[ cor_strict ] let @xmath1 be a fan in @xmath50 such that @xmath10 is convex support of dimension @xmath5 .
then every primitive collection for @xmath1 has at most @xmath372 elements .
this follows immediately from proposition [ thm_strict ] since any maximal proper subset @xmath476 is linearly independent and hence has at most @xmath5 elements .
therefore @xmath477 has at most @xmath372 elements .
proposition [ thm_strict ] and corollary [ cor_strict ] are trivial in the simplicial case .
let @xmath1 be a fan in @xmath50 with convex support of dimension @xmath5 , and let @xmath37 be a simplicial refinement with @xmath478 .
given @xmath143 , let @xmath479 .
the following convexity result will be useful .
[ lemma_subfab_convex ] let @xmath26 be a non - simplicial cone in @xmath1 and take an interior wall @xmath164 of @xmath328 with @xmath480 , @xmath481 .
then @xmath482 is convex .
since @xmath483 , @xmath164 divides @xmath26 into two convex subcones @xmath484 with @xmath485 .
we may assume @xmath486 .
given @xmath487 , it follows easily that the line segment @xmath488 lies in @xmath489 .
[ corollary_subfab_convex ] in the situation of lemma [ lemma_subfab_convex ] , let @xmath128 be the two element set @xmath490 thus @xmath128 consists of the generators of @xmath491 not lying in the wall @xmath492 . then @xmath128 is a primitive collection for @xmath37 . first note that @xmath128 is contained in neither @xmath493 nor @xmath494 .
since @xmath128 is contained in the convex set @xmath482 , it follows that @xmath128 is contained in no cone of @xmath37 .
thus @xmath128 is a primitive collection for @xmath37 since has it only has two elements . as in the proof of theorem [ theorem_mori_nonsimp ]
, a primitive collection for @xmath37 has type a when it is contained in a cone of @xmath1 .
hence the primitive collection for @xmath37 constructed in corollary [ corollary_subfab_convex ] has type a. the idea is that these two - element primitive collections define @xmath24 inside @xmath415 .
let @xmath1 be a fan in @xmath50 with convex support of dimension @xmath5 and let @xmath37 be a simplicial refinement with @xmath495 .
then @xmath496 the inclusion @xmath497 is obvious since elements of @xmath24 are linear on cones of @xmath1 and a type a primitive collection is contained in such a cone . for the opposite inclusion , take @xmath451 such that @xmath498 for all two element type a primitive collections for @xmath37 .
for each @xmath499 , there is @xmath500 such that @xmath501 for @xmath104 .
it suffices to show that @xmath502 for cones @xmath503 that lie in the same cone @xmath26 of @xmath1 and intersect in a wall @xmath504 .
this is the situation of corollary [ corollary_subfab_convex ] , where @xmath505 and @xmath506 is a two element type a primitive collection for @xmath37 .
we label the elements of @xmath128 so that @xmath507 and @xmath508 . since @xmath489 is convex by lemma [ lemma_subfab_convex ] , it contains @xmath509 .
we may assume @xmath510 without loss of generality .
then @xmath511 since @xmath512 , it follows that @xmath502 .
this completes the proof . in the fan @xmath1 pictured in figure [ figure : counterex_def ] in example [ nonsimpl_ex ] , we saw that every primitive collection for @xmath1 came from a primitive collection for a simplicial subdivision of @xmath1 .
in general , if @xmath37 is a simplicial subdivision of @xmath1 with @xmath161 , we say that a primitive collection @xmath128 for @xmath1 is _ supported _ on @xmath37 if @xmath128 is also a primitive collection for @xmath37 .
we now prove that all primitive collections for @xmath1 are supported on such simplicial subdivisions .
here is the precise result .
[ prim_lemma ] let @xmath1 be a fan in @xmath50 with convex support of dimension @xmath5 and let @xmath128 be a primitive collection for @xmath1 .
then there exists a simplicial refinement @xmath37 with @xmath161 such that @xmath128 is a primitive collection for @xmath37 .
furthermore , if @xmath1 is quasi - projective , then @xmath37 can be chosen to be quasi - projective . by proposition [ thm_strict ] ,
every proper subset @xmath128 is linearly independent . in particular ,
if @xmath143 , then @xmath330 is a proper subset of @xmath128 ( since @xmath128 is a primitive collection ) and hence is linearly independent .
thus we can apply theorem [ thrm_simpl_ref_general ] to obtain a simplicial refinement @xmath37 such that @xmath330 generates a cone of @xmath37 for all @xmath143 .
the theorem also allows us to assume @xmath37 is quasi - projective whenever @xmath1 is .
we claim that @xmath128 is a primitive collection for @xmath37
. first note that if @xmath128 were contained in a cone of @xmath37 , then it would be contained in a cone of @xmath1 , which we know to be false .
now let @xmath457 be a proper subset of @xmath128 .
then @xmath457 is contained in a cone @xmath513 , so that @xmath514 .
since @xmath330 generates a cone of @xmath37 , it follows that @xmath457 is contained in a cone of @xmath37 .
hence @xmath128 is a primitive collection for @xmath37 . when @xmath1 is non - simiplicial , it may be impossible to find a _
single _ simplicial refinement @xmath37 such that _ every _ primitive collection for @xmath1 is also primitive for @xmath37 . in figure
[ figure : counterex_def ] from example [ nonsimpl_ex ] , we see two primitive collections @xmath515 and @xmath516 , but there is no simplicial refinement @xmath37 of @xmath1 with @xmath161 that supports both @xmath293 and @xmath294 .
in this section we explore an open question about primitive collections . in @xcite ,
casagrande raises the question of whether @xmath36 is defined by primitive inequalities when @xmath1 is not quasi - projective .
here is a classic example .
[ ex_simplicial_nonqp ] the following example of a non - projective smooth complete fan is taken from fulton @xcite .
consider the fan @xmath1 in @xmath276 with seven minimal generators : @xmath517 the cones of @xmath1 are obtained by projecting from the origin through the triangulated polytope shown in figure [ figure : ex_simplicial_nonqp ] .
the fan @xmath1 has 15 walls and 10 maximal cones .
( 0,0 ) # 1#2#3#4#5 @font ( 2519,2700)(1149,-2543 ) ( 3668,-1463)(0,0)[lb ] ( 2888,-2498)(0,0)[lb ] ( 3203,-892)(0,0)[lb ] ( 1149,-1703)(0,0)[lb ] ( 2408 , 37)(0,0)[lb ] ( 1651,-938)(0,0)[lb ] ( 2520,-1432)(0,0)[lb ] the seven primitive collections for @xmath1 and their associated primitive relations are : @xmath518 by , a convex function @xmath48 satisfies the primitive inequalities : @xmath519 notice that adding up the third , fourth and sixth inequalities yields an equality , hence we have 3 equalities : @xmath520 to see what this says about the nef cone @xmath38 , note that @xmath521 assume @xmath522 .
then the three equalities give @xmath523 .
define @xmath524 and @xmath525 .
then inequalities imply @xmath526 and @xmath527 .
it follows that @xmath38 is contained in the 2-dimensional cone pictured in figure [ figure : ex_simpl_nqp_nef ] .
( 0,0 ) # 1#2#3#4#5 @font ( 2163,2010)(1789,-1873 ) ( 3691,-1658)(0,0)[lb ] ( 1921 , 29)(0,0)[lb ] ( 3541,-458)(0,0)[lb ] since @xmath56 has dimension 4 and @xmath38 has dimension at most two , we see that @xmath0 is non - projective since the nef cone does not have maximal dimension .
it is also easy to see that the cone in figure [ figure : ex_simpl_nqp_nef ] actually equals the nef cone @xmath38just show that the generators of this cone are nef .
for example , when @xmath528 , note that @xmath1 is a refinement of the complete fan @xmath529 with 1-dimensional generators @xmath530 .
the toric variety of @xmath529 is @xmath531 , and the class corresponding to @xmath532 is the pullback of an ample divisor on @xmath531 , hence nef on @xmath0 . for @xmath533 ,
one proceeds similarly by noting that @xmath1 is a refinement of the projective non - simplical fan @xmath534 with 1-dimensional generators @xmath535 .
other more substantial examples can be found in chapter 7 of scaramuzza s thesis @xcite . based on this ,
we make the following conjecture , which we credit to casagrande .
[ casagrande ] let @xmath0 be a simplicial toric variety coming from the fan @xmath1 in @xmath50 .
if @xmath10 is convex of dimension @xmath5 , then @xmath51 besides the evidence provided by examples , we also have the theoretical result of casagrande ( * ? ? ?
5.6 ) , which states that if a smooth complete non - projective toric variety @xmath0 has a toric blow - up @xmath536 with @xmath537 projective , then conjecture [ casagrande ] holds for @xmath0 .
we stated conjecture [ casagrande ] for the simplicial case because of the following result .
[ if_conjecture ] assume that conjecture [ casagrande ] is true .
let @xmath0 be a non - simplicial toric variety of a fan @xmath1 in @xmath50 such that @xmath10 is convex of dimension @xmath5 .
then : @xmath51 furthemore , every primitive collection for @xmath1 is supported on a simplicial refinement @xmath37 of @xmath1 satisfying @xmath335 . the first part of the proposition follows since the proof of theorem [ theorem_mori_nonsimp ] ( the non - simplicial case of our main theorem ) requires the existence of a simplicial refinement @xmath37 of @xmath1 with the following properties : * @xmath161 . *
@xmath538 is described using primitive inequalities .
if we assume conjecture [ casagrande ] , then the second bullet is automatically true , which means that corollary [ cor_simpl_ref_general ] gives the needed simplicial subdivision of @xmath1 . for the final assertion ,
observe that the proof of proposition [ prim_lemma ] applies without change since the first part of theorem [ thrm_simpl_ref_general ] gives the required simplicial refinement of @xmath1 without needing to assume quasi - projective .
one way to think about conjecture [ casagrande ] is that once this conjecture is proved , the results of this paper would apply to any fan in @xmath50 whose support is convex of dimension @xmath5there would be no requirement that @xmath1 be quasi - projective .
however , the proofs of the simplicial case given in theorem [ theorem_mori_primitive ] make essential use of extremal rays , which exist only in the quasi - projective case .
the result of casagrande ( * ? ? ?
5.6 ) mentioned above is a good first step , but it is likely that some significantly new ideas will be needed to prove conjecture [ casagrande ] in general .
this paper is based on the second author s phd thesis at the university of massachusetts @xcite , written under the direction of eduardo cattani and the first author .
both authors are very grateful to eduardo cattani for many stimulating discussions .
the first author would also like to thank anna scaramuzza for asking some interesting questions about primitive collections .
we would also like to thank cinzia casagrande for bringing @xcite to our attention and for giving a counterexample to a conjecture in the original version of the paper .
d. cox , _ recent developments in toric geometry _
, in _ algebraic geometry santa cruz 1995 _ ( j. kollr , r.lazarsfeld and d. morrison , eds . ) , proceedings of symposia in pure mathematics * 62 * , amer .
soc . , providence , ri , 1997 , 390435 | this paper studies batyrev s notion of primitive collection .
we use primitive collections to characterize the nef cone of a quasi - projective toric variety whose fan has convex support , a result stated without proof by batyrev in the smooth projective case .
when the fan is non - simplicial , we modify the definition of primitive collection and explain how our definition relates to primitive collections of simplicial subdivisons .
the paper ends with an open problem . |
turbulence , hailed as one of the last major unsolved problems of classical physics , has been the subject of numerous publications as researchers seek to understand the underlying physics , structures , and mechanisms inherent to the flow @xcite . the standard test problem for wall - bounded studies
historically has been turbulent channel flow because of its simple geometry and computational efficiency .
even though much insight has been achieved through the study of turbulent channel flow , it remains an academic problem because of its infinite ( computationally periodic ) spanwise direction .
the next simplest geometry is turbulent pipe flow , which is of interest because of its real - world applications and its slightly different dynamics .
the three major differences between turbulent pipe and channel flow are that pipe flow displays a log layer , but overshoots the theoretical profile until a much higher reynolds numbers ( @xmath1 for a channel versus @xmath2 for a pipe @xcite ) , has an observed higher critical reynolds number , and is linearly stable to an infinitesimal disturbance @xcite .
unfortunately , few direct numerical simulations of turbulent pipe flow have been carried out because of the complexity in handling the numerical radial singularity at the origin .
although the singularity itself is avoidable , its presence causes standard high - order spectral methods to fail to converge exponentially . as a result
, only a handful of algorithms found in the literature for turbulent pipe flow ; the first reported use for each algorithm is listed in table [ radialdiscr ] .
these algorithms typically use a low - order expansion in the radial direction . only shan et al .
@xcite using concentric chebyshev domains ( `` piecemeal '' ) and loulou et al .
@xcite using basis spline ( b - spline ) polynomials provide a higher - order examination of turbulent pipe flow . using a spectral element method
, this study provides not only a high - order examination but the first exponentially convergent investigation of turbulent pipe flow through direct numerical simulation ( dns ) .
[ radialdiscr ] with this dns result , one of the studies that can be performed with the full flow field and time history it provides is an analysis based on an orthogonal decomposition method . in such methods , the flow is expanded in terms of a natural or preferred turbulent basis .
one method used frequently in the field of turbulence is karhunen - love ( kl ) decomposition , which uses a two - point spatial correlation tensor to generate the eigenfunctions of the flow .
this is sometimes referred to as proper orthogonal decomposition , empirical orthogonal function , or empirical eigenfunction analysis .
work in this area was pioneered by lumley , who was the first to use the kl method in homogeneous turbulence @xcite .
this method was later applied to turbulent channel flow in a series of papers by ball , sirovich , and keefe @xcite and sirovich , ball , and handler @xcite , who discovered plane waves and propagating structures that play an essential role in the production of turbulence through bursting or sweeping events . to study the interactions of the propagating structures ,
researchers have examined minimal expansions of a turbulent flow @xcite .
these efforts have led to recent work by webber et al .
@xcite , who examined the energy dynamics between kl modes and discovered the energy transfer path from the applied pressure gradient to the flow through triad interactions of kl modes .
this present study uses a spectral element navier - stokes solver to generate a globally high - order turbulent pipe flow data set .
the karhunen - love method is used to examine the turbulent flow structures and dynamics of turbulent pipe flow .
the direct numerical simulation of the three - dimensional time dependent navier - stokes equations is a computationally intensive task .
by fully resolving the necessary time and spatial scales of turbulent flow , however , no subgrid dissipation model is needed , and thus a turbulent flow is calculated directly from the equations .
dns has one main advantage over experiments , in that the whole flow field and time history are known , enabling analyses such as the karhunen - love decomposition .
because of the long time integration and the grid resolution necessary for dns , a high - order ( typically spectral ) method is often used to keep numerical round - off and dissipation error small .
spectral methods and spectral elements use trial functions that are infinitely and analytically differentiable to span the element .
this approach decreases the global error exponentially with respect to resolution , in contrast to an algebraic decrease with standard methods such as finite difference or finite element methods @xcite .
this study uses a spectral element navier - stokes solver that has been developed over the past 20 years @xcite to solve the navier - stokes equations : @xmath3 in the above equations , @xmath4 is the velocity vector corresponding to the radial , azimuthal , and streamwise direction respectively ; re@xmath5 is the reynolds number ; @xmath6 is the radius of the pipe ; @xmath7 is the kinematic viscosity ; and @xmath8 is the shear velocity based upon the wall shear stress @xmath9 and density @xmath10 .
this solver employs a geometrically flexible yet exponentially convergent spectral element discretization in space , in this case allowing the geometry to be fitted to a cylinder .
the domain is subdivided into elements , each containing high - order ( usually 1113 ) legendre lagrangian interpolants .
the resulting data localisation allows for minimal communication between elements , and therefore efficient parallelization .
time discretization is done with third - order operator splitting methods , and the remaining tensor - product polynomial bases are solved by using conjugate gradient iteration with scalable jacobi and hybrid schwarz / multigrid preconditioning @xcite .
spectral elements are effective in cylindrical geometries @xcite and their use elegantly avoids the radial singularity at the origin . the mesh is structured as a box near the origin and transitions to a circle near the pipe walls ( figure [ grid ] ) , maintaining a globally high - order method at both the wall and the origin .
in addition to avoiding the numerical error associated with the singularity , the method also avoids the time - step restriction due to the smaller element width at the origin of a polar - cylindrical coordinate system , which could lead to potential violations of the courant - friedrichs - levy stability criteria .
= 2 in each slice , as shown in figure [ grid ] , has 64 elements , and there are 40 slices stacked in the streamwise direction , adding up to a length of 20r .
each element has 12th - order legendre polynomials in each direction for re@xmath0 .
this discretization results in 4.4 million degrees of freedom . near the wall , the grid spacing normalised by wall units ( @xmath11 ) denoted by the superscript @xmath12 is @xmath13 and @xmath14 , where @xmath15 is
the radius and @xmath16 is the azimuthal angle . near the centre of the pipe
, the spacing in cartesian coordinates is @xmath17 .
the streamwise grid spacing is a constant @xmath18 throughout the domain where @xmath19 is the streamwise coordinate .
benchmarking was performed at re@xmath21 with the experiments and dns of eggels et al . @xcite and the dns of fukagata and kasagi @xcite . for this higher reynolds number flow , 14th - order polynomials
were used , giving grid spacings near the wall of @xmath22 and * @xmath23 , @xmath24 * in the centre , and @xmath25 .
eggels et al .
and fukagata and kasagi used a spectral fourier discretization in the azimuthal and axial directions and then a 2nd - order finite difference discretization in the radial direction .
also , both groups used a domain length of 10r and grid sizes in their dns studies of @xmath26 for @xmath15 , @xmath16 , and @xmath19 directions , respectively .
the turbulent flow was tripped with a solenoidal disturbance , and the simulation was run until transition was complete .
this was estimated by examining the mean flow and root - mean - squared statistics over sequential periods of 100 @xmath27 until a statistically steady state was achieved , marked by a change of less than 1 percent between the current and previous period . in this case , it took @xmath28 to arrive at a statistically steady state .
convergence was checked by examining the total fluctuating energy after 1000 time steps for increasing polynomial orders 8 , 10 , 12 , and 14 with a fixed number of elements ( 2560 ) .
the absolute value of the difference of the total fluctuating energy for each simulation and the value obtained using 16th order polynomials was plotted versus the total number of degrees of freedom in figure [ convergence ] .
the exponential decay of the error with increasing degrees of freedom shows that our spectral element algorithm achieves geometric or exponential convergence .
=3 in the mean flow profiles in figure [ mean_180 ] correspond well with the hot wire anemometry ( hwa ) results of eggels et al .
@xcite , but as seen in the root - mean - squared ( rms ) statistics shown in figure [ rms_180 ] , the spectral element calculation shows a lower peak @xmath29 and higher peak @xmath30 and @xmath31 compared to the hwa , particle image velocimetry ( piv ) , and laser doppler anemometry ( lda ) results .
these results are in contrast to those with channel flow , as reported by gullbrand @xcite , where the 2nd order finite difference methods undershoot the spectral method wall - normal velocity rms . when compared to the experimental results of eggels et al .
@xcite , the spectral element results are in better agreement than the 2nd - order finite difference results .
we note that eggels et al .
report that their ( piv ) system had difficulties capturing @xmath32 near the wall and near the centerline of the pipe due to reflection , which could explain the deviation of all of the dns results in that area . .
the theoretical line is the law of the wall @xmath33 and the log layer @xmath34 .
deviations from the log layer are expected in turbulent pipe flow until much higher reynolds numbers.,width=4 ] a second major difference our spectral element method and the previous pipe dns using 2nd - order finite difference is the domain size . with the spectral element method , which results in a global high - order convergence , a domain size of 10r yielded unphysical results in the flow , as seen in the bulge in the azimuthal @xmath30 in figure [ l10 ] at @xmath35 and @xmath36 .
this unphysical result arose even with a more refined grid , and is therefore not a function of under - resolution .
however , when the domain of the spectral element method was extended to 20r , this problem disappeared .
we surmise that the 2nd - order finite difference method dissipated some large - scale structures after 10r that the higher - order spectral element case appropriately resolves .
this undissipated structure , because of the periodic boundary conditions , then re - enters the inlet and causes the unphysical bulges in the azimuthal rms profile .
this result is also supported by the work of jimnez @xcite in turbulent channel flow ( re@xmath37=180 ) that shows large - scale structures that extend well past the domain size of @xmath38 .
this benchmark confirms that the spectral element algorithm , at the given grid resolution and domain size , will generate the appropriate turbulent flow field and time history to perform the kl decomposition .
= 4 in for completeness , the karhunen - love decomposition method is briefly described here , but for more detail , see @xcite .
the karhunen - love method is the solution of the two - point velocity correlation kernel equation defined by @xmath39 where @xmath40 is the fluctuating component of the velocity and where the mean @xmath41 is determined by averaging over both homogeneous planes and time .
the angle brackets represent an average over many time steps , on the order of @xmath42 , to sample the the entire attractor .
the @xmath43 denotes the outer product , establishing the kernel as the velocity two - point correlation between every spatial point @xmath44 and @xmath45 . for turbulent pipe flow , with two homogeneous directions providing translational invariance in the @xmath16 ( azimuthal ) and @xmath19 ( streamwise ) direction , eq .
( [ two_point ] ) becomes @xmath46 thus , given the kernel in eq .
( [ two_point_pipe ] ) , the eigenfunctions have the form @xmath47 where @xmath48 is the azimuthal wavenumber and @xmath49 the streamwise wavenumber .
the determination of @xmath50 is then given by @xmath51 where the @xmath52 denotes the complex conjugate , @xmath53 is the eigenvalue , and @xmath54 is the fourier transform of the fluctuating velocity in the azimuthal and axial direction . in the present problem ,
2,100 snapshots of the flow field were taken , corresponding to one snapshot every eight viscous time steps ( @xmath55 ) .
the results of each snapshot were projected to an evenly spaced grid with @xmath56 points in @xmath57 and @xmath19 , respectively .
the fourier transform of the data was then taken and the kernel assembled .
this kernel was averaged over every snapshot to generate the final kernel to be decomposed .
since the dimension of k is 303 ( given by three velocity components on 101 radial grid points ) there are 303 eigenfunctions and eigenvalues for each fourier wavenumber pair @xmath58 .
the eigenfunctions are ranked in descending order with the quantum number @xmath59 to specify the particular eigenfunction associated with the corresponding eigenvalue .
thus , it requires a triplet @xmath60 to specify a given eigenfunction . the eigenfunctions @xmath61 are complex vector functions and are normalised so that the inner product of the full eigenfunction @xmath62 is of unit length , namely @xmath63 , where @xmath64 is the kronecker delta .
the eigenvalues physically represent the average energy of the flow captured by the eigenfunction @xmath62 , @xmath65 we note that the eigenfunctions , as an orthogonal expansion of the flow field , retain the properties of the flow field , such as incompressibility and boundary conditions of no slip at the wall .
the pipe flow is statistically invariant under azimuthal reflection , @xmath66 and taking advantage of it reduces the total number of calculations as well as the memory and storage requirements .
we note that , because of its geometry , turbulent channel flow has two more symmetries a vertical reflection , and a x - axis rotation that are not present in the pipe , since a negative radius is equivalent to a 180 degree rotation in the azimuthal direction .
a major consequence of this symmetry is that the resulting eigenfunctions are also symmetric , and the modes with azimuthal wave number @xmath48 will be the azimuthal reflection of the modes with wave number @xmath67 , @xmath68 thus reducing the total computational memory needed for this calculation .
the kl method provides an orthogonal set of eigenfunctions that span the flow field . as such
, the method allows the flow field to be represented as an expansion in that basis , @xmath69 with @xmath70 since the fourier modes are orthogonal to each other , equation [ coeff ] becomes @xmath71 with @xmath72 being the fourier transform of @xmath73 in the azimuthal and streamwise direction with wavenumbers @xmath48 and @xmath49 , respectively .
the time history of the eigenfunctions can be used to examine their interactions , such as the energy interaction examined by webber et al .
@xcite and bursting events by sirovich et al .
this section presents the results of our analysis of turbulent pipe flow based on kl decomposition .
dns was performed for @xmath74 .
the reynolds number based on the centreline velocity is @xmath75 and based on the mean velocity is @xmath76 .
this range is above the observed critical reynolds number for pipe flow and exhibits self - sustaining turbulence , as seen from the fluctuating time history of the mean velocity shown in figure [ mean_time ] .
the velocity profile , seen in figure [ mean_profile ] , shows the mean velocity with respect to wall units away from the wall ( @xmath77 ) .
the profile fits the law of the wall but fails to conform to the log law , and as mentioned in section 1 , the log law is not expected for turbulent pipe flow until much higher reynolds number .
= 4 in = 4 in the rms velocity fluctuation profiles and the reynolds stress profile is shown in figures [ rms ] and [ reynolds ] .
the streamwise fluctuations peak near @xmath78 .
the azimuthal and axial velocities show a weaker peak near @xmath79 and @xmath35 , respectively , and then remain fairly flat throughout the pipe .
the reynolds stress @xmath80 has a maximum of 0.68 at @xmath81 . [
cols="^,^ " , ] the energy spectra of the propagating mode subclasses are shown in figure [ distribe ] .
this shows that the tail wavenumber end of the inertial range is populated primarily by the lift modes and the low end of the spectra is populated predominantly by the wall modes .
the physical meaning of this is that the energy starts at the wall with large scale structures and is lifted away from the wall into the outer region in small scale structures , represented by the lifting modes .
the effect of higher radial quantum number @xmath82 is more zero crossings of the velocities in the radial direction .
the effect on the location of coherent vorticity and the number of vortex cores scales with the radial quantum number @xmath59 , making the subclasses invariant with @xmath59 . the wall and lift modes retain their characteristics , in that even with more vortices , they remain close to the wall for the wall mode and close to the centreline for the lift modes , shown in figures [ 623non ] through [ 265non ] .
cc & + & +
although the karhunen - love method yields a preferred or natural basis for turbulence , one must be careful in the conclusions drawn from the results , as any structure or feature can be reconstructed from any given orthogonal basis .
the strength of this method is that it is the most optimal basis , and by focusing on the most energetic modes a low order dynamical representation of the flow is realised .
this low order dynamical system reveals three important results , and paints a picture of the energy dynamics in turbulent pipe flow .
the first result is the constant phase speed of the propagating modes .
this was also found in studies of turbulent channel flow , as the structures represented by the basis advect with a constant group velocity , the same average velocity of burst events @xcite .
the normal speed locus of the propagating waves is shown in figure [ locus ] for the propagating modes found in the top 50 most energetic modes .
for this , the phase speed @xmath83 is plotted in the direction @xmath84 .
the locus is nearly circular , which is evidence that these structures propagate as a wave packet or envelope that travels with speed of * @xmath85 * , the point at which the circle intersects the x - axis in figure [ locus ] , which corresponds to the mean velocity in the buffer region at @xmath86 . .
the circle is a least - squares fit to the data and represents that the wave packets are acting together as a group with speed of 8.41 @xmath87 , the point where the circle intersects the y - axis.,width=4 ] the second interesting result is that none of the modes with azimuthal wave number @xmath88 exhibit any travelling waves near the wall , and are without exception a streamwise or inclined streamwise vortex in the outer region .
the reason is that the @xmath88 mode does not allow for a near - wall travelling wave as found by kerswell @xcite , and as such , the basis expansions for these modes have only near - centreline structures .
the third result of note is that the different kl structures , as observed from our results , can be seen qualitatively as an expansion that represents the horseshoe ( hairpin ) vortex representation of turbulent structures .
this horseshoe structure is supported by a large number of researchers in the turbulence community as the self - sustaining mechanism for turbulence @xcite . in the representative horseshoe structure ( see , for example ,
the figure by theodorsen @xcite ) , the structures found in the kl can be seen .
the wall modes represent the leg structure and its perturbation near the wall .
the lift modes represent the structure lifting off the wall .
the asymmetric modes represent the secondary and tertiary horseshoes that are formed .
the turn modes represent the spanwise head of the horseshoe . as mentioned before , since the kl decomposition forms a basis , any flow can be recreated from these eigenmodes , so this result
is reported as an interesting qualitative result , and as an understanding that the propagating modes are the structures of interest in turbulence . the final dynamical picture of the kl modes in summarising the work done by kerswell @xcite , webber et al .
@xcite , sirovich et al .
@xcite , and this current work is as follows .
webber et al . found that the energy enters the flow from the pressure gradient to the shear modes .
the shear modes , interacting with the roll modes present in fully developed turbulence transfers the energy from the shear modes to the roll modes . shown in this study , and theoretically by kerswell
, the roll modes decay much more slowly than the propagating modes . because of this slow decay
, these streamwise rolls provide an energy storage role in the turbulence , similar to a capacitance .
this allows enough time for the roll to interact with the propagating wave packet shown by sirovich et al . because this interaction requires a finite time to occur , the energy storage and slow decaying nature of the streamwise rolls play an integral role in the self - sustaining nature of turbulence .
it is also surmised that the travelling modes found by kerswell and other researchers are represented in the kl formulation as the most unstable wave packet of kl modes .
following again the work of webber et al . and applying the classifications found in this study , the majority of the energy from the roll modes are transferred to the wall modes .
the wall modes then interact with themselves through an asymmetric mode catalyst , and with the lift modes through a ring mode catalyst .
physically , this gives a picture of wall turbulence energy being transferred near the wall and then lifted up to the outer layer .
this interaction between the wall and the lift modes is what populates the inertial range shown in figure [ distribe ] .
the energy is ultimately dissipated by viscosity , faster for the higher wavenumber modes .
we have presented the use of the karhunen - love expansion method with the results of a globally high - order direct numerical simulation of turbulent pipe flow .
the results reveal the structure of the turbulent pipe flow as propagating ( 80.58% total energy ) and non propagating modes ( 19.42% total energy ) .
the propagating modes are characterised by a constant phase speed and have four distinct classes : wall , lift , asymmetric , and ring modes .
these propagating modes form a travelling wave envelope , forming a circular , normal - speed locus , with advection speed of @xmath85 corresponding to the mean velocity in the buffer region near @xmath86 .
the non propagating modes have two subclasses : streamwise roll and shear modes .
these represent the energy storage and mean fluctuations of the turbulent flow , respectively .
the energy is transferred from the streamwise rolls to the wall modes first , and then later to the lift modes , physically representing the energy transfer from the wall to the outer region .
this eigenfunction expansion , using both their structure and their time - dependent coefficients , provides a framework for understanding the dynamics of turbulent pipe flow , and will provide a basis for further analysis and comparison leading to understanding the mechanism of drag reduction by spanwise wall oscillation @xcite and the mechanism of relaminarization @xcite .
we gratefully acknowledge the use of the virginia tech terascale computing facility and the teragrid san diego supercomputing facility .
eggels , j.g.m . , unger , f. , weiss , w.h .
, westerweel , j. , adrian , r.j . , friedrich , r. , and nieuwstadt , f.t.m . , 1994 .
fully developed turbulent pipe flow : a comparison between direct numerical simulation and experiment .
_ j. fluid mech . ,
_ * 268 * , 175209 .
fischer , p.f .
, ho , l.w . , karniadakis , g.e , ronouist , e.m . , and patera , a.t . , 1988 .
recent advances in parallel spectral element simulation of unsteady incompressible flows .
_ comput . & struct . , _ * 30 * , 217231 .
tufo , h.m . , and fischer , p.f . , 1999 .
terascale spectral element algorithms and implementations , _ proceedings of the acm / ieee sc99 conference on high performance networking and computing , _ ieee computer soc . , gordon bell prize paper . | the results of an analysis of turbulent pipe flow based on a karhunen - love decomposition are presented .
the turbulent flow is generated by a direct numerical simulation of the navier - stokes equations using a spectral element algorithm at a reynolds number re@xmath0 .
this simulation yields a set of basis functions that captures 90% of the energy after 2,453 modes .
the eigenfunctions are categorised into two classes and six subclasses based on their wavenumber and coherent vorticity structure . of the total energy ,
81% is in the propagating class , characterised by constant phase speeds ; the remaining energy is found in the non propagating subclasses , the shear and roll modes .
the four subclasses of the propagating modes are the wall , lift , asymmetric , and ring modes .
the wall modes display coherent vorticity structures near the wall , the lift modes display coherent vorticity structures that lift away from the wall , the asymmetric modes break the symmetry about the axis , and the ring modes display rings of coherent vorticity .
together , the propagating modes form a wave packet , as found from a circular normal speed locus .
the energy transfer mechanism in the flow is a four step process .
the process begins with energy being transferred from mean flow to the shear modes , then to the roll modes .
energy is then transfer ed from the roll modes to the wall modes , and then eventually to the lift modes .
the ring and asymmetric modes act as catalysts that aid in this four step energy transfer .
physically , this mechanism shows how the energy in the flow starts at the wall and then propagates into the outer layer . corresponding author .
e - mail : [email protected] ] direct numerical simulation , karhunen - love decomposition , turbulence , pipe flow , mechanism |
highly energetic partons propagating through the hot and dense qgp rapidly lose their energy and momentum as the energy is deposited in the medium .
measurements of two- and three - particle correlations in heavy - ion collisions show a complete suppression of the away - side jet , whereas for lower @xmath0 a double peak structure is observed in the two - particle correlation function @xcite . one possible and promising origin of these structures
is assumed to be the interaction of fast partons with the soft matter which generates collective motion of the medium in form of mach cones .
@xcite . for this purpose
we investigate the propagation and formation of mach cones in the microscopic transport model bamps ( boltzmann approach of multiparton scatterings ) @xcite in the limit of vanishing mass and very small shear viscosity over entropy density ratio @xmath1 of the matter .
two different scenarios for the jet are used .
in addition , by adjusting @xmath1 , the influence of the viscosity on the profile of the mach cone and the corresponding two - particle correlation is explored for the first time .
the results presented are based on a recent publication @xcite .
shock waves are phenomena which have their origin in the collective motion of matter . in a simplified one - dimensional setup shock waves
have already been studied within the framework of bamps for the perfect fluid limit @xcite .
furthermore bamps calculations have demonstrated that the shock profile is smeared out when viscosity is large .
it was also found that a clear observation of the shock within the short time available in hic requires a small viscosity . in the following
we study the evolution of `` mach cone''-like structures with different scenarios of the jet - medium interaction by using the parton cascade bamps .
we focus on investigation of mach cone evolution in absence of any other effects - i.e. we neglect such effects as initial fluctuations or expansion , which are however relevant in hic .
we use a static box with @xmath2 mev and binary collisions with an isotropic cross section .
furthermore , we keep the mean free path @xmath3 of the medium particles constant in all spatial cells by adjusting the cross section according to @xmath4 , where @xmath5 is the particle density . the related shear viscosity for isotropic binary collisions is given by @xmath6 @xcite .
for different viscosities extracted from calculations shown in fig .
[ fig : edensity_viscdifsourceterms ] .
the results are shown in the for the jet ( a ) and ped ( b ) scenario for @xmath7 gev / fm . ]
the mach cones studied here are induced by two different sources .
the first of them we refer to as the pure energy deposition scenario ( ped ) @xcite .
this is simulated by a moving source depositing momentum end energy isotropically according to the thermal distribution @xmath8 .
the second source we refer to as jet .
this is simulated by a highly massless particle ( jet ) which has only momentum in @xmath9-direction , i.e. @xmath10 .
after each timestep the energy of the jet particle is reset to its initial value . for both scenarios
the sources are initialized at @xmath11 fm / c at the position @xmath12 fm and propagate in @xmath9-direction with @xmath13 , i.e. with the speed of light . in fig .
[ fig : edensity_viscdifsourceterms ] we show the mach cone structure for both ped scenario ( upper panel ) and jet scenario ( lower panel ) with @xmath14 , @xmath15 and @xmath16 from left to right , respectively .
we show a snapshot at @xmath17 . the energy deposition rate is fixed to @xmath7 gev / fm . in both scenarios ,
ped and jet , for @xmath14 ( left panel ) , we observe a conical structure , but with obvious differences .
the ped case with the isotropic energy deposition induces a spherical shock into back region ; this structure is missing in the jet scenario because of the high forward peaked momentum deposition .
another difference is that in the jet scenario a clearly visible head shock appears .
this in turn is missing in the ped scenario .
furthermore a ( anti)-diffusion wake is induced by the jet ( ped ) scenario .
adjusting the shear viscosity over entropy density ratio @xmath18 we observe a smearing out of the mach cone structure . for a sufficient high @xmath19 the conical structure in both scenarios disappears .
this is true for shock fronts as well as for the ( anti- ) diffusion wake .
the difference between the ped and the jet case is that as @xmath1 increases , in the ped scenario the resulting `` mach cone '' solution covers approximately the same spatial region regardless of a value of @xmath1 , while in the jet case the structure is concentrated more and more near the projectile as the viscosity increases . in fig .
[ fig : num_visc_tpc ] we show the two - particle correlations extracted from bamps calculations of the mach cones shown in fig .
[ fig : edensity_viscdifsourceterms ] . for the jet scenario ( a ) and sufficiently small @xmath14
we observe only a peak in direction of the jet . the typical double peak structure , which has been proposed as a possible signature of the mach cone in hic ,
can only be observed for the ped scenario ( b ) and small @xmath1 .
however , the ped scenario has no correspondence in heavy - ion physics . for the jet scenario , which is a simplified model of jets depositing energy and momentum , a double peak structure never appears .
this is due to the strong formation of a head shock and diffusion wake .
in summary , the evolution of mach cones induced by two different source terms , ped and jet , were investigated using a microscopic transport model . the development of mach cones is observed in case the viscosity of matter is small enough .
in addition , the effects of viscosity of the matter were shown by adjusting the shear viscosity over entropy density ratio @xmath1 from @xmath20 to @xmath16 .
a clear and unavoidable smearing of the profile depending on a finite ratio of shear viscosity to entropy density is clearly visible .
investigating the corresponding two - particle correlations we see that mach cones can not be connected to double peak structures by any realistic picture of jets in hic .
the authors are grateful to the center for the scientific computing ( csc ) at frankfurt for the computing resources .
this work was supported by the helmholtz international center for fair within the framework of the loewe program launched by the state of hesse . | we investigate in a microscopical transport model the evolution of conical structures originating from the supersonic projectile moving through the matter of ultrarelativistic particles . using different scenarios for the interaction between projectile and matter , and different transport properties of the matter , we study the formation and structure of mach cones .
furthermore , the two - particle correlations for different viscosities are extracted from the numerical calculations and we compare them to an analytical approximation .
in addition , by adjusting he cross section we investigate the influence of the viscosity to the structure of mach cones . |
the bose - einstein condensation of a mixture of different hyperfine states of the same element ( first realized by the jila group @xcite ) offers the possibility to transfer atoms from an internal state to another one in a macroscopic matter wave .
this feature has driven a rich body of experimental studies of phenomena such as the formation of spin domains , vortices and other nonlinear structures @xcite , internal josephson effect @xcite , formation of squeezed and entangled states @xcite , motion of spin impurities @xcite , persistent currents @xcite , effective gauge potentials @xcite and spin - orbit coupled systems @xcite , which has itself opened an avenue of new researches : observation of a superfluid hall effect @xcite ; of zitterbewegung @xcite ; of spin hall effect @xcite ; of tunable landau - zener transitions @xcite ; of a dicke - type phase transition @xcite ; of the softening of a roton - like dispersion relation @xcite ... in some of the above cited works , the change of internal state is only due to spin - dependent collisions , but in others the coupling is externally driven by a combination of radiofrequency and microwave fields @xcite or by using raman coupling lasers @xcite . in the present study we concentrate on an effective spin @xmath0 system in which two internal states are coherently coupled by an external potential . in such a system ,
the coupling explicitly breaks the @xmath1 symmetry originating from the irrelevance of global phase factors of each the two components : the relative phase is no longer free and only remains a @xmath2 symmetry for the global phase of the spinor . as a consequence
, the two - branched spectrum of the system has a single goldstone mode and the other branch is gaped .
the mean field dynamics of the system is described by two coupled gross - pitaevskii equations accounting for intra and inter - species collisions , for the external coupling field and also possibly for a spin - orbit term .
the ground state of the system and the associated possible phase transitions and the elementary excitations have been theoretically studied in refs .
@xcite and @xcite , as long as a rich variety of nonlinear structures ( refs . @xcite and @xcite ) .
the reason for the protean aspect of the theoretical approaches of the system lies in the fact that its dynamics is described by a nonintegrable set of coupled gross - pitaevskii equations which do not admit simple integrable equations as limiting cases . even in the simpler case of a spinor condensate in the absence of spin - orbit and raman coupling ,
the integrable limit is the so - called manakov system ( obtained when all the nonlinear interaction constants are equal ) which does not pertain to the well - studied ablowitz - kaup - newell - segur hierarchy and for which all the types of solutions are not yet fully classified ( see e.g. , refs .
the aim of the present work is to partially clarify the rich nonlinear behavior of the system by presenting a systematic study revealing how nonlinear effects modify the elementary excitations of the system .
the paper is organized as follows : the model , its ground state and linear excitations are described in sec .
[ model ] .
we then use a singular perturbation theory to describe in section [ sec : upperbranch ] how excitations in the upper branch of the dispersion relation are affected by nonlinear effects .
the method is exposed in subsections [ subeps1 ] , [ sec : order2 ] and [ order3 ] and the results are summarized and discussed in subsection [ final - upper ] . the technique used in sec .
[ sec : upperbranch ] can also be employed for describing the effects of nonlinearity on the lower branch of the spectrum . however , for this branch another approach can be used which is more appropriate in the long wave length limit .
this is explained in sec .
[ sec : lowerbranch ] and we show in subsections [ kdvregime ] and [ gardner_regime ] how to deal with this issue . the general doctrine is presented in subsection [ quartic ] where we also discuss the different regimes accessible in present days experiments .
our conclusions are summarized in sec .
[ conclusion ] and some technical aspects are detailed in appendices [ app1 ] and [ app2 ] .
we consider a one - dimensional system described by a two - component spinor order parameter @xmath3 ( where the superscript @xmath4 denotes the transposition ) obeying the following coupled gross - pitaevskii equations @xmath5
\alpha_2 \psi_\uparrow^*\psi_\downarrow & & \alpha_1 |\psi_\downarrow|^2 \\ \end{pmatrix } \psi\ ; , \ ] ] where @xmath6 is the single particle hamiltonian : @xmath7 @xmath8 and @xmath9 being pauli matrices .
this corresponds to a system with equal contribution of rashba and dresselhaus coupling , as realized in spin - orbit coupled condensates ( see , e.g. , ref .
@xcite ) . in eq .
, @xmath10 is the interspecies interaction coefficient , and for simplicity we have assumed equal intraspecies interaction : @xmath11 . in the following
we will consider the case of repulsive intraspecies interaction : @xmath12 .
it is convenient to re - parametrize the spinor wave - function @xcite : @xmath13 \sin\frac{\theta}2\,\operatorname{e}^{i\varphi/2 } \\ \end{pmatrix}\ ; .\ ] ] here @xmath14 denotes the total density of the condensate and @xmath15 has the meaning of the velocity potential of its in - phase motion ; the angle @xmath16 is the variable describing the relative density of the two components ( @xmath17 ) and the phase @xmath18 is the potential of their relative ( counter - phase ) motion .
accordingly , the densities of the components of the condensate are given by @xmath19 their velocities are defined as @xmath20 it will also appear convenient to define the following velocity fields @xmath21 equation , expressed in terms of the real fields @xmath22 , @xmath23 , @xmath24 and @xmath25 , reads [ mod5 ] @xmath26_x- \tfrac12(\rho\,\phi_x)_x \ ; , \\ - \phi_t=&- \tfrac12 \cot\theta\frac{(\rho\,\theta_x)_x}{\rho } + \tfrac12 \left(\frac{\rho_x^2}{2\rho^2}-\frac{\rho_{xx}}{\rho}\right ) + \nonumber \\ & \tfrac14[\phi_x^2+\theta_x^2+(\varphi_x+2k_0)^2]+\nonumber \\ & ( \alpha_1+\alpha_2)\rho+\omega\ , \frac{\cos\varphi}{\sin\theta } \ ; , \label{mod5b } \\
-\theta_t=&\tfrac12 \phi_x\theta_x + \frac{1}{2\rho } [ \rho(\varphi_x+2 k_0)\sin\theta]_x + \nonumber \\ & \omega\,\sin\varphi \ ; , \\
\varphi_t=&\frac{1}{2\sin\theta}\frac{(\rho\,\theta_x)_x}{\rho } -\tfrac12 \phi_x(\varphi_x+2k_0)+\nonumber \\ & ( \alpha_1-\alpha_2)\rho \cos\theta-\omega\,\cos\varphi\cot\theta \ ; , \label{mod5d } \end{aligned}\ ] ] where we have used units such that @xmath27 .
in all the following we will assume that the different parameters of the hamiltonian are fixed in such a way that the ground state of the system corresponds to a configuration in which both components are homogeneous ( @xmath28 and @xmath29 ) , in phase ( @xmath30 ) , stationary ( @xmath23 , @xmath31 , @xmath24 and @xmath32 are time independent ) with equal densities ( @xmath33 @xcite ) .
in this case , one obtains @xmath34 , where @xmath35 is the chemical potential . in this expression
we have used the notation @xmath36 .
it will also appear convenient to define @xmath37 and to introduce a rescaled density @xmath38 . in the absence of spin - orbit coupling ( @xmath39 ) this ground state is stable provided @xmath40 @xcite .
for a spin - orbit coupled system , this ground state is denoted as the `` single minimum '' or `` zero momentum '' or `` phase iii '' ground state .
it is the true ground state of the system in a region of parameters which is schematically depicted in fig .
[ fig.phases ] ( adapted from ref .
@xcite ) . and @xmath41 .
for each phase the inset represent a typical density pattern .
the boundary between phases iii and ii corresponds to @xmath42 . ] although the present work is devoted to the study of nonlinear effects in phase iii , we note that the methods we use also apply with unessential modifications in phase ii which is a spin polarized phase where the system condensates in a single plane wave state with non - zero momentum .
phase i ( the so - called striped phase ) which has a modulated ground state density deserves a special treatment .
a first insight in the dynamics of the system can be obtained by linearizing eqs . .
for simplifying the notations we introduce the column vector @xmath43 being the ground state value of @xmath44 .
we write @xmath45 where @xmath46 describes a small departure of the fields @xmath47 , @xmath22 , @xmath24 and @xmath25 from their ground state values . inserting this ansatz into one
obtains at first order in @xmath48 a system of the form @xmath49 where @xmath50 this equation being linear one can expand @xmath46 on a basis of plane waves of wave - vector @xmath51 and angular frequency @xmath52 .
this amounts to look for solutions of of the form @xmath53 + \operatorname{c.c.}$ ] , where `` @xmath54 '' stands for `` complex conjugate '' and @xmath55 is a constant column vector whose entries are possibly complex .
one then obtains a system of linear equations which reads @xmath56 where @xmath57 the system has non - trivial solutions only if the determinant of @xmath58 vanishes .
this fixes the dispersion relation of the elementary excitations , with two branches @xmath59 which are represented in fig .
[ fig.dispers ] .
they are solutions of @xmath60\\ & + \frac{k^2}{2 } \left[\frac{k^2}{2}+\omega+g_2 - 2 k_0 ^ 2\right]\times \\ & \left[\left(\frac{k^2}{2}+\omega\right ) \left(\frac{k^2}{2}+g_1\right)-k^2 k_0 ^ 2\right]\ ; . \end{split}\ ] ] ( upper branch ) and @xmath61 ( lower branch ) , solutions of eq . .
the figure is drawn in the case @xmath62 , @xmath63 , @xmath64 and @xmath65 .
the ( red ) dashed line represent the long wavelength expansion .
the thin ( blue ) lines represent the spectrum of the single particle hamiltonian @xmath6 [ cf .
they are obtained by taking @xmath66 in eq . : in this case one obtains @xmath67^{1/2}$ ] . ]
the upper branch @xmath68 is gaped , with a dispersion relation of the form @xmath69 the lower branch @xmath70 is not gaped : it accounts for the goldstone mode corresponding to the spontaneous breaking of the global @xmath2 symmetry of the system .
one sees in fig .
[ fig.dispers ] that the upper branch is not qualitatively affected by interaction effects , contrarily to the lower branch whose long wavelength dispersion relation would be quadratic in the absence of interaction and becomes linear in its presence .
the lower branch admits , for the positive @xmath51 portion of the spectrum , the following expansion ( corresponding to linear waves propagating in the positive-@xmath71 direction ) : @xmath72 where @xmath73 is the sound velocity , and the parameter @xmath74 verifies @xmath75\;. \end{split}\ ] ]
we study in the present section how nonlinear effects modify the structure of an excitation propagating in the upper branch of the spectrum . for instance
, one can anticipate that nonlinear terms cause some modulations or anharmonicities of this wave , and make it interact with the other branch of the spectrum . instead of the simple linear analysis of section [ model ] ( eqs . , and following ) ,
we perform here a singular perturbative expansion by writing the term @xmath46 in eq . under the form ( see , e.g. , @xcite ) : @xmath76 in this expansion @xmath77 is a small parameter .
@xmath78 are multiscale coordinates aiming at describing the slow spatial and temporal modulations of a wave packet of finite amplitude .
@xmath79 in is the same as in and we make the following ansatz for the form of the @xmath80 term : @xmath81 where @xmath82 this means that we assume that the @xmath83 solution of consists in a slowly varying contribution ( @xmath84 , @xcite ) plus an oscillating term with a smoothly varying amplitude @xmath85 .
we will see below that the non - oscillating contribution @xmath84 is necessary for the consistency of the approach , meaning that nonlinearity not only modifies the shape of a finite amplitude wave but also affects the background on top of which the wave propagates .
we enforce a behavior of type only at order @xmath86 ; it then will be automatically verified at higher orders , with also possible contributions from higher harmonics : see for the form of the @xmath87 solution .
the multiscale analysis consists in considering that the time variables @xmath88 , @xmath89 and @xmath90 ( and also the spatial coordinates @xmath71 and @xmath91 ) are independent .
one thus writes @xmath92 the method applies for any value of @xmath93 , provided one remains in phase iii , but the general expressions are quite cumbersome : for legibility we present the computation in the simpler case @xmath39 . at this order , eq
. reads [ as already obtained in eq . ]
@xmath94 where @xmath95 is defined in . using the matrix @xmath58 of eq . and defining @xmath96 by @xmath97 one can re - write eq . as @xmath98 for to have nontrivial solutions we need to impose @xmath99 ( we already have @xmath100 ) .
as was seen in section [ model ] , this determines the dispersion relation .
we study here a wave propagating in the upper branch of the spectrum , that is , in the expression for @xmath101 , one has @xmath68 .
we then obtain for the solutions @xmath102 and @xmath103 of eqs .
expressions of the form : @xmath104 and @xmath105 where @xmath106 . at this point @xmath107 and @xmath108 in expressions and are still unknown , but we already collected some useful pieces of information on the form of the wave : we see that @xmath109 and that @xmath110 is proportional to @xmath111 . in the case
@xmath112 , @xmath113 and @xmath114 are non zero , but both are proportional to @xmath111 , as well as @xmath110 . at this order eq
. reads @xmath116=\\ & \overline{c}_0(x , t_1,t_2 ) + \\ & \left[\widetilde{c}_1(x , t_1,t_2 ) \operatorname{e}^{i \beta(x , t)}+\operatorname{c.c.}\right ] + \\ & \left[\widetilde{c}_2(x , t_1,t_2 ) \operatorname{e}^{2 i \beta(x , t)}+ \operatorname{c.c.}\right ] \ ; . \end{split}\ ] ] with @xmath117 @xmath118 and @xmath119 in expressions and we have used the same notation @xmath120 as in .
the precise expressions , and for @xmath121 , @xmath122 and @xmath123 result from the formulas and for @xmath102 and @xmath85 .
since the operator @xmath124 is linear , the solution of equation consists of three contributions , one for each of the source terms .
hence @xmath125 is of the form @xmath126\\ & + \left[\widetilde\xi^{(2)}_2(x , t_1,t_2)\operatorname{e}^{2i\beta } + \operatorname{c.c.}\right]\ ; , \end{split}\ ] ] the different components being solutions of @xmath127 @xmath128 and @xmath129 where @xmath130 is defined similarly to @xmath58 in eq . and
@xmath96 in eq .
: @xmath131 equation is easily solved because @xmath132 @xcite .
we do not write its solution explicitly , but it is necessary for next order in @xmath86 : it contributes to the r.h.s . of eq .
, in particularly to the expression for the coefficient @xmath133 . solving eqs . and
is more complicated than solving eq . because @xmath134 and @xmath135 .
hence , if @xmath121 is not in the image space of @xmath96 ( or if @xmath122 is not in the image space of @xmath58 ) one can not find a solution .
one must thus impose that @xmath121 is in the image space of @xmath96 and that @xmath122 is in the image space of @xmath58 .
this can be done conveniently through the following technique .
let us define @xmath136 and @xmath137 such that @xmath138 multiplying by the transposed row vector @xmath139 and by @xmath140 one obtains @xcite @xmath141 the first of these equations is trivially satisfied .
the second imposes that @xmath142 which implies the important physical result that the envelope of the wave packet propagates with the group velocity @xmath143 .
once eq . is satisfied , the compatibility condition is fulfilled and one can solve eqs . and .
one obtains @xmath144 and @xmath145 at this order one obtains an equation whose form is quite similar to that of eq . with additional harmonics : @xmath147= \\ &
\overline{d}_0(x , t_1,t_2 ) + \\ & \left[\widetilde{d}_1(x , t_1,t_2 ) \operatorname{e}^{i \beta(x , t)}+\operatorname{c.c.}\right ] + \\ & \left[\widetilde{d}_2(x , t_1,t_2 ) \operatorname{e}^{2 i \beta(x , t)}+ \operatorname{c.c.}\right ] + \\ & \left[\widetilde{d}_3(x , t_1,t_2 ) \operatorname{e}^{3 i \beta(x , t)}+ \operatorname{c.c.}\right]\ ; . \end{split}\ ] ] we need not write the expressions for @xmath148 and @xmath149 because they are not necessary to determine the dynamic of @xmath150 .
the terms @xmath151 and @xmath133 read ( remember that for legibility we give the explicit expressions only in the case @xmath39 ) @xmath152 and @xmath153 in the above expression for @xmath133 the quantities @xmath154 and @xmath155 are defined as @xmath156\ ; , \end{split}\ ] ] and @xmath157 \ ; , \end{split}\ ] ] where @xmath158 following the same method as in section [ sec : order2 ] , we write @xmath159 + \\ & \left[\widetilde\xi^{(3)}_2(x , t_1,t_2 ) \operatorname{e}^{2 i \beta(x , t)}+ \operatorname{c.c.}\right]+\\ & \left[\widetilde\xi^{(3)}_3(x , t_1,t_2 ) \operatorname{e}^{3 i \beta(x , t)}+ \operatorname{c.c.}\right ] \ ; . \end{split}\ ] ] we will not need to consider the contribution of the second and third harmonics in and .
but the contributions of the first harmonic ( @xmath160 and @xmath161 ) and of the `` zero''-harmonic ( @xmath162 and @xmath163 ) are important . reinserting expression in yields : @xmath164 and @xmath165 again , for solving eq .
one must make sure that @xmath151 is in the image space of @xmath96 : this yields @xmath166 which writes @xmath167 where @xmath168 is the speed of sound [ cf .
eq . ] and @xmath169 the solution of reads ( computations are explained in appendix [ app1 ] ) : @xmath170 where @xmath171 ^ 2-c^2}\ ; .\ ] ] expression combined with eq .
shows that @xmath172 this result shows that the deformation of the background propagates with the group velocity , as does the envelope of the wave [ which obeys the same equation , cf . ] . finally , for being able to solve eq .
we need @xmath133 be in the image space of @xmath58 : @xmath173 this reads @xmath174 \frac{\widetilde\theta^{(1)}}{2}\ ; , \end{split}\ ] ] where @xmath175 .
one can re - express the term @xmath176 using eq . .
one then obtains a nonlinear schrdinger equation ( nls ) for @xmath108 : @xmath177 with @xmath178 one has reached a point where the approach is self - contained , as far as the first order term @xmath179 in expansion is concerned .
one just needs to return to the actual variables @xmath71 and @xmath88 using the reverse of transformations @xcite .
we give below final formulas valid even when @xmath112 .
a nonlinear wave packet propagating in the upper branch is described by a set of fields @xmath44 of the form with @xmath180 \ ; .\ ] ] the component @xmath181 of the envelope @xmath182 is solution of @xmath183 where @xmath184 is the space coordinate in a frame moving at the group velocity .
once @xmath185 has been determined , the component @xmath186 of the background deformation is obtained as @xmath187 the other components of the background and of the envelope are given by @xmath188 and @xmath189 where @xmath190 and @xmath191 -\frac{i \left((2 g_2+k^2 + 2 \omega ) ( k^2 + 2 \omega ) + 4 k^2 k_0 ^ 2 - 4 \omega_+(k ) ^2\right)}{4 k k_0 \left(k^2+\omega \right ) } \\[1 mm ] 1\\[1 mm ] \frac{i \left(k^2 ( 2 g_2+k^2 - 4 k_0 ^ 2 + 2 \omega ) + 4 \omega_+^2(k)\right)}{4 \left(k^2+\omega \right)\omega_+(k ) } \end{pmatrix}\ ; .\ ] ] we do not write here the explicit forms of @xmath192 and @xmath193 for @xmath194 because they are too cumbersome . however , it is important for subsequent discussions to stress that eq
. still holds for @xmath112 , but with a numerator @xmath195 whose expression is different from the one given in eq . for the case @xmath39 . on the other hand , the formulas , , and are valid even when @xmath112 . note that @xmath196 is identical to @xmath197 defined in and that @xmath198 reduces to @xmath199 defined in when @xmath39 . in this case , the first two components of @xmath198 cancel and the nonlinear structure corresponds to a polarization signal , with oscillations of @xmath200 and @xmath201 preserving a fixed total density .
the nonlinear schrdinger equation describes the spatio - temporal evolution of the envelope wave which is advected by the group velocity @xmath202 while dispersion and nonlinearity give corrections to the dynamics of the wave train , in particular for large times .
it has been obtained through a multiscale expansion assuming the existence of well separated spatial and temporal scales .
the two spatial scales are the wave length @xmath203 ( associated with the @xmath71-dependence of the phase @xmath101 ) and the length @xmath204 characteristic of the spatial variations of the envelope of the wave packet ( associated to coordinate @xmath91 ) .
these length scales should be widely different and this corresponds to defining our small parameter as @xmath205 the three time scales legitimating the introduction of the three different time coordinates @xmath88 , @xmath89 and @xmath90 are : @xmath206 where @xmath207 is the period of the carrier wave . the characteristic time @xmath208 is associated to the group motion of the envelope . the time
@xmath209 accounts for the fact that the envelope not only propagates with the group velocity , but also changes form because of higher order dispersive effects and of nonlinearity ( both effects typical balance in a nonlinear wave such as the soliton solutions discussed below ) . from eqs .
and one can check that when the condition is fulfilled one has @xmath210 , thus legitimating _ a posteriori _ the introduction of the three time coordinates .
when @xmath193 is positive , periodic wave trains with constant amplitude formed in the upper branch of the spectrum are dynamically stable .
they can support nonlinear excitations such as dark solitons . in this case @xmath211 solution of is of the form @xmath212\ ; , \ ] ] where @xmath213 is the amplitude of the wave train and @xmath214 $ ] ; @xmath215 is the dimensionless velocity of the dark soliton , cf .
the argument @xmath216 in is @xmath217 and @xmath218 . in the case where @xmath219 , wave trains in the upper branch are dynamically unstable ( they experience a modulational instability , see below ) , but one may observe stable bright envelope solitons , for which the solution of is of the form @xmath220 where @xmath221 is a positive real parameter governing the amplitude of the soliton .
once @xmath211 is known , the corresponding values of the other fields describing the system are then given by , and .
the density profiles of typical envelope solitons are plotted in fig .
[ fig - soliton ] . ,
@xmath222 and @xmath223 as functions of @xmath224 for a dark envelope soliton ( upper plot ) and a bright envelope one ( lower plot ) .
the system s parameters are the same as in fig .
[ fig - geff1 ] .
for both plots @xmath225 .
the dark soliton is plotted for @xmath226 ( @xmath227 ) and @xmath228 ( black soliton ) and the bright soliton for @xmath229 ( @xmath230 ) . ] in order to get better insight on the type of dynamics described by the envelope nls equation , we show in fig .
[ fig - geff1 ] how the effective nonlinear constant @xmath231 depends on @xmath51 .
for the choice of parameters @xmath39 , @xmath232 , @xmath233 and @xmath65 . ]
we see that @xmath193 starts at low @xmath51 with a negative value , and since @xmath234 , this means that wave trains in the upper branch experience a modulational instability , see , e.g. , ref .
@xcite and references therein .
modulational instability in bose - einstein condensates with repulsive interaction have already been studied in the presence of an external optical lattice potential @xcite and for the counterflow of two miscible species @xcite .
here we consider a scenario closer to the original bejamin - feir configuration @xcite , where nonlinearity destabilizes a periodic wave - train through generation of spectral sidebands [ see the discussion below , around eqs . and ] . when discussing the physical origin of the modulational instability in the present context , it is interesting to note that @xmath193 diverges and changes sign for a value of @xmath51 which is denoted as @xmath235 in fig .
[ fig - geff1 ] . for this value of @xmath51 ,
the system displays a so called long wave - short wave resonance @xcite . in the present configuration
this corresponds to a case where the wave in the upper branch ( with wave vector @xmath51 ) decays into two waves , one in the same branch with a similar wave vector ( @xmath236 ) , and an other one in the lower branch , with a small wave vector ( @xmath237 , the `` long wave '' ) .
the conditions of conservation of momentum and energy read @xmath238 and @xmath239 since @xmath237 is small one can expand the first term of the r.h.s . of as : @xmath240 , and also write @xmath241 .
hence , the phenomenon occurs at @xmath242 such that @xmath243 meaning that the condition of resonance is that the group velocity of the short wave is equal to the phase velocity of the long wave . the location of the resonance is clearly seen in fig . [ fig - geff1 ] , at a value of @xmath51 in exact agreement with the value @xmath235 determined by . from the derivation leading to the nls eq . , one can locate the mathematical origin of the resonance phenomenon in eq . , where @xmath192 as given by clearly diverges exactly at resonance .
the phenomenological analysis just presented assumes that this divergence corresponds to a transfer of excitation from the upper branch to the lower one , but one should ascertain that this is indeed the case in our mathematical treatment .
indeed , it might seem from eq . that the divergence is connected to a resonance with a deformation of the background ( of @xmath244 ) which might not be exactly connected to the lower branch of excitation .
a first clue of this connection comes from the l.h.s .
of eq . itself
: in this equation , the zero mode of the operator acting on @xmath244 corresponds to a dispersion relation which is the long wave length approximation of the lower branch : @xmath245 .
the second and final reason explaining why in this context , @xmath244 indeed represents the lower branch excitation comes from the very reason for its appearance in : it originates from eq .
, more precisely , from the specific form of @xmath246 which is tailored to be representative of the kernel of @xmath96 . and , as can be checked by a comparison of the forms and definitions of @xmath58 and @xmath96 , @xmath96 is the @xmath247 limit of @xmath58 when @xmath70 : hence the background contributions in the ansatz ( and in the higher order terms ) is indeed a low @xmath51 contribution in the lowest branch and the divergence of @xmath192 in indeed corresponds to a resonance between the upper branch and the ( long wavelength limit of ) the lower branch .
it is remarkable that the occurrence of the long wave - short wave resonance is connected to a disappearance of the modulational instability of the upper branch : as one can see from fig .
[ fig - geff1 ] the nonlinear parameter @xmath193 is positive when @xmath51 is larger than @xmath235 and wave trains in the upper branch are thus stable when their wave - vector is larger than the one of the long wave - short wave resonance . in order to appreciate the origin of this phenomenon one
first needs to get some physical insight on the cause of the modulational instability .
since the reasoning presented below is quite general , and for simplifying the notations , we will here for a moment denote the dispersion relation as @xmath248 instead of @xmath249 .
if one studies a wave train with wave vector @xmath51 and constant ( real ) amplitude @xmath250 , one finds from that the corresponding @xmath185 is equal to @xmath251 .
then , a perturbative treatment of eq . readily shows ( see , e.g. , refs .
@xcite ) that small amplitude modulations of the carrier wave with relative wave vector @xmath237 and angular frequency @xmath252 obey the dispersion relation @xmath253 if @xmath193 is negative , @xmath252 will be imaginary ( for low enough values of @xmath237 ) , meaning that the wave train is dynamically unstable .
the value @xmath254 of @xmath237 corresponding to the largest imaginary part of @xmath252 , i.e. , to the greatest growth rate of the perturbations , verifies @xmath255 one gets here a confirmation that the wavetrain is unstable when @xmath256 is negative .
this corresponds to the so - called lighthill - benjamin - feir criterion of modulational instability @xcite , which can be given the following intuitive interpretation : one assumes that a wave - train of finite amplitude @xmath250 corresponds to the renormalized dispersion relation @xmath257 the initial carrier wave at angular frequency @xmath248 and wave - vector @xmath51 may decay into two side bands according to the following process : @xmath258 where @xmath254 as given by enforces the energy and momentum conservation relations in the process , as can be checked analytically ( by an expansion in @xmath254 ) and is graphically demonstrated in fig .
[ fig - mi ] .
it is clear that , when @xmath259 , the geometrical construction of fig .
[ fig - mi ] is only possible if @xmath219 : in this case the wave train is modulationally unstable . at @xmath260 .
two initial elementary excitations @xmath261 can decay into @xmath262 and @xmath263 provided @xmath254 verifies the above construction , i.e. , that the two black points are the intersections of the straight solid line with the renormalized dispersion relation @xmath264 .
an expansion of @xmath265 at second order in the vicinity of @xmath51 shows that ( i ) the white point is the middle of the two black points ( and this automatically implies momentum and energy conservation in the process ) and that ( ii ) @xmath254 defined by the above construction verifies eq . .
] hence we understand why the change of sign of @xmath193 observed in fig .
[ fig - geff1 ] when @xmath51 crosses @xmath235 changes the stability of the wave train .
now it remains to understand the physical reason for this change of sign .
actually , the reason for it becomes clear when one focuses on the nonlinear term in brackets in eq . .
the first part of this term [ with the @xmath154 and @xmath155 contributions ] is a genuine nonlinear self - interaction , but the second part is proportional to @xmath266 , i.e. , to @xmath267 which is a doppler contribution to the energy of an excitation moving over a background of velocity @xmath268 . for @xmath269 ,
the momentum @xmath237 imparted to the lower branch is negative , and the corresponding value of @xmath268 is also , as physically clear and mathematically demonstrated by the fact that in this case @xmath270 [ see eq . ] .
it so happens that this doppler contribution is dominant over the self - interaction terms , and , as a result , @xmath271 . on the contrary , for
@xmath272 the momentum imparted to the lower branch is positive , @xmath273 and @xmath274 .
this ends our discussion of the behavior of @xmath275 around @xmath276 and the explanation for the disappearance of the modulational instability when @xmath277 .
besides the long wave - short wave resonance , one can notice an other resonant - like structure in fig .
[ fig - geff1 ] .
it corresponds to a generation of second harmonic according to the three waves process @xmath278 the condition of conservation of momentum and energy in the above process determines the value of the resonant wave vector @xmath279 in excellent agreement with the location of the divergence of @xmath193 observed in fig .
[ fig - geff1 ] . in the vicinity of @xmath279
our approach fails ( and the envelope nls equation is not relevant ) because the determinant of @xmath130 vanishes , contrarily to what has been stated after eq . , and the procedure that has been used for determining @xmath280 from eq .
is incorrect . in this case
the assumption that higher - order harmonics have a very small contribution is wrong .
the fact that second harmonic generation is associated with vanishing of the determinant of @xmath130 is an immediate result of the definition and of energy conservation in the process : at resonance one has @xmath281 .
the determinant of this last matrix is zero , because , for any @xmath282 , @xmath283 , since @xmath284 is one of the dispersion relations of the system . for concluding the discussion , it is interesting to notice that , besides the second harmonic generation identified in fig .
[ fig - geff1 ] , there exists an other possible generation of second harmonics , which only involves excitations of the upper branch : @xmath285 this new process should induce a divergence of @xmath193 at the wave vector @xmath286 which ensures energy conservation in the process .
indeed , in this case we have a linear system @xmath287 that has a zero determinant because @xmath288 is one of the dispersion relations of the system .
for the set of parameters corresponding to fig .
[ fig - geff1 ] , this second harmonic generation should occur at @xmath289 .
it is then surprising that this resonance is not seen in this figure .
however , it is clearly seen when @xmath112 ( see fig .
[ fig - geff2 ] ) , at the value predicted by the conservation of energy in .
for the choice of parameters @xmath290 , @xmath232 , @xmath233 and @xmath65 .
the location of the resonances is determined by momentum and energy conservation in the processes ( for @xmath235 ) , ( for @xmath279 ) and ( for @xmath291 ) . ] actually , in the case where @xmath39 , at @xmath286 one has @xmath292 , and the divergent factor involved in the determination of @xmath280 from eq .
( and which results in a divergence in the expression of @xmath193 ) is canceled by an other contribution .
this can be easily understood by noticing that @xmath95 defined in eq .
is a block matrix when @xmath39 : @xmath293 where @xmath294 and @xmath295 are @xmath296 matrices accounting for the lower and the upper excitation branches .
we are here interested in second harmonic generation , i.e. , in the specific matrix @xmath297 . in this case
, we denote the matrices @xmath294 and @xmath295 as @xmath298 and @xmath299 and their inverses are @xmath300 where `` adj '' denotes the adjugate matrix .
the divergence of @xmath301 is associated with the divergence of the denominator in , corresponding to energy conservation in the process . in the special case @xmath39 , the solution of eq .
reads : @xmath302 where @xmath123 is given by eq . when @xmath39 : in this case its last two components are zero .
eq . then reads @xmath303 \mathbb{m}_{2-}^{-1 } \begin{pmatrix } i\,\delta\ , k^2 \\ \frac{1}{2 } \left ( \frac{k^2 - 2\,\omega}{k^2 + 2\,\omega}\,g_2 - k^2 - 2\,\omega \right ) \end{pmatrix}\\ \mathbb{m}_{2+}^{-1 } \begin{pmatrix } 0 \\ 0 \end{pmatrix}=\begin{pmatrix } 0 \\ 0 \end{pmatrix } \end{pmatrix*}\left(\widetilde\theta^{(1)}\right)^2 \ ; , \ ] ] and the possible divergence of the denominator of @xmath304 is masked .
this is the reason for the inhibition of the second harmonic generation process when @xmath39 .
we now study the propagation of a sound pulse which , in a linear approximation , would lie on the lower excitation branch . the method used in section [ sec : upperbranch ]
can be employed in the present case .
it yields for the nonlinear coefficient @xmath193 a behavior represented in fig .
[ fig - geff - lb ] . for the envelope nls equation describing a wave packet propagating in the lower excitation branch .
the curves are drawn for different values of @xmath305 ; the other parameters are @xmath306 , @xmath232 and @xmath233 . ]
the nonlinear coefficient diverges at large wave - length .
this is due to the fact that , for the lower branch , the analog of the coefficient @xmath192 diverges when @xmath247 since @xmath307 . in this case
the nonlinear time @xmath308 associated to eq . diverges indicating that nonlinear structures form extremely rapidly .
@xmath309 may become even smaller that the period of the wave ( except for waves of extremely small amplitude ) and in this case the technique of the envelope nls fails . in this long wave length limit one can suggest an alternative method consisting in deriving equations for the interacting fields themselves instead of an effective equation for the envelope .
this method is based on the following reasoning : in the linear regime and at the level of accuracy at which the expansion holds , any of the components of @xmath46
@xmath310 say satisfies the linear equation @xmath311 where the last term describes a small dispersive correction to the propagation with constant velocity @xmath168 . if the amplitude @xmath310 is small but finite and such that this term has the same order of magnitude as the leading nonlinear correction to ( which is typically quadratic in @xmath310 ) , then nonlinear effects can not be omitted for correctly describing the propagation of the pulse . in this regime
one can try to derive an equation of the type with additional terms taking into account weak nonlinear effects .
the most natural extension of is a korteweg - de vries ( kdv ) equation in which a nonlinear term of the form @xmath312 accounts for a dependence in density fluctuations ( @xmath313 ) of the velocity of sound .
it now is appropriate to work in a reference frame moving at the speed of sound @xmath168 , and to use @xmath314 and @xmath88 as coordinates . in order
that the derivatives in equations of type appear at the same order , we define @xmath315 where @xmath77 will henceforth be a small positive parameter .
the choice of the specific powers @xmath316 and @xmath317 in ( instead of @xmath77 and @xmath318 for instance ) will make sure that the derivatives in equations of type appear at the same order as the quadratic nonlinear contribution [ @xmath319 , see eq .
below ] . in terms of the new variables @xmath320 and @xmath321 and of the velocities @xmath322 and @xmath323 defined in eq . , the system reads [ kdv2 ] @xmath324_\xi \nonumber \\ -\ ; & \tfrac12 \epsilon^{1/2 } \big[n\left(u - v\cos\theta \right)\big]_\xi \ ; , \\
\epsilon^{3/2}u_\tau = \ ; & \epsilon^{1/2}c\,u_\xi - \epsilon^{1/2 } k_0 \ , v_\xi \nonumber \\ -\ ; & \epsilon^{1/2 } \left[\frac{\epsilon}{4 } \frac{n_\xi^2}{n^2 } - \frac{\epsilon}{2 } \frac{n_{\xi\xi}}{n } - \epsilon \frac{\cot \theta}{2 } \frac{\left[n \ , \theta_\xi \right]_\xi}{n}\right]_\xi \nonumber \\
-\;&\epsilon^{1/2 } \left[\frac{\epsilon\theta_\xi^2+u^2+v^2}{4}+ g_1 n \right]_\xi \nonumber \\
+ \;&\epsilon^{1/2}\theta_\xi\frac{\omega \cos\theta}{\sin^2\theta } \cos\varphi + \frac{\omega}{\sin\theta } \ ,
v\ , \sin\varphi \ ; , \\
\epsilon^{3/2}\theta_\tau = \ ; & \epsilon^{1/2}c \ , \theta_\xi - \omega \ , \sin\varphi \nonumber \\ -\ ; & \tfrac12 \epsilon^{1/2 } \left(u \theta_\xi + \frac{\left[n\,(v+2k_0)\sin\theta \right]_\xi}{n } \right ) \ ; , \\ \epsilon^{3/2}v_\tau=\ ; & \epsilon^{1/2}c\,v_\xi - \epsilon^{1/2 } k_0 \ , u_\xi \nonumber \\ + \ ; & \epsilon^{1/2 } \left[\frac{\epsilon}{2 \sin \theta } \frac{\left[n \ , \theta_\xi \right]_\xi}{n}-\frac{u\,v}{2 } + g_2 n \cos\theta\right]_\xi \nonumber \\
+ \;&\epsilon^{1/2}\theta_\xi\ , \frac{\omega \cos\varphi}{\sin^2\theta } + \omega \ ,
v\ , \cot\theta\ , \sin\varphi \;.\end{aligned}\ ] ] we perform a multi - scale analysis by expanding @xmath325 in the following way : @xmath326 in agreement with the definitions and we have @xmath327 taking into account expansion and the leading order for @xmath22 and @xmath25 we thus have @xmath328 once the anstze and are inserted back into , the leading term is @xmath329 and simply yields @xmath330 ( and thus @xmath331 ) . at next order ( @xmath317 ) one obtains @xmath332 u_\xi^{(1)}\\[1 mm ] \theta_\xi^{(1)}\\[1mm]\varphi^{(2)}\end{pmatrix } = 0\ ; , \ ] ] where @xmath333 since @xmath334 , eq . has non trivial solutions .
the kernel of @xmath335 is one - dimensional ; as a result , the solution of eq .
is of the form : @xmath336 u^{(1)}_\xi \\[1 mm ] \theta^{(1)}_\xi \\[1 mm ] \varphi^{(2 ) } \end{pmatrix } = \begin{pmatrix } 1 \\[1mm]\frac{g_1}{c } \\[1 mm ] \frac{1}{c}\frac{k_0 g_1}{\omega + g_2 } \\[1 mm ] \frac{k_0}{\omega } \frac{\omega+g_1+g_2}{\omega+g_2 } \end{pmatrix } n^{(1)}_\xi \equiv r\,n^{(1)}_\xi \;.\ ] ] we also need ( for later use ) to determine the column vector @xmath337 such that @xmath338 this fixes @xmath339 at order @xmath340 we obtain @xmath341 u_\xi^{(2)}\\[1mm]\theta_\xi^{(2 ) } \\[1mm]\varphi^{(3)}\end{pmatrix } = \begin{pmatrix } a_1\\[1 mm ] a_2\\[1mm]a_3\\[1mm]a_4\end{pmatrix } \ ; , \ ] ] where @xmath342_\xi - k_0 [ n^{(1 ) } \theta^{(1)}]_\xi \ ; , \\ a_2=\ ; & u^{(1)}_\tau + \tfrac12 u^{(1 ) } u^{(1)}_\xi - \omega \ , \theta^{(1 ) } \theta^{(1)}_\xi \\ & + \tfrac12 n^{(1)}_{\xi\xi\xi } + k_0 v^{(2)}_\xi \ ; , \\ a_3=\ ; & \theta^{(1)}_\tau+\tfrac12 u^{(1)}\theta_\xi^{(1 ) } -\tfrac12 v_\xi^{(2)}\\ & + k_0\theta^{(1)}\theta_\xi^{(1)}+k_0 n^{(1)}n_\xi^{(1)}\ ; , \\ a_4=\ ; & -g_2[n^{(1 ) } \theta^{(1)}]_\xi+ \tfrac12 \,\theta^{(1)}_{\xi\xi\xi } -c \,v^{(2)}_\xi\ ; . \end{split}\ ] ] performing the substitution , we can express the system in the following way : @xmath343 u_\xi^{(2)}\\[1mm]\theta_\xi^{(2)}\\[1mm]\varphi^{(3)}\end{pmatrix } = c_\tau \ , n_\tau^{(1 ) } + c_3\,n^{(1)}_{\xi\xi\xi } + c_{\mathrm{nl}}\ , n^{(1)}n^{(1)}_\xi \;,\ ] ] with @xmath344\frac{g_1}{c } \\[1 mm ] \frac{1}{c}\frac{k_0 g_1}{\omega + g_2 } \\[1 mm ] 0 \end{pmatrix } \ ; , \ ] ] @xmath345 and @xmath346 left multiplication of eq . by @xmath347
gives @xmath348 equation is a consistency condition : eq .
admits a solution only if the column vector @xmath349 is in the image space of @xmath335 , which is implied by ( we used the same technique in sec . [
sec : upperbranch ] , see eqs . , and ) .
explicitly , eq . reads @xmath350 where @xmath74 is the third order coefficient of the dispersion relation .
going back to the original coordinates @xmath71 and @xmath88 and denoting @xmath351 , we obtain the kdv equation @xmath352 where @xmath353 once the solution of eq .
is found , the other field variables can be obtained using relations which we rewrite here for completeness in the final notation : @xmath354 note that when @xmath310 becomes of order of the nonlinear coefficient in the kdv eq .
, that is when @xmath355 , the level of accuracy accepted here is not sufficient : the cubic nonlinear terms ( @xmath356 ) neglected in the present treatment have the same order of magnitude as the quadratic term in eq . .
in this limit we have to consider the next order of approximation . as advocated in sec .
[ kdvregime ] , cubic nonlinearities become important when @xmath357 is small : their contributions can therefore be calculated from the system choosing parameters such that @xmath358 .
this is achieved when @xmath359 for this choice of parameters , the sound velocity reads @xmath360 in this case the system can sustain long wavelength perturbations only if @xmath361 , that is if @xmath362 , which we assume henceforth ( see however the discussion at the end of sec .
[ quartic ] and appendix [ app2 ] ) . in this regime
the coordinates defined in are no longer appropriate for the description of non - linear excitations .
one should instead perform the computations with the new coordinates : @xmath363 then the system rewrites : [ gar3 ] @xmath364_\xi \nonumber \\ -\ ; & \tfrac12 \epsilon \big[n\left(u - v\cos\theta \right)\big]_\xi \ ; , \\
\epsilon^3 u_\tau = \ ; & \epsilon c\,u_\xi - \epsilon k_0 \ , v_\xi \nonumber \\ -\ ; & \epsilon \left[\frac{\epsilon^2}{4 } \frac{n_\xi^2}{n^2 } - \frac{\epsilon^2}{2 } \frac{n_{\xi\xi}}{n } - \epsilon^2 \frac{\cot \theta}{2 } \frac{\left[n \ , \theta_\xi \right]_\xi}{n}\right]_\xi \nonumber \\ -\;&\epsilon \left[\frac{\epsilon^2\theta_\xi^2+u^2+v^2}{4}+ g_1 n \right]_\xi \nonumber \\
+ \;&\epsilon\theta_\xi\frac{\omega \cos\theta}{\sin^2\theta } \cos\varphi + \frac{\omega}{\sin\theta } \ ,
v\ , \sin\varphi \ ; , \\
\epsilon^3\theta_\tau = \ ; & \epsilon c \ , \theta_\xi - \omega \ , \sin\varphi \nonumber \\ -\ ; & \tfrac12 \epsilon \left(u \theta_\xi + \frac{\left[n\,(v+2k_0)\sin\theta \right]_\xi}{n }
\right ) \ ; , \\ \epsilon^3 v_\tau=\ ; & \epsilon c\,v_\xi - \epsilon k_0 \ , u_\xi \nonumber \\ + \ ; & \epsilon \left[\frac{\epsilon^2}{2 \sin
\theta } \frac{\left[n \ , \theta_\xi \right]_\xi}{n}-\frac{u\,v}{2 } + g_2 n \cos\theta\right]_\xi \nonumber \\ + \;&\epsilon \theta_\xi\,\frac{\omega \cos\varphi}{\sin^2\theta } + \omega \ , v\ , \cot\theta\ , \sin\varphi \;.\end{aligned}\ ] ] we perform a multiscale analysis using the anstze ; @xmath322 and @xmath323 now read : @xmath365 the leading term in is now @xmath83 and reads , as previously , @xmath330 . the next order @xmath366 is described by the same system as in eq . .
substituting @xmath367 , @xmath368 and @xmath369 by their expression in @xmath370 defined in , the order @xmath371 then reads : @xmath372 u_\xi^{(2 ) } \\[1mm]\theta_\xi^{(2)}\\[1mm]\varphi^{(3)}\end{pmatrix}= c_{\mathrm{nl } } \ , n^{(1 ) } n^{(1)}_\xi \ ; , \ ] ] where @xmath373 is defined in eq . .
since in this subsection @xmath93 is fixed such that the non - linearity ( @xmath374 ) in eq .
cancels , the choice of parameter automatically implies @xmath375 and from eq .
one can only deduces that @xmath376 u_\xi^{(2)}\\[1mm]\theta_\xi^{(2)}\\[1mm]\varphi^{(3 ) } \end{pmatrix } = \begin{pmatrix } 0 \\ 0 \\ \frac{2\sqrt{g_1(-g_2)}}{\omega+g_2 } \\
\frac{g_1 \left(3 g_2 \omega -g_2 ^ 2 + 2 \omega ^2\right)-g_2 \left(g_2+\omega \right){}^2}{\sqrt{2 } g_2 \omega \left(g_2+\omega \right ) } \end{pmatrix } n^{(1 ) } n^{(1)}_\xi \;.\ ] ] eq .
is thus not conclusive and the expansion at order @xmath371 is not sufficient to describe the dynamic of non - linear excitations . at next order [ @xmath377 , taking into account the formula ,
we obtain : @xmath378 u_\xi^{(3)}\\[1mm]\theta_\xi^{(3)}\\[1mm]\varphi^{(4)}\end{pmatrix } = c_\tau n_\tau^{(1 ) } + d_{\mathrm{nl}}\ , { n^{(1)}}^2n^{(1)}_\xi + c_3\,n^{(1)}_{\xi\xi\xi } \;,\ ] ] where @xmath379 and @xmath380 are defined in eqs . , and @xmath381 left multiplication of eq . by @xmath347 gives @xmath382
which reads @xmath383 going back to the original coordinates @xmath71 and @xmath88 we obtain the modified kdv ( mkdv ) equation @xmath384 where @xmath385 in the regime where @xmath386 is not exactly zero , but of order @xmath310 , we also have to take into account the quadratic nonlinearity of eq . , which finally yields @xmath387 this is the gardner equation describing the evolution of nonlinear polarization pulses in a coherently coupled two - component condensate in the limit where the parameters of the system are close to satisfy the condition .
this can be considered as an intermediate region where the quadratic and cubic nonlinearities make contributions of the same order of magnitude in the wave dynamics . in the limit of very small @xmath386 ,
the quadratic nonlinearity effects can be neglected , the nonlinear polarization waves are correctly described by the modified kdv equation . if instead @xmath386 is large , then the cubic nonlinearity effects are negligible and the evolution of nonlinear polarization pulses is described by the kdv equation .
note that for consistency reasons , the value of the sound velocity in has to evaluated as not being exactly equal to @xmath388 , but has to include corrections @xmath389 .
once the solution of the gardner equation has been found , the other field variables can be expressed in terms of @xmath310 by the formulas with account of .
the parameters of the system can be chosen in such a way that not only @xmath386 , but also @xmath390 cancels .
this is achieved for the choice with the additional constrain @xmath391 which ensures that @xmath392 .
note that for having a positive value of @xmath393 , one must have @xmath40 , but this condition is automatically fulfilled in phase iii [ cf . the definition of the sound velocity ] .
computations very similar to the ones exposed in secs .
[ kdvregime ] and [ gardner_regime ] now lead to a higher order mkdv equation @xmath394 where @xmath395 finally we can write the general form which is able to account for choices of parameters such that @xmath396 and @xmath397 : @xmath398 with @xmath399 eq .
is known as a generalized kdv equation @xcite . at this point
, it might be helpful to remind the strategy followed in the present section : we study excitations of the lower branch of the spectrum , which , in the linear regime and in the long wave limit , are described by eq . . in order to analyze how nonlinearity affects these excitations
, we consider the modifications of the pulse propagation velocity induced by the nonlinear effects : @xmath400 with an expansion of the form @xmath401 .
the multi - scale analysis consists in rescaling the variable @xmath402 in the following way : @xmath403 transforming eq . into @xmath404
the analysis amounts to determine the coefficients @xmath405 ; this has been done in eqs . , and eq . .
the first correction is @xmath406 : @xmath407 the approach is coherent if all orders in @xmath77 in eq . are identical , i.e. , if the stretched variables in are chosen with @xmath408
. the parameter @xmath409 can vanish or become small for a particular value of @xmath93 ; the first order correction @xmath406 is then no longer sufficient , and we must consider the correction @xmath410 which corresponds , by the same argument as the one used after eq . , to @xmath411 ; these are the exponents used in sec .
[ gardner_regime ] . if @xmath412 is also small , one has to consider the next order , as done in the beginning of the present section .
the different orders considered and their regime of relevance are recalled in table [ table1 ] .
.list of the different nonlinear equations describing the weakly nonlinear and weakly dispersive dynamics of excitations which , in the linear regime , pertain to the lower dispersion branch .
the right column shows their successive regime of relevance of the equations .
the conditions @xmath358 and @xmath392 are precisely defined in eqs . and .
note that each row assumes that the regime of relevance of the upper rows is fulfilled .
[ cols="<,<",options="header " , ] it is appropriate to discuss if the different regimes identified in table [ table1 ] can be reached with current days experimental realization of spin - orbit coupled becs . the references @xcite consider the two states @xmath413 of a @xmath414rb bose - einstein condensate in the @xmath415 hyperfine structure .
the @xmath416-wave scattering lengths are ( in units of the bohr radius ) @xmath417 and @xmath418 .
since @xmath419 the simplifying assumption of a common value of the nonlinear coupling @xmath420 in is legitimate ( in this expression @xmath421 is the angular frequency corresponding to a tight harmonic radial trapping which ensures a quasi-1d behavior of the condensate @xcite ) . besides , if necessary
, the present formalism can be extended to take into account the fact that @xmath422 and @xmath423 are not equal , see ref .
note that the recoil energy is typically @xmath424 khz ( it is monitored by the wavelength and the relative angle of the raman lasers ) , whereas the interaction energy @xmath425 is of order @xmath426 khz ( it depends on the value of the radial trapping frequency and on the linear atomic density ) .
one has @xmath427 , hence @xmath428 , and it seems from the discussion in the beginning of section [ gardner_regime ] that one can never reach the interesting regime where @xmath358 and where the nonlinear modulations of an excitation formed in the lower branch is described by gardner equation .
however , @xmath429 is small ( since @xmath430 and @xmath431 are so close ) and , as explained in appendix [ app2 ] , generalizing the present approach by taking into account the small detuning @xmath432 from the raman resonance not considered in the main text
one can show that , even with a positive @xmath429 , it is possible to reach a regime where the nonlinearity coefficient @xmath386 cancels by correctly fixing the value of @xmath432 .
however , for keeping the discussion simple we will only consider here the case of a small negative @xmath429 .
the more relevant case of a small positive @xmath429 in the presence of a small detuning is presented in appendix [ app2 ] .
the main conclusions are similar in both cases .
then , for negative @xmath429 , the condition leads to @xmath433 , which corresponds to a value of the raman coupling frequency @xmath305 typical in present days experiments .
we recall however that for the system to remain in the good side of the boundary between phase iii and phase ii one needs to impose @xmath434 , which is verified by the above choice of @xmath305 , but not by a large extend .
hence , one can reach a regime where the lower excitation branch is described by gardner dynamics , but this is obtained at the expense of getting close to the phase iii phase ii boundary .
away from this boundary , the lower branch has a kdv dynamics .
in the present paper we have described how nonlinearity affects the dynamics of elementary excitations of a coherently coupled bose - einstein condensate .
excitations in the upper branch of the spectrum display a modulational instability .
as discussed in the text , this instability is stabilized by a low wave short wave resonance : the momentum imparted by the wave train formed in the upper branch to excitations in the lower branch has a stabilizing effect when it has the same sign that the velocity of the wave train .
we also showed that the system can experience second harmonic generation , and that this mechanism may be inhibited by symmetry effects ( namely by the complete separation between density and polarization modes which occurs when @xmath39 ) .
excitations in the lower branch are stable . in the long wave length limit
they are affected by nonlinear effects in a manner which can generically be described by kdv dynamics .
for some specific configuration of the system s parameters ( close to the phase ii
phase iii boundary ) one has to use gardner equation instead .
it is interesting to note that the gardner regime is realized in the lower branch which is a mode mainly corresponding to density waves : hence the wide range of nonlinear excitations of gardner s equation ( see , e.g. , ref .
@xcite ) can be generated by means of a simple scalar potential , whereas for non - coherently coupled two component condensates , where the gardner regime is obtained for a polarization mode @xcite , this can be achieved only thanks to a polarization potential @xcite .
we thank g. martone and a. recati for fruitful discussions .
amk thanks laboratoire de physique thorique et modles statistiques ( universit paris - sud , orsay ) where this work was started , for kind hospitality .
this work was supported by the french anr under grants n@xmath435 anr-11-idex-0003 - 02 ( inter - labex grant qeage ) and anr-15-ce30 - 0017 ( haralab project ) .
in this appendix we briefly explain how the solution of eq .
is obtained .
let us assume that it is of the form @xmath436 where @xmath192 is a constant ( i.e. , it depends on @xmath51 , but not on @xmath437 ) .
one first remarks that @xmath438 it follows from this result and from the ansatz that @xmath439 ^ 2 \partial_x |\widetilde\theta^{(1)}|^2 $ ] , and of course @xmath440 [ this is a direct consequence of ] . hence @xmath441
^ 2-c^2\bigg)\ , \partial_x \left|\widetilde\theta^{(1)}\right|^2 \ ; . \end{split}\ ] ] equating the r.h.s . of this expression to the r.h.s .
of determines the value of @xmath192 as given in eq . .
in this appendix we rapidly present the treatment of the lower excitation branch for a spin - orbit coupled condensate ( @xmath194 ) in the case where the system experiences a finite detuning @xmath442 from the raman resonance .
the single particle hamiltonian @xmath6 has now an additional contribution : @xmath443 .
the system is not modified , except for eq . which now reads @xmath444 the ground state value of the fields is no longer given by eq .
one has now @xmath445 where @xmath446 is a variational parameter . minimizing the energy per particle one obtains @xcite
@xmath447 the same result can be obtained in a different manner , by using dynamical arguments : one keeps @xmath446 as a free parameter , and one studies the linear excitations of the system . by demanding that the system is dynamically stable , i.e. , that the frequency of elementary excitations remains real , one obtains the result . in the presence of a finite @xmath432 ,
the ground state value @xmath448 is no longer exactly equal to @xmath449 as in and @xmath450 is not given by eq . .
instead one has [ from and ] @xmath451 and @xmath452 eq .
determines the value of @xmath448 .
depending on the system s parameters it has either four or two solutions . in the first case
, only one corresponds to the minimum of the energy per particle ( the other is the maximum ) and the system can be considered to be in the single minimum phase iii . in the second case
there are two non equivalent minima and the system is in phase ii . in the regime where @xmath453 is larger than @xmath454 and where @xmath455 , one can show that the boundary between these two regimes corresponds to @xmath456 in the case @xmath457 the solution of is @xmath33 if @xmath458 and corresponds to the standard transition line between phases ii and iii which is reproduced in figure [ fig.phases ] .
it is important to stress that in the presence of a finite detuning @xmath432 the second order phase transition from phase iii to phase ii strictly speaking disappears because the system does not cross any phase transition line when @xmath305 varies @xcite .
for instance , the velocity of sound vanishes at the transition region when @xmath457 ( cf .
eq . ) , whereas it remains finite when @xmath459 ( cf . fig .
[ fig.sound-omega ] ) . also , when @xmath459 , even in what has been identified above as the single minimum phase , the system has a small spin polarization and condensates into a state with a small but finite momentum .
the matrix @xmath335 of eq . now reads @xmath460 where , for gaining space , we have written @xmath461 and @xmath462 instead of @xmath463 and @xmath464 . in formula
the sound velocity @xmath168 is not given by : it now depends on @xmath432 .
it can be determined through the computation of the dispersion relation in the system , or more simply just by imposing the cancellation of @xmath465 . for non zero @xmath432 the ground state breaks galilean invariance and in our 1d configuration one obtains two velocities of sound , one for each direction of propagation .
the first one denoted as @xmath466 corresponds to waves propagating in the same direction as the ground state ( for which @xmath467 ) and the other ( @xmath468 ) propagates in the opposite direction .
a typical case is displayed in fig .
[ fig.sound-omega ] .
note that this figure is interrupted at low values of @xmath305 in order to prevent the system to get into phase i. as a function of @xmath305 .
the system s parameter are are @xmath306 , @xmath232 and @xmath233 .
when @xmath457 , the transition from phase iii to phase ii occurs at @xmath469 , and above @xmath470 one is in phase iii with a single sound velocity . when @xmath459 one can show that @xmath471 and @xmath472
this is the reason why the sign of @xmath432 is not specified in the figure . ]
following the procedure exposed in section [ kdvregime ] one can determine the form of the kdv equation which describes how long wavelength excitations propagating along the lower branch of the spectrum are affected by nonlinearity when @xmath459 .
in this case the nonlinear parameter @xmath386 is different from the value given by expression ( which corresponds to the @xmath457 case ) .
we do not write the explicit form of @xmath386 when @xmath459 because it is too cumbersome . instead , we rather plot @xmath386 as a function of @xmath305 for different values of @xmath432 in figure [ fig.gamma1-delta ] . as a function of @xmath432 for several values of @xmath305 .
the system s parameter are @xmath306 , @xmath232 and @xmath233 .
the curve is discontinuous when @xmath473 because in this case the system meets a first order phase transition at @xmath457 . for the present choice of parameters
@xmath474 can be canceled by changing the value of @xmath432 when @xmath475 or 2.5 , but not when @xmath65 . ] as for the sound velocity , the value of the nonlinear coefficient depends of the direction of propagation of the wave .
we denote as @xmath474 ( @xmath476 ) the value of @xmath386 corresponding to wave - trains propagating in the same ( the opposite ) direction than the momentum of the ground state .
one has the symmetry relation @xmath477 . when @xmath478 the nonlinear coefficient is discontinuous at @xmath457 , as @xmath448 and @xmath446 are , and this corresponds to the crossing of the first order transition line in the plane @xmath479 @xcite .
one sees in the figure that there exist values of @xmath432 for which the nonlinear coefficient cancels even for @xmath480 , provided @xmath305 is not too large .
it is important to notice that this cancellation of @xmath386 is obtained for a positive @xmath429 , contrarily to what occurs when @xmath457 .
note however that for @xmath65 the non - linear coefficient can not be canceled by imposing a finite value of the detuning @xmath432 .
when @xmath386 cancels , the effective nonlinear dynamics of the system is no longer described by a kdv equation , but rather by a gardner equation .
j. stenger , s. inouye , d.m .
stamper - kurn , h .- j .
miesner , a.p .
chikkatur , and w. ketterle , nature ( london ) * 396 * , 345 ( 1998 ) ; h .- j .
miesner , d. m. stamper - kurn , j. stenger , s. inouye , a. p. chikkatur , and w. ketterle phys .
82 * , 2228 ( 1999 ) ; m. r. matthews , b. p. anderson , p. c. haljan , d. s. hall , c. e. wieman , and e. a. cornell , phys .
lett . * 83 * , 2498 ( 1999 ) ; z. dutton , m. budde , c. slowe , l. vestergaard hau , science * 293 * , 663 ( 2001 ) ; n. s. ginsberg , j. brand , and l. v. hau , phys .
94 * , 040403 ( 2005 ) ; l. e. sadler , j. m. higbie , s. r. leslie , m. vengalattore , and d. m. stamper - kurn , nature ( london ) * 443 * , 312 ( 2006 ) ; k. c. wright , l. s. leslie , a. hansen , and n. p. bigelow , phys .
rev . lett . * 102 * , 030405 ( 2009 ) ; d. m. weld , p. medley , h. miyake , d. hucul , d. e. pritchard , and w. ketterle , phys .
lett . * 103 * , 245301 ( 2009 ) ; l. s. leslie , a. hansen , k. c. wright , b. m. deutsch , and n. p. bigelow , phys .
lett . * 103 * , 250401 ( 2009 ) .
p. bhi , m. f. riedel , j. hoffrogge , j. reichel , t. w. hnsch , and p. treutlein , nat . phys . * 5 * , 592 ( 2009 ) ; c. gross , t. zibold , e. nicklas , j. estve , and m. k. oberthaler , nature ( london ) * 464 * , 1165 ( 2010 ) ; m. f. riedel , p. bhi , y. li , t. w. hnsch , a. sinatra , and p. treutlein , nature ( london ) * 464 * , 1170 ( 2010 ) ; c. gross , h. strobel , e. nicklas , t. zibold , n. bar - gill , g. kurizki , and m. k. oberthaler , nature ( london ) * 480 * , 219 ( 2011 ) .
s. palzer , c. zipkes , c. sias , and m.khl , phys .
103 * , 150601 ( 2009 ) ; t. fukuhara , a. kantian , m. endres , m. cheneau , p. schau , s. hild , d. bellem , u. schollwck , t. giamarchi , c. gross , i. bloch , and s. kuhr , nat .
* 9 * , 235 ( 2013 ) .
lin , r. l. compton , a. r. perry , w. d. phillips , j. v. porto , and i. b. spielman , phys .
102 * , 130401 ( 2009 ) ; z. fu , p. wang , s. chai , l. huang , and j. zhang , phys .
a * 84 * , 043609 ( 2011 ) ; m. w. ray , e. ruokokoski , s. kandel , m. mttnen , and d. s. hall , nature * 505 * , 657 ( 2014 ) .
lin , k. jimnez - garca , and i. b. spielman , nature ( london ) * 471 * , 83 ( 2011 ) ; j .- y .
zhang , s .- c .
ji , z. chen , l. zhang , z .- d .
du , b. yan , g .- s .
pan , b. zhao , y .- j .
deng , h. zhai , s. chen , and j .- w .
pan , phys .
109 * , 115301 ( 2012 ) ; s .- c .
ji , j .- y .
zhang , l. zhang , z .- d .
du , w. zheng , y .- j .
deng , h. zhai , s. chen , and j .- w .
pan , nat .
* 10 * , 314 ( 2014 ) . c. qu , c. hamner , m. gong , c. zhang , and p. engels , phys . rev .
a * 88 * , 021604(r ) ( 2013 ) ; l. j. leblanc , m. c. beeler , k. jimnez - garca , a. r. perry , s. sugawa , r. a. williams , and i. b. spielman , new j. phys . *
15 * , 073011 ( 2013 ) .
m. a. khamehchi , y. zhang , c. hamner , t. busch , and p. engels , phys .
a * 90 * , 063624 ( 2014 ) ; s .- c .
ji , l. zhang , x .- t .
xu , z. wu , y. deng , s. chen , and j .- w .
pan , phys .
lett . * 114 * , 105301 ( 2015 ) .
m. r. matthews , d. s. hall , d. s. jin , j. r. ensher , c. e. wieman , e. a. cornell , f. dalfovo , c. minniti , and s. stringari , phys .
rev . lett . * 81 * , 243 ( 1998 ) ; d. s. hall , m. r. matthews , c. e. wieman , and e. a. cornell , phys . rev . lett .
* 81 * , 1543 ( 1998 ) .
k. c. wright , l. s. leslie , and n. p. bigelow , phys .
a * 77 * , 041601(r ) ( 2008 ) ; k. c. wright , l. s. leslie , and n. p. bigelow , phys .
a * 78 * , 053412 ( 2008 ) ; j. higbie and d. m. stamper - kurn , phys .
88 * , 090401 ( 2002 ) .
p. b. blakie , r. j. ballagh , and c. w. gardiner , j. opt .
b : quantum semiclass . opt . * 1 * , 378 ( 1999 ) ; c. p. search and p. r. berman , phys .
a * 63 * , 043612 ( 2001 ) ; p. tommasini , e. j. v. de passos , a. f. r. de toledo piza , m. s. hussein , and e. timmermans , phys .
a * 67 * , 023606 ( 2003 ) ; c. lee , w. hai , l. shi , and k. gao , phys . rev .
a * 69 * , 033611 ( 2004 ) ; m. abad and a. recati , eur .
j. d * 67 * , 148 ( 2013 ) . c. wang , c. gao , c .- m .
jian , and h. zhai , phys .
lett . * 105 * , 160403 ( 2010 ) ; t .-
ho and s. zhang , phys . rev .
107 * , 150403 ( 2011 ) ; y. zhang , l. mao , and c. zhang , phys .
* 108 * , 035302 ( 2012 ) ; w. zheng and z. li , phys .
a * 85 * , 053607 ( 2012 ) ; y. li , l. p. pitaevskii , and s. stringari , phys .
lett . * 108 * , 225301 ( 2012 ) ; g. i. martone , y. li , l. p. pitaevskii , and s. stringari , phys . rev . a * 86 * , 063621 ( 2012 ) ; y. zhang , g. chen , and c. zhang , scientific reports * 3 * , 1937 ( 2013 ) ; w. zheng , z .- q .
yu , x. cui , and h. zhai , j. phys .
b : at . mol .
. phys . * 46 * , 134007 ( 2013 ) ; y. li , g. i. martone , l. p. pitaevskii , and s. stringari , phys .
lett . * 110 * , 235302 ( 2013 ) .
d. t. son and m. a. stephanov , phys
a * 65 * , 063621 ( 2002 ) ; j. j. garca - ripoll , v. m. prez - garca , and f. sols , phys .
a * 66 * , 021602(r ) ( 2002 ) ; k. kasamatsu , m. tsubota , and m. ueda , phys .
lett . * 93 * , 250406 ( 2004 ) ; b. deconinck , p. g. kevrekidis , h. e. nistazakis , and d. j. frantzeskakis , phys .
a * 70 * , 063605 ( 2004 ) ; v. a. brazhnyi and v. v. konotop , phys .
e * 72 * , 026616 ( 2005 ) ; i. m. merhasin , b. a. malomed , and r. driben ; j. phys .
. phys . * 38 * , 877 ( 2005 ) ; k. nakamura , a. kohi , h. yamasaki , v. m. prez - garca , and v. v. konotop , epl * 80 * , 50005 ( 2007 ) .
m. merkl , a. jacob , f. e. zimmer , p. hberg , and l. santos , phys .
lett . * 104 * , 073603 ( 2010 ) ; x .- q . xu and j. h. han , phys .
lett . * 107 * , 200401 ( 2011 ) ; t. kawakami , t. mizushima , m. nitta , and k. machida , phys .
lett . * 109 * , 015301 ( 2012 ) ; o. fialko , j. brand , and u. zlicke , phys .
a * 85 * , 051605(r ) ( 2012 ) ; v. achilleos , d. j. frantzeskakis , p. g. kevrekidis , and d.e .
pelinovsky , phys .
lett . * 110 * , 264101 ( 2013 ) ; v. achilleos , j. stockhofe , p. g. kevrekidis , d. j. frantzeskakis , and p. schmelcher , epl * 103 * , 20002 ( 2013 ) ; c. polymilis , k. hizanidis , and d. j. frantzeskakis , phys .
e * 58 * , 1112 ( 1998 ) ; p.l .
christiansen , j. c. eilbeck , v .
z. enolskii , and n. a. kostov , proc .
london a * 456 * , 2263 ( 2000 ) ; c. eilbeck , v. z. enolskii , and n. a. kostov , j. math .
phys . * 41 * , 8236 ( 2000 ) ; a. m. kamchatnov and v. v. sokolov , phys . rev . a * 91 * , 043621 ( 2015 ) .
y. li , g. i. martone , and s. stringari , chapter in volume iii of the _ annual review of cold atoms and molecules _ , ed
. k. w. madison , k. bongs , l. d. carr , a. m. ray and h. zhai , p. 201
( world scientific , singapore , 2015 ) .
we use the simple property that the image space of a given matrix @xmath95 is orthogonal in the sense of the usual scalar product to the kernel of @xmath483 , to which @xmath337 belongs by .
this property is easily demonstrated in @xmath484 : let @xmath485 be a column vector in the image space of a @xmath486 matrix @xmath95 ; this means that there exists a column vector @xmath487 such that @xmath488 .
let @xmath337 be a column vector in the kernel of @xmath483 .
this means that @xmath489 .
then it is clear that the scalar product @xmath490 ( since it reads @xmath491 ) .
v. v. konotop and m. salerno phys .
a * 65 * , 021602(r ) ( 2002 ) ; a. smerzi , a. trombettoni , p. g. kevrekidis , and a. r. bishop , phys . rev . lett . * 89 * , 170402 ( 2002 ) ; b. wu , q. niu , new j. phys . * 5 * , 104 ( 2003 ) ; m. machholm , c.j .
pethick , h. smith , phys .
a * 67 * , 053613 ( 2003 ) ; m. cristiani , o. morsch , n. malossi , m. jona - lasinio , m. anderlini , e. courtade , and e. arimondo , opt .
express * 12 * , 4 ( 2004 ) ; l. fallani , l. de sarlo , j. e. lye , m. modugno , r. saers , c. fort , and m. inguscio phys .
lett . * 93 * , 140406 ( 2004 ) ; l. de sarlo , l. fallani , j. e. lye , m. modugno , r. saers , c. fort , and m. inguscio phys .
a * 72 * , 013603 ( 2005 ) . c. k. law ,
c. m. chan , p. t. leung , and m .- c .
chu , phys .
a * 63 * , 063612 ( 2001 ) ; m. a. hoefer , j. j. chang , c. hamner , and p. engels phys .
a * 84 * , 041605(r ) ( 2011 ) ; c. hamner , j. j. chang , p. engels , and m. a. hoefer phys
lett . * 106 * , 065302 ( 2011 ) .
the situation is similar to the magnetization transition in a pure ferromagnet : in the absence of external field the system has a line of first order transition which terminates with a critical point . for fixed finite external magnetic field
, one does not cross this line by varying temperature .
it can be crossed at fixed temperature by varying the external field . in our case
, the equivalent line corresponds in the @xmath479 plane to the segment @xmath457 , @xmath493 which is a line of first order phase transition ending with a critical point at @xmath457 and @xmath494 .
this line can not be crossed at fixed non zero @xmath432 by changing @xmath305 , but can be at fixed @xmath305 by changing @xmath432 . | we consider a quasi - one - dimensional two - component bose - einstein condensate subject to a coherent coupling between its components , such as realized in spin - orbit coupled condensates .
we study how nonlinearity modifies the dynamics of the elementary excitations .
the spectrum has two branches which are affected in different ways .
the upper branch experiences a modulational instability which is stabilized by a long wave - short wave resonance with the lower branch .
the lower branch is stable . in the limit of weak nonlinearity and small dispersion
it is described by a korteweg - de vries equation or by the gardner equation , depending on the value of the parameters of the system . |
the de sitter metric is an early solution of the cosmological einstein s equations which has the same degree of symmetry as the flat minkowski solution and it can be seen as a one - parameter deformation of the latter which involves a fundamental length @xmath1 . the corresponding space - time may be represented by a @xmath0-dimensional one - sheeted hyperboloid @xmath2 embedded in a minkowski ambient space @xmath3 , and the invariance group is @xmath4 , the lorentz group of the ambient space .
this group plays the same role as the poincar group on @xmath5 .
models of quantum field theory ( qft ) on the de sitter space - time have been studied very often in the past , since the symmetry properties of this universe can be used as a guideline which greatly helps in the otherwise difficult task of quantizing fields in a gravitational background ; moreover , the radius @xmath1 may also be regarded as providing a ( de sitter covariant ) _ infrared cutoff _ for minkowskian qft s , whose removal regenerates automatically poincar covariance . on the physical side ,
there has been a great revival of interest for these models at the beginning of the eighties , when the de sitter metric appeared in the inflationary cosmological scenario . according to the latter ,
the universe undergoes a de sitter phase in the very early epochs of its life @xcite .
a possible explanation of phenomena having their origin in the very early universe then relies on an interplay between space - time curvature and thermodynamics and a prominent role is played by the mechanisms of symmetry breaking and restoration in a de sitter qft . in view of the physical importance of these models , and also of the great theoretical significance of
de sitter space - time , an enormous amount of work has been done concerning quantum field theory on this space - time and many mathematical aspects of the latter have been extensively developed ( see @xcite and references therein . ) most of this work has been devoted to the construction of quantum fields satisfying a klein - gordon type equation , since even this simplest case already sets non - trivial problems when the background space - time is curved . in fact ,
qft on a curved space - time is plagued by the phenomenon of non - uniqueness of the vacuum state . in the free field case such ambiguities immediately appear since
, whenever a second quantization procedure is applied , it is based on mode expansions related to a given choice of local coordinates ( which generally cover only a part of the whole space ) . at a deeper level of investigation
the considerable relativization of the concept of energy operator , reduced to be ( on curved space - time ) the infinitesimal generator of some one - parameter group of isometries interpreted as the proper time - evolution for the chosen system of coordinates of a certain observer , makes it impossible to characterize unambiguously the vacuum state of qft as a fundamental state for the energy in the usual sense ; indeed , in general _ no true spectral condition can be satisfied by a qft on a curved space - time and this is related to the proliferation of quantizations_. in the case of de sitter space - time , where such ambiguities are present as in the general case , it has however been discovered that the _ hadamard condition _ selects a unique vacuum state for quantum fields satisfying the de sitter covariant klein - gordon equation ( see @xcite for two distinct senses in which this is true ) .
the hadamard condition postulates that the short distance behavior of the two - point function of the field should be the same for klein - gordon fields on curved space - time as for the corresponding minkowskian free field .
the fact that the hadamard condition selects a unique vacuum state for linear fields has actually been established for a wide class of space - times with bifurcate killing horizons , of which de sitter space - time is an example ( see @xcite and references therein ) ; in the de sitter case the preferred vacuum state thus selected coincides with the so called `` euclidean '' @xcite or `` bunch - davies '' @xcite vacuum state , and it singles out one vacuum in the two - parameter family of quantizations constructed in @xcite .
however , the too special character of the hadamard condition if one wishes to deal with _ general interacting fields _ leads one to search another explanation of the existence of preferred de sitter vacuum states in the global structure of this space - time manifold and , as we shall see , of the corresponding complex quadric @xmath6 ( described by eq .
( [ hyp ] ) , with @xmath7 varying in @xmath8 ) . in this connection ,
the feature of the hadamard condition which will remain of general value for de sitter qft is the requirement that the two - point function is the boundary value of an analytic function `` from the good side '' ( i.e. the so - called @xmath9-rule . )
the program which we have in mind concerning qft in de sitter space - time deals with the possibility of developping a general approach based on a set of fundamental principles , which should be completely similar to the wightman approach for minkowskian fields .
this is by no means obvious , since in the latter many important concepts are based on the fourier representation of space - time , namely the energy - momentum space .
therefore one major objective to be reached would be to dispose of a global de sitter - fourier calculus which would allow one to implement such important notions as wave propagation , `` particle states '' , etc .. however in a preliminary step , one can start from the idea that support properties in fourier space are closely connected ( via paley - wiener theorems ) to analyticity properties in the original space .
in fact , let us recall that the basic momentum space properties of minkowskian qft ( in particular the spectral condition ) can be equivalently expressed as specific analyticity properties of the @xmath10-point correlation functions @xmath11 in complex minkowski space ; moreover it is known that ( in view of the `` reconstruction theorem '' @xcite ) any general sequence of distributions @xmath12 satisfying the relevant analyticity and positivity properties determine a unique wightman field acting in the associated `` gns - vacuum '' representation . a major success of this analytic framework has been the derivation of the celebrated kllen - lehmann representation of general two - point functions which expresses the latter as ponderate superposition of `` elementary '' two - point functions satisfying a klein - gordon equation with arbitrary mass @xmath13 .
the following important facts , connected with this representation , have inspired the first steps of our approach to general qft in the de sitter universe .
\i ) _ the global holomorphic structure of the two - point function _ ( both in complex space - time variables and in complex energy - momentum variables ) is _ of maximal type _ with respect to the principles of locality , covariance and positivity of the energy spectrum ; we mean that the ( physical sheet ) holomorphy domain of the two - point function , represented by a cut - plane in the relevant lorentz - invariant variable , could not be enlarged without imposing very special conditions ( such as local commutativity at time - like separation or lacunae in the energy spectrum ) .
one can thus say that at the level of the two - point function , these general principles are completely encoded in the analyticity properties exhibited by the kllen - lehmann formula , which holds for all interacting local fields in the vacuum representation .
\ii ) being given any general two - point function @xmath14 satisfying the kllen - lehmann representation there exists a unique _ generalized free field _ which obeys the general principles and admits @xmath14 as its two - point function ( it is obtained via the gns construction by imposing that for this field all the truncated @xmath10-point functions @xmath15 , with @xmath16 , and @xmath17 vanish ) .
\iii ) as a by - product of ii ) , the free fields can be introduced by the previous construction starting from the `` elementary '' two - point functions which satisfy a klein - gordon equation .
this way of introducing the free fields presents the advantage of being independent of any particular quantization procedure ; it moreover exhibits directly the various properties of the field from those of its two - point function .
in this paper we will show that all the previous facts have their exact counterparts in the de sitter space - time thanks to the introduction of relevant analyticity domains on the complex quadric @xmath6 and to the ( related ) existence of a global fourier calculus on de sitter space - time whose primary steps are presented here .
the starting point of our axiomatic approach to de sitter qft was precisely the discovery of the existence of two - point kernels on @xmath6 enjoying maximal analyticity properties ( similar to those of minkowskian two - point functions mentioned in i ) ) and equipped with causal discontinuities of retarded and advanced type on the real de sitter space @xmath18 .
such holomorphic kernels had been introduced in the context of the theory of complex angular momentum in minkowskian qft @xcite , but the relevance of this global analytic framework for describing general two - point functions on the de sitter universe was first advocated by one the authors in @xcite , where ( as an application of the results of @xcite ) an integral representation of kllen - lehmann type was proved under a certain growth condition related to the dimension of the corresponding de sitter space - time .
this paper presents several new achievements ( partly described in @xcite ) , which open the way to a genuine axiomatic approach to de sitter qft , and can be shortly described as follows .
the results of @xcite concerning the global structure of general two - point functions are completed not only in themselves ( for example , the positivity of the weight of the kllen - lehmann representation is proven here ) but by the study of various field - theoretical aspects which they imply .
in particular , having in view a possible general approach to de sitter qft , we treat the physical interpretation of the maximal analyticity property of two - point functions from the viewpoint of the `` energy content '' of the underlying qft s ( as a substitute to the spectral condition ) .
of course , without any information on the @xmath10-point functions for @xmath16 , one can not infer a physical property of general interacting fields from the structure of the two - point function alone .
this is why it was reasonable to construct and study at first the _ generalized free fields _ , the latter being uniquely determined by the two - point functions according to the gns procedure , as in the minkowskian case ( see ii ) above ) .
moreover , such a construction presents the major interest of covering the free field case as a by - product , since the maximal analytic framework applies to all the elementary two - point functions which satisfy the de sitter - klein - gordon equation ( as in iii ) ) .
our general result can then be stated as follows : in the maximal analytic framework which we consider , the two - point functions determine generalized free fields in a _ preferred representation _ which is characterized by a _
well - defined kms condition _ formulated in appropriate regions of de sitter universe associated with geodesic observers .
the well - known occurrence of hawking s thermal effects on de sitter space - time @xcite is given here a strikingly simple geometrical interpretation ; the kms condition is actually the `` natural substitute '' to the spectral condition when maximal analyticity on the temporal geodesics remains satisfied , the linear geodesics of the flat case being here replaced by hyperbolae .
the maximal analytic framework also sheds light on the euclidean properties of the ( generalized ) free fields considered ; in fact all the two - point functions here considered define schwinger functions by restriction from @xmath6 to the sphere @xmath19 of `` imaginary times '' @xmath20 in eq .
( [ hyp ] ) ) which plays the same role as the euclidean space in complex minkowski space . in the free field case , this allows us to identify our model with the pioneering euclidean formulation of the preferred vacua given in @xcite ( and of course also coincides with the preferred vacua selected by the hadamard condition ) .
it is perhaps worthwhile to stress that properties of analytic continuation are at the basis of every treatment of de sitter field theories based on the functional integral on the euclidean sphere ; this includes the constructive approach to qft on de sitter space - time ( in this connection see @xcite ) or the application that de sitter theories may find in minkowskian constructive field theory , since the de sitter radius @xmath1 provides a natural infrared cutoff ( the euclidean space is compact ! ) . without the appropriate analyticity properties all results inferred by euclidean methods would not be relevant for the real de sitter universe .
of course in the general interacting case one can not expect maximal analyticity properties to hold for @xmath10-point functions ( @xmath16 ) and again we are left with the task of finding a suitable general approach to de sitter qft ; to be relevant this approach should assure the possibility of going to the euclidean sphere , solving there the field equations , coming back to real de sitter world ( and possibly taking the limit @xmath21 if the de sitter radius is used as a cut - off ) . at the end of this paper
we will present a set of axioms which is based on the previous ideas and which represents one possible solution to this problem . in a forthcoming paper @xcite
we will show that the euclidean sphere ( minus the coinciding points ) _ does belong _ to the extended domain of analyticity of the @xmath10-point functions satisfying these axioms .
a major contribution of this paper is the introduction of global fourier variables playing the role of mass and momentum variables , and of relevant laplace and fourier - type transformations ; in the latter the exponentials will be replaced by appropriate eigenfunctions of the laplace - beltrami operator which can be identified with _
de sitter plane waves_. the introduction of de sitter plane waves in relevant tube domains of @xmath6 , first presented by one of the author in @xcite , yields a new integral representation of the free field two - point function ( and thereby a previously unknown integral representation of legendre - gegenbauer functions in their complex domain ) : this representation , is the analogue of the standard plane - wave expansion of @xmath22 in minkowski space ; it also allows one to control in a very suggestive way the limit of de sitter qft to minkowski qft when the de sitter radius tends to infinity .
in fact it is at basis of our study of the free field in its various aspects .
the family of analytic plane waves we will introduce also suggests the definition of a suitable fourier type transformation , which maps test functions on @xmath18 to homogeneous functions on its asymptotic cone ; similar transforms had also been introduced in the context of harmonic analysis on @xmath18 in the framework of the @xmath23 representation theory @xcite , but with our definitions we can derive new mathematical results more adapted to our analytic framework based on the complex quadric @xmath24 . besides it will be seen in a further work @xcite that by applying our fourier transformation to the special case of @xmath4-invariant retarded propagators one recovers the _ laplace transform _ of the latter , which coincides with the so - called spherical laplace transform of invariant volterra kernels introduced in @xcite .
this laplace transform , which depends on a single variable interpreted as a complex mass , is the genuine analogue of the minkowskian two - point green function in the invariant mass variable : its property of being holomorphic in a half - plane and the characterization of the euclidean schwinger function by the values of this laplace transform at discrete imaginary masses ( which is a generalization of the so - called imaginary time formalism ( itf ) of thermal qft ) @xcite are recalled here in view of their importance . in @xcite , all these notions will be developed so as to produce a genuine fourier calculus on de sitter space , which could hopefully be applicable to the @xmath10-point functions of a general field theory , to investigations on the notion of particle and to the computation of feynman amplitudes . here
we make use of the fourier transform to describe the hilbert space structure of the ( generalized ) free field theories and the associated representation of the de sitter group . in particular we prove the validity of the reeh - schlieder property , which is another important aspect of the analyticity framework which we wish to stress .
the latter will be established here for de sitter ( generalized ) free fields by arguments which are very close in their spirit to the proof of the original theorem in minkowski space and differ from the recent proof of @xcite , applicable to free fields on the class of ( not necessarily analytic ) ultrastatic spacetimes .
the principle of uniqueness of analytic continuation ( applied under the refined form of the `` edge - of - the - wedge '' theorem in a complex manifold ) together with the use of the ( analytic ) fourier transformation we have introduced appear to be the main reasons for the validity of this property , and we believe that this would also hold for a large class of manifolds equipped with a holomorphic structure , provided the two - point functions considered are boundary values of holomorphic functions from relevant tuboid domains ( see appendix a ) . as a matter of fact , properties of this nature ( extended to the @xmath10-point functions ) could represent a reasonable `` microlocal''substitute to the spectral condition for a suitable class of manifolds .
one can hope that there exists a formulation of such an axiomatic approach in the framework of the analytic wave - front set theory @xcite ( where a local substitute of the analytic fourier transformation would be available ) in the spirit of the one proposed in @xcite ( for general space - time manifolds ) in the framework of the @xmath25 wave - front set theory @xcite .
our paper is organized as follows : section 2 is devoted conjointly to the presentation of general two - point functions on de sitter space - time enjoying maximal analyticity properties and to the construction and characterization of models of generalized free fields associated with such two - point functions . in
2.1 and 2.2 we fix the notations and introduce the fock representation for fields satisfying only the requirements of locality and de sitter covariance ; in 2.3 we introduce the notion of `` normal analyticity '' for general two - point functions and show that this is sufficient to recover the maximal analyticity ( perikernel property ) of @xcite ; for the associated generalized free fields this is equivalent to postulating the _ geodesic spectral condition _ or the _ geometric kms condition _ of 2.4 , supplemented by an appropriate `` antipodal condition '' when the field correlation functions are considered as given on the whole space - time @xmath18 . sections 3 and 4
continue the study of section 2 under two different respects and can therefore be read independently .
section 3 is a development of the general study of two - point functions , which is based on various results of @xcite . in
3.1 we introduce a feynman - type algebra and a substitute to wick - rotation in the space - time variables .
the laplace transform of the retarded propagators is described in 3.2 .
we devote 3.3 to describe the connection between this laplace transform at imaginary masses and the legendre expansion of the schwinger function .
this substitute to the wick rotation property in momentum space ( called in @xcite the `` froissart - gribov property '' ) is a generalized form ( invariant under the de sitter group ) of the connection between rtf and itf in the thermal representations of qft in minkowski space @xcite . in
3.4 we present an integral representation of general two - point functions satisfying appropriate growth properties at infinity in @xmath18 ; for a specially interesting class of two - point functions , it takes the form of a genuine kllen - lehmann representation .
section 4 is devoted to a detailed study of the de sitter - klein - gordon fields , which are exhibited as a class of basic elements of the general framework of section 2 .
analogues of the plane - wave solutions of the klein - gordon equation are introduced in 4.1 following @xcite for the de sitter - klein - gordon equation , and a corresponding analysis of the linear quantum field is given in 4.2 .
the use of the tubular domains of the complex hyperboloid is crucial for obtaining the relevant `` plane - wave representation '' of the two - point function , identified to a gegenbauer function of the first kind . in mathematical terms , we introduce in 4.3 a new version of the fourier transformation on the hyperboloid ( a purely mathematical study of this transformation will be developed in @xcite ) .
this transformation allows us to completely characterize in 4.4 the hilbert space of `` one - particle '' states and the corresponding irreducible unitary representations of the de sitter group .
we will also discuss in a simple and direct way the contraction of the unitary irreducible representations of the de sitter group ( giving rise to _ reducible _ representations of the poincar group @xcite ) .
in section 5 we complete our discussion of general two - point functions . in
5.1 we prove the positivity property of the kllen - lehmann representation obtained in 3.4 for the relevant class of de sitter two - point functions by using the fourier type transform introduced in 4.3 .
we then discuss in 5.2 the flat limit of de sitter gff : we show that the plane - wave representation introduced in 4.2 tends by contraction to the corresponding fourier representation of the minkowskian wightman function ; the same property is derived for de sitter general two - point functions having a kllen - lehmann decomposition .
it follows from this analysis that group - theoretical contraction is not a suitable procedure to draw consequences on the structure of the limiting quantum field theory ; in particular no problem of negative energy arises in the latter .
5.3 is devoted to show the reeh - schlieder property for de sitter gff s .
we conclude section 5 by some remarks on the qft s obtained by taking the restriction to @xmath18 of a minkowski qft in the ambient space .
in our final section 6 we outline a possible wightman - type axiomatic approach to the theory of interacting fields on the de sitter space - time , which is formulated in terms of global analyticity properties of the @xmath10-point functions of the fields .
consequences of this axiomatic approach will be discussed in a forthcoming paper @xcite . throughout our work ,
the wightman functions of the fields are described as distributional boundary values of analytic functions from relevant domains of the complex de sitter manifold , which we call `` tuboids '' @xcite ; a general survey of these mathematical notions is given in our appendix .
it is useful to describe the main properties of the de sitter space - time @xmath26 ( see eq . ( [ hyp ] ) ) in terms of the minkowskian structure of the ambient space @xmath27 , whose scalar product is denoted by @xmath28 with as usual @xmath29 @xmath26 is then equipped with a _ causal _ ordering relation ( induced by that of @xmath27 ) : let @xmath30 then , for @xmath31 the future ( resp .
past ) cone of a given point or `` event '' @xmath7 in @xmath26 will be denoted by @xmath32 ( resp.@xmath33 ) @xmath34 the `` light - cone''@xmath35,x namely the boundary set of @xmath36 is the union of all linear generatrices of @xmath26 containing the point @xmath7 ( equivalently : @xmath37 .
two events @xmath7 and @xmath38 of @xmath26 are in `` acausal relation '' , or `` space - like separated '' if @xmath39 , i.e. if @xmath40 .
the ( pseudo)-distance in the de sitter space - time is implicitely defined by @xmath41 for @xmath7 and @xmath38 time - like separated ; it is defined by @xmath42 for @xmath7 and @xmath38 space - like separated and such that @xmath43 ; it is not defined for @xmath7 and @xmath38 space - like separated and such that @xmath44 .
correspondingly , for any pair @xmath45 with @xmath46 , there exists a unique arc of hyperbola in @xmath26 in the two - plane determined by @xmath47 and @xmath38 , which is a time - like geodesic with end - points @xmath7 and @xmath38 , and @xmath48 ( defined by eq . ( [ distance ] ) ) is the proper time which elapses from @xmath38 to @xmath7 for an observer sitting on this geodesic .
when @xmath7 and @xmath38 are space - like separated , there exists a unique ( connected ) geodesic with end - points @xmath7 and @xmath38 if and only if @xmath43 ; in this case , this geodesic is an arc of ellipse in the two - plane determined by @xmath47 and @xmath38 , and @xmath48 ( defined by eq .
( [ distance1 ] ) ) is the spatial distance between @xmath7 and @xmath38 . in both cases ,
the distance is defined as an `` angle '' ( respectively hyperbolic or circular ) multiplied by @xmath1 .
the relativity group of the de sitter space - time is the pseudo - orthogonal group @xmath49 leaving invariant the scalar product ( [ product ] ) , and each of the sheets of the cone @xmath50 : @xmath51 this group is denoted by @xmath52 .
the @xmath52-invariant volume form on @xmath26 will be denoted by @xmath53 , and normalized as follows ( in leray s notations @xcite ) : @xmath54 since the group @xmath52 acts in a transitive way on @xmath26 , it will be convenient to distinguish a _ base point _
@xmath55 in @xmath26 , which will play the role of the origin in minkowski space - time ; we choose the point @xmath56 , and consider the tangent space @xmath57 to @xmath26 , namely the hyperplane @xmath58 as the @xmath0-dimensional minkowski space - time ( with pseudo - metric @xmath59 ) onto which the de sitter space - time can be contracted in the zero - curvature limit .
let then @xmath60 be the subgroup of @xmath52 which leaves the point @xmath61 invariant : @xmath60 is isomorphic to the lorentz group @xmath62 of @xmath57 and its orbits are the @xmath63-dimensional hyperboloids @xmath64 with equations @xmath65 in the corresponding hyperplanes @xmath66 . for @xmath67 ,
each hyperboloid @xmath64 represents a set of points @xmath7 in @xmath26 which are equidistant from the base point @xmath61 , namely @xmath68 where ( in view of eqs .
( [ distance ] ) and ( [ distance1 ] ) ) : for @xmath69 , @xmath70 ; for @xmath71 , @xmath72 . the fields @xmath73 which we wish to consider on @xmath18 are expected to be @xcite operator - valued distributions on @xmath18 acting in a hilbert space @xmath74 and satisfying the following property of _ local commutativity _ ( on a suitable dense domain ) :
@xmath75=0 $ ] , if @xmath76 and @xmath38 are spacelike separated in the sense of 2.1 .
moreover , there should exist a continuous unitary representation @xmath77 of the de sitter group @xmath52 acting on @xmath74 under which the field @xmath78 is transformed covariantly , i.e. @xmath79 ( we limit ourselves here to scalar bosonic fields ) and a distinguished vector @xmath80 , cyclic for the polynomial algebra of field operators and invariant under the representation @xmath77 of @xmath52 , which we call `` the vacuum '' . for a more elaborate formulation of this theoretical framework ,
one is led to consider the borchers - uhlmann @xcite algebra @xmath81 of terminating sequences of test - functions @xmath82 on @xmath18 , where the @xmath83 are @xmath25 test - functions with compact support on the cartesian product of @xmath10 copies of @xmath18 .
this @xmath84-algebra @xmath81 ( in which the involution is @xmath85 is equipped in a natural way with a representation @xmath86 of the de sitter group @xmath52 , given by @xmath87 , and there exists a `` locality '' ideal @xmath88 in @xmath89 , which is defined as in the minkowskian case ( but in terms of the de sitter space - like separation ) .
each scalar local field theory is then uniquely defined by a positive linear functional @xmath90 on @xmath81 satisfying the following two properties : \i ) @xmath52-_invariance _ : @xmath91 , @xmath92 \ii ) _ locality _ : @xmath93 , for every @xmath94 . while the functional @xmath90 is defined in terms of the quantum field @xmath73 by the sequence of its `` wightman functions '' @xmath95 n\in { \bbb n}\}$ ] , the field theory is conversely reobtained from the functional @xmath90 via the g.n.s .- type construction of a triplet @xmath96 satisfying all the required ( covariance and locality ) properties ; in the latter , the representation @xmath97 ( of @xmath98 contains the field @xmath73 through the relation @xmath99 . for every open set @xmath100 of @xmath18 the corresponding polynomial algebra @xmath101 of the field @xmath78
is then defined as the subalgebra of @xmath102 whose elements @xmath103 are such that for all @xmath104 supp@xmath105 .
the set @xmath106 is a dense subset of @xmath107 . in this framework
a _ generalized free field _ on @xmath18 can then be defined by a functional @xmath108 , such that @xmath109 and all the truncated functions @xmath110 vanish for @xmath16 .
the theory is then entirely encoded in the two - point function @xmath111 which should be a distribution on @xmath112 satisfying the following conditions : a ) _ positivity _ : @xmath113 @xmath114 b ) _ locality _ : @xmath115 _ for every space - like separated pair _
@xmath45 ; c ) _ covariance _ : @xmath116 , @xmath92 . in this case , the g.n.s .
triplet associated with the functional @xmath90 can be constructed explicitly : it is the fock representation of a field @xmath73 satisfying the following commutation relations : @xmath117 = c(x , y ) \times { \bf 1 } \label{cr}\ ] ] where the commutator function is @xmath118 the hilbert space @xmath107 of the representation can be described as the hilbertian sum @xmath119 $ ] ( with s denoting the symmetrization operation ) , in which @xmath120 and @xmath121 is defined as follows : a regular element @xmath122 of @xmath121 is a class of functions @xmath123 in @xmath124 modulo the functions @xmath125 such that @xmath126g(y)d\sigma(x)d\sigma(y)=0 $ ] ; for such element @xmath127 the corresponding norm is defined by the formula @xmath128 the full hilbert space @xmath121 is then obtained by completion ( for this norm ) of the space of regular elements .
an analogous description follows for the regular elements @xmath129 of the various subspaces @xmath130 .
each field operator @xmath131 can then be decomposed into `` creation '' and `` annihilation '' parts @xmath132 , the action of the latter on the dense subset of `` regular elements '' of the form @xmath133 being specified by @xmath134 @xmath135 @xmath136 the latter formulae imply the commutation relations ( [ cr ] ) and therefore ( in view of condition b ) on @xmath137 ) the field @xmath73 satisfies the property of local commutativity .
finally , condition c ) implies the @xmath52-covariance of the field @xmath73 , the representation @xmath77 of @xmath52 in @xmath74 ( namely the one trivially induced by @xmath86 ) being unitary in view of the same condition c ) and of formula ( [ hjk ] ) .
the previous construction of the fock representation of generalized free fields has not necessitated ( a priori ) any postulate playing the role of the positivity of the energy spectrum ; it is in fact independent of the properties of the underlying space - time and differs from the usual presentation of gff in minkowski spacetime ( see e.g. @xcite ) in which the positivity of the energy spectrum is built - in via the use of momentum - space and the choice of a spectral measure @xmath138 . in the case of de sitter space - time , where no momentum - space interpretation is given a priori
, we shall keep from the minkowskian case the idea that the _ properties of analytic continuation _ of the theory ( i.e. of the two - point function ) in the _ complexified space - time _ are directly related to the energy content of the model considered ; we shall therefore take as a postulate certain analyticity properties of @xmath139 on the complexified de sitter space - time @xmath6 which are completely similar to those implied by the usual spectral condition in the complex minkowski space - time @xmath140 . the interpretation of these properties in terms of the energy content of our models of gff on de sitter universe
will then be given in the next 2.4 ; moreover we think that , as in minkowski space , such global analyticity of the two - point functions , should also hold for _ general _ interacting fields on de sitter space - time .
let us first introduce some geometrical notions concerning the complex hyperboloid @xmath141 equivalently described as the set @xmath142 we define the following open subsets of @xmath143 : @xmath144 where t@xmath145 are the so - called forward and backward tubes with support contained in @xmath146 or in @xmath147 , obtained in connection with the spectral condition in wightman qft in @xmath27 @xcite .
] in @xmath140 .
@xmath148 ( resp .
@xmath149 ) can be characterized as the set of points @xmath150 such that @xmath151 or @xmath152 ( equivalently @xmath153 ) and @xmath154 ( resp . @xmath155 . in the same way as @xmath156 contains the `` euclidean subspace '' @xmath157 of the complex minkowski space - time @xmath140 , one easily checks that @xmath158 contains the sphere @xmath159 .
we will also call @xmath148 , @xmath149 and @xmath160 the `` forward '' and `` backward tubes '' and the `` euclidean sphere '' of @xmath143 . in order to complete the parallel with the tubes @xmath161 and @xmath162 of minkowski space ,
it is important to show that @xmath163 and @xmath164 are domains of @xmath143 from which one can take the boundary value on @xmath26 of analytic functions ( in the sense of distributions ) in a well - defined way ; the relevant geometrical notion which generalizes those of tube and local tube is the notion of _ tuboid _ in a complex manifold ( see definition a.3 , the corresponding notion of boundary value being fully characterized in theorem a.2 . ) .
the `` tubes '' @xmath148 and @xmath149 are tuboids above @xmath26 in @xmath143 , whose profiles are respectively @xmath165 and @xmath166 , where : @xmath167 @xmath168 denoting the tangent space to @xmath169 at the point x , identified here to a hyperplane of the vector space @xmath170 .
similarly the set @xmath171 is a tuboid above @xmath172 in @xmath173 , whose profile is @xmath174 .
[ prop1 ] * proof .
* for each point @xmath175 we introduce the subset @xmath176 of the tangent space to @xmath18 at the point @xmath177 : @xmath178 ( @xmath179 is a @xmath0-dimensional minkowskian subspace of @xmath3 ) .
correspondingly , we put @xmath180 ; @xmath181 is an open subset of the tangent bundle @xmath182 to @xmath18 .
we then introduce the following mapping @xmath183 from @xmath181 into @xmath143 : @xmath184 in view of eq .
( [ hypec1 ] ) it is clear that @xmath183 is a one - to - one @xmath185-mapping from @xmath181 onto the following subset @xmath186 of @xmath143 ( considered as a @xmath187-dimensional real manifold ) : @xmath188 one moreover checks that at each point @xmath189 of @xmath181 , the corresponding tangent mapping @xmath190 is of maximal rank , so that @xmath183 is in fact a diffeomorphism from @xmath181 onto @xmath186 . if one now introduces the following open and connected subset @xmath191 of @xmath181 , one easily deduces from eq .
( [ delta ] ) that @xmath192 ; since @xmath183 is a diffeomorphism , @xmath193 is an open connected subset of @xmath6 , which also entails the connectedness of @xmath163 ; the tubes @xmath163 and @xmath164 are therefore domains of @xmath6 . in order to show that @xmath163 and @xmath164 are tuboids above @xmath18 with respective profiles @xmath194 and @xmath195 ( defined by eq . ( [ profiles ] ) ) ,
one just has to notice that the _ global _ diffeomorphism @xmath183 provides ( in the neighbourhood of all points @xmath196 ) _ admissible local _ diffeomorphisms in the sense of definition a.1 ( conditions a ) and b ) of the latter are satisfied in view of eq .
( [ etre ] ) ) . properties a ) and
b ) of definition a.3 are then satisfied by @xmath148 and @xmath149 as a by - product of the equalities @xmath197 and @xmath198 ( resulting from the previous analysis ) .
the statement concerning the set @xmath199 is a direct application of the previous facts and of proposition a.2 .
+ we can now formulate the following postulate of `` normal analyticity '' for general two - point functions on de sitter space - time , which supplements the postulates a ) , b ) , c ) stated in 2.2 : d ) _ postulate of normal analyticity : + the two - point function @xmath200 is the boundary value ( in the sense of distributions according to theorem a.2 ) of a function @xmath201 which is analytic in the domain @xmath202 of @xmath203 . _ from this postulate and from the locality and covariance conditions satisfied by @xmath204 , we deduce the following properties of @xmath205 : ( property of maximal analyticity ) 1 ) @xmath201 can be analytically continued in the `` cut - domain '' @xmath206 where `` the cut '' @xmath207 is the set @xmath208 ; + 2 ) @xmath201 satisfies in @xmath209 the complex covariance condition : @xmath210 for all @xmath211 , the complexified of the group @xmath212 ; + 3 ) the `` permuted wightman function '' @xmath213 is the boundary value of @xmath201 from the domain @xmath214 .
[ maxan ] * proof .
* in view of eq .
( [ tubi1 ] ) , the tubes @xmath148 and @xmath149 are ( like @xmath215 and @xmath216 ) invariant under @xmath52 ; the covariance postulate c ) then implies the corresponding covariance property for the analytic function w in its domain @xmath199 , namely @xmath217 in fact , the boundary value on @xmath112 ( from @xmath199 ) of every derivative @xmath218 of @xmath205 with respect to any one - parameter subgroup of @xmath52 is the corresponding distribution @xmath219 , which is zero , in view of postulate c ) ; this entails ( by a refinement of the analytic continuation principle given in corollary a.1 ) that @xmath220 , and therefore proves eq .
( [ covzero ] ) .
the extension of eq .
( [ covzero ] ) to the complexified group @xmath211 then follows by analytic continuation in the group variables .
this implies that the function @xmath205 can be analytically continued in the domain @xmath221 ; the fact that this set coincides with @xmath209 is a by - product of the study of the extended tube in two vector variables @xmath222 ( in @xmath140 ) ( see e.g. @xcite , where it is shown that the section of @xmath222 by @xmath223 , namely @xmath224 , is described by the cut - plane @xmath225 in the third lorentz - invariant variable @xmath226 , where @xmath227 ) .
the relation ( [ covcom ] ) then obviously holds in @xmath209 as a consequence of ( [ covzero ] ) , with @xmath228 . in view of d ) , it is clear that the _ permuted wightman function _
@xmath229 is the boundary value of a function @xmath230 which is analytic in the tuboid @xmath231 ; then the same argument as above applied to @xmath232 shows that @xmath232 can also be analytically continued in the domain @xmath209 in which it satisfies the covariance condition @xmath233 .
now the domain @xmath209 of @xmath234 contains a real open set @xmath235 ( of @xmath236 ) in which @xmath237 .
this is the set of space - like separated points @xmath238 , where the previous equality holds as a consequence of the locality condition b ) .
it follows that the analytic functions @xmath205 and @xmath232 coincide , which implies property 3 ) and ends the proof of the proposition .
+ properties 1 ) and 2 ) of proposition [ maxan ] characterize @xmath205 as being an _ invariant perikernel _ on @xmath6 in the sense of [ bv-2 ] ( with domain @xmath239 ) . by using the transitivity of the group @xmath240 on @xmath6
, one can identify any such perikernel @xmath201 with an @xmath241-invariant function @xmath242 , @xmath61 being the base point @xmath243 introduced in 2.1 , and @xmath244 the subgroup of @xmath240 leaving invariant @xmath61 ; @xmath245 is analytic in the domain @xmath246 in view of its @xmath241-invariance property , @xmath247 can also be identified with a function @xmath248 analytic in the cut - plane @xmath249 by putting @xmath250 ( with @xmath251 ) . @xmath252 and
@xmath253 will be called reduced forms of @xmath205 . according to theorem a.2 , the postulate d ) of normal analyticity implies that the analytic function @xmath201 is locally of moderate growth , i.e.
is bounded by an inverse power of the distance from its real boundary @xmath112 .
it then follows that the analytic functions @xmath254 and @xmath255 are themselves locally of moderate growth near their ( respective ) real boundary sets @xmath18 and @xmath256 . in view of theorem a.2 ,
the function @xmath254 thus admits two distribution - like boundary values @xmath257 and @xmath258 on @xmath18 from the respective tubes @xmath259 and @xmath148 such that formally @xmath260
= @xmath261 and @xmath262 = @xmath263 . the corresponding jump of @xmath252 , namely the distribution @xmath264 ( such that @xmath265 )
, has its support contained in @xmath266 , see fig.1 ) ; in the open set @xmath267 the distributions @xmath257 and @xmath258 coincide and are identical with the restriction of the analytic function @xmath252 to @xmath268 .
the distributions @xmath257 , @xmath258 and @xmath269 are invariant under the subgroup @xmath60 of @xmath52 , but their identification with the corresponding boundary values @xmath270 , @xmath271 ( resp .
the jump c=@xmath270-@xmath271 ) of the analytic function of moderate growth @xmath272 on ( resp . across ) @xmath256 must be done carefully , namely separately on the open half - line @xmath273-\infty,1[$ ] , considered as the projection of @xmath268 , and on the open half - line @xmath2731,\infty[$ ] , considered as the projection of the interior either of @xmath274 or of @xmath275 .
the previous analysis closely follows the corresponding presentation of two - point functions in minkowski space - time ( from the viewpoint of analytic functions and of the affiliated distributions ) ; here , the transition to the @xmath60-invariant variable @xmath276 ( see 2.1 ) is the exact analogue of the passage from the vector variable @xmath277 to the lorentz - invariant variable @xmath278 in the minkowskian case .
this analysis will also be similarly completed by the following results for the green functions and for the euclidean propagators . the retarded propagator @xmath279 ( resp .
its reduced form @xmath280 .
is introduced by splitting the support of @xmath281 ( resp .
@xmath282 ) corresponding to the formal definitions @xmath283
@xmath284 @xmath285 has its support contained in the `` future cone ''
@xmath274 of @xmath18 ; its rigorous definition as an @xmath60-invariant distribution is obtained in the local coordinates @xmath286 along the same line as in the minkowskian case ( see e.g. @xcite ) ; in fact the singular character of @xmath269 at the point @xmath287 necessitates a two - step procedure for splitting its support , namely : + i ) @xmath288 is defined on a subspace of test - functions vanishing up to a certain derivation order at @xmath287 ; + ii ) @xmath288 is then continued ( up to a distribution with support at @xmath287 ) to the whole space @xmath124 by virtue of the hahn - banach theorem . the other green functions and their reduced forms
are defined in terms of @xmath289 and @xmath288 by the usual formulae , namely : + -advanced propagator : @xmath290 ( resp .
@xmath288-@xmath292 ) ; + -chronological propagator : @xmath293 ( resp .
@xmath294 = @xmath295+@xmath258 ) ; + -antichronological propagator : @xmath296 ( resp .
@xmath297 = @xmath257+@xmath298 ) . + as for the distributions @xmath299 and c considered on the whole real line ( in @xmath300 ) , they are respectively the one - variable representatives of @xmath60-invariant distributions @xmath301 , @xmath294 and @xmath301- @xmath294= @xmath302 ) , the latter being formally equal to @xmath303 ; however this calls for the following _ although apparently absent from the definition of @xmath299 and c ( directly obtained from boundary values of the analytic function w(@xmath300 ) ) , the problem splitting the support ( necessitating a two - step procedure of the type described above ) reappears in the reconstruction of @xmath294 and @xmath297 as distributions on @xmath18 from their one - variable representatives @xmath304 .
_ _ as in the minkowskian space - time , the two - point function of a hermitian field satisfies the identity @xmath305 .
+ it follows that the function @xmath201 is real analytic , i.e. for @xmath306 @xmath307 , and that the green functions @xmath279 and @xmath308 are real valued . _ _ the euclidean propagators or schwinger functions _ it is easily seen that the intersection of the domain @xmath209 with the `` euclidean subset '' @xmath309 of @xmath310 is the set @xmath311 , where @xmath183 is the set of `` coinciding points '' , namely @xmath312 : in fact , if @xmath313 , one has @xmath314 , which implies that @xmath315 iff @xmath316 .
it thus follows from proposition 2.2 that the `` euclidean propagator '' @xmath317 is defined as an analytic function in @xmath318 ; moreover , since @xmath201 is of moderate growth with respect to the distance of @xmath319 from @xmath320 ( due to the distribution character of the boundary values of w ) , s is itself of moderate growth on @xmath318 with respect to the distance of @xmath319 from @xmath321 .
as in the minkowskian case , a standard argument based on the construction of a suitable primitive of s , shows that s can be extended as a distribution to the whole euclidean space @xmath322 .
s can be seen as a distribution kernel on @xmath19 , with singular support on the diagonal , which is moreover invariant under the subgroup @xmath323 of @xmath240 , in view of the @xmath240-invariance of w. it follows that s admits the reduced forms @xmath324 and @xmath325 defined by @xmath326 , @xmath327 for @xmath328 ; @xmath329 is in fact a distribution on @xmath19 with singular support at @xmath61 , invariant under the stabilizer of @xmath330 , and is also obtained by taking the restriction of @xmath252 to @xmath19 ; similarly one can say that @xmath331 . in this subsection we shall study the gff s
whose two - point functions satisfy the general postulates a ) , b ) , c ) , d ) ( this is what is meant when we will say that these gff satisfy a ) , b ) , c ) , d ) ) .
as seen in 2.2 a ) , b ) , c ) imply that they are scalar local fields acting in a hilbert space . in order to investigate the enrgy content of such fields , which is encoded in d ) , we shall adopt the viewpoint of an observer moving on the geodesic of the point @xmath61 contained in the @xmath332-plane , namely the branch of hyperbola @xmath333 where @xmath334 .
the set of all events of @xmath335 which can be connected with the observer by the reception and the emission of light - signals is the region : @xmath336 this region admits two boundary parts @xmath337 and @xmath338 , respectively called the `` future '' and `` past horizons '' of the observer with geodesic @xmath339 , namely @xmath340 since the parameter @xmath341 of the representation ( [ geod ] ) is interpreted as the proper time of the observer sitting on the geodesic @xmath339 ( see ( 2.1 ) ) , one can call `` time - translation group relative to @xmath339 '' the one - parameter subgroup @xmath342 of @xmath52 ( isomorphic to the one - parameter lorentz subgroup of @xmath4 ) whose transformations are hyperbolic rotations parallel to the @xmath332-plane .
the action of @xmath342 on the domain @xmath343 is defined below .
let @xmath344 be any point in @xmath345 represented as follows : @xmath346 for @xmath347 , one puts @xmath348 = x(t+\tau,\vec{\underline{x}})\equiv x^{t } , \label{old}\ ] ] ( with @xmath349 in this notation ) .
@xmath342 thus defines a group of isometric automorphisms of @xmath343 whose orbits are all branches of hyperbolae of @xmath343 in two - dimensional plane sections parallel to the @xmath332-plane ( see @xcite for a general discussion of this kind of structure ) .
note however , that in this family of orbits of @xmath342 , @xmath339 is the only one which is a geodesic of @xmath18 , so that the time - translation interpretation of the group @xmath342 is , strictly speaking , only relevant for observers sitting on @xmath339 ( or in a neighbourhood of @xmath339 which is small with respect to the radius @xmath1 of @xmath18 ) .
one can also consider the action of the group @xmath342 on the horizons @xmath350 which are foliated by linear orbits of this group ( these orbits are the light rays on @xmath18 which are parallels to the asymptots of @xmath339 ) .
we can now establish the following property of the generalized free fields on @xmath18 : ( geometrical kms condition ) + for every pair of bounded regions @xmath351 , @xmath352 of @xmath343 , all the ( two - point ) correlation functions between elements of the corresponding polynomial algebras @xmath353 , @xmath354 of a generalized free field on @xmath18 satisfying the previous postulates a ) , b ) , c ) , d ) enjoy a kms condition with respect to the time - translation group @xmath342 whose temperature parameter is equal to @xmath355 .
[ kms ] * proof .
* we first consider the correlation functions with respect to @xmath356 between one - component elements @xmath357 of @xmath358 , @xmath359 , namely @xmath360 where we have put @xmath361{(x)}=f(x^{-t})$ ] . in view of eq .
( [ old ] ) , we can then write @xmath362 which defines @xmath363 as a @xmath185 function of @xmath341 .
let us now consider , for every point @xmath364 @xmath365 in @xmath366 the corresponding ( complex ) curve @xmath367 of @xmath368 defined as follows : @xmath369 where @xmath370 denotes the complexified form of the parametric representation ( [ parahor ] ) .
the curve @xmath367 is a complex hyperbola whose representation enjoys the following properties ( easily checked from eq .
( [ parahor ] ) ) : + i ) ( @xmath371)-periodicity along the imaginary axis of the @xmath341-plane ( since @xmath372 + ii ) for each @xmath341 in the strip @xmath373 ( resp .
@xmath374 ) the corresponding point @xmath375 belongs to the domain @xmath148 ( resp .
@xmath149 ) of @xmath143 ; + iii ) for each @xmath341 such that @xmath376 the point @xmath377 is real and such that @xmath378 . for all such points
one then checks that @xmath379 ; + iv ) if @xmath380 then there exists a finite interval @xmath381 containing the origin in the real @xmath341-axis on which @xmath382 remains negative . from these geometrical properties and from the analyticity properties of @xmath201 ,
one then deduces that the restriction of w to the complex curve @xmath367 defines a @xmath383-periodic analytic function of @xmath341 , namely @xmath384 whose domain is the periodic cut - plane @xmath385 it must be noted that the _ extended domain _ @xmath209 of @xmath205 is used here , since the sets @xmath386 and @xmath289 ( whose traces in @xmath367 yield respectively @xmath387 , @xmath388 and @xmath389 ) belong to the _ boundaries _ of the primitive `` tubes of normal analyticity '' @xmath199 and @xmath231 .
however these sets are contained in the extended domain @xmath209 .
moreover it is easily checked that the boundary values @xmath390 of @xmath391 on @xmath256 ( defined as distributions in the variable @xmath341 , for each fixed values of @xmath392 ) represent the restriction to @xmath393 of the corresponding two - point wightman distributions , namely @xmath394 therefore the @xmath395 periodicity of the function @xmath391 induced by the geometry of the complex hyperbola @xmath367 implies in turn the following k.m.s .- type condition : @xmath396 since all these properties of the function @xmath391 hold for any choice of the point @xmath397 in @xmath366 , one easily obtains the corresponding result for the integral @xmath398 the latter is analytic and periodic in the cut plane @xmath399 and its boundary values @xmath400 and @xmath401 satisfy the corresponding kms relations : @xmath402 one notices that the cut - plane @xmath403 is connected if @xmath351 and @xmath404 are space - like separated ( since then @xmath405 contains the point @xmath406 ) . in the general case one
considers correlation functions between arbitrary elements @xmath407 of @xmath408 ( i=1,2 ) ; this can easily be seen to reduce to the previous one - component case ; in fact any general two - point function @xmath409 can always be decomposed ( via the standard contraction formalism of fock space ) as a polynomial of elementary two - point functions @xmath410 ( integrated with the appropriate test functions of the variables @xmath411 ) .
since the previously described analytic structure ( including @xmath395-periodicity and kms condition of the boundary values ) in the @xmath341-plane is stable under multiplication and integration over the real parameters @xmath412 , one concludes that there exists an analytic function @xmath413 , analytic and periodic in @xmath414 and such that @xmath415 @xmath416 + it must be noted that in the course of the proof of this kms property , the following additional property , relating by analytic continuation the region @xmath417 with the antipodal region @xmath418 @xmath419 has been obtained as a by - product ( see the property iii ) of @xmath420 in this proof ) .
( antipodal condition ) + under the condition that all the test functions have their support contained in @xmath417 , the following identities hold @xmath421 @xmath422 and more generally @xmath423 @xmath424 where the sequence @xmath425 is localized in the antipodal region @xmath426 .
[ prantipod ] we shall now introduce an `` energy operator '' @xmath427 associated with the geodesic @xmath339 by considering in @xmath107 the continuous unitary representation @xmath428 t\in { { \bbb r}}\}$ ] of the time - translation group @xmath342 and its spectral resolution @xmath429 this defines ( on a certain dense domain of @xmath107 ) the unbounded operator @xmath430 for any pair of one - component vector states @xmath431 , @xmath432 , the corresponding correlation function given in ( [ mamma ] ) can be written as follows : @xmath433 which shows that @xmath400 is a continuous and bounded function . in view of eq .
( [ unitario ] ) it can be expressed as the fourier transform of the bounded measure @xmath434 we also obtain similarly that @xmath401 is the fourier transform of @xmath435 eqs .
( [ tutor ] ) and ( [ dicendo ] ) are valid for arbitrary test - functions @xmath436 and @xmath437 in @xmath124 .
now , if @xmath436 and @xmath437 belong to @xmath438 , the functions @xmath400 and @xmath401 satisfy the kms relations ( [ gms ] ) , and their fourier transforms satisfy ( as bounded measures ) the following relation , which is equivalent to ( [ gms ] ) : @xmath439 since the kms relations hold similarly for the correlation functions between arbitrary vector states @xmath440 , @xmath441 of d ( see the end of the proof of proposition 2.3 . )
we could reproduce completely the previous argument and computations in the general case and therefore obtain the following ( geodesic spectral condition ) + for every pair of states @xmath440 , @xmath441 in @xmath442 , the corresponding matrix elements of the spectral measure @xmath443 satisfy the following relation @xmath444 [ pr4 ] _ this geodesic spectral condition gives a precise content to the statement that in the region @xmath343 corresponding to an observer living on the geodesic @xmath339 , the energy measurements ( relative to this observer ) give exponentially damped expectation values in the range of negative energies . in the limit of flat space - time the l.h.s .
( [ tyty ] ) would be equal to zero for @xmath445 , which corresponds to recovering the usual spectral condition of `` positivity of the energy '' .
_ _ a statement similar to proposition [ pr4 ] making use of the same unitary group @xmath446 would hold for vector states @xmath447 @xmath448 in @xmath449 , i.e. states localized in the antipodal region of @xmath450 .
however , as far as the corresponding correlation functions are concerned , the kms analyticity strip would be replaced by @xmath451 and eqs .
( [ gms ] ) by the following ones : @xmath452 whose equivalent form for the fourier transforms is @xmath453 in writing a geodesic spectral condition of the form ( [ tyty ] ) for such states in @xmath454 the factor @xmath455 ( at the r.h.s . )
should therefore be replaced by @xmath456 , which would correspond to an exponential damping of @xmath457 in the range of positive @xmath90 s .
it must however be remarked that for an observer sitting on the geodesic @xmath458 ( antipodal to @xmath459 ) the natural time and energy variables should be respectively equal to @xmath460 and @xmath461 , since the corresponding time - translation group @xmath462 ( obtained from @xmath342 e.g. by a conjugation of the form @xmath463 , @xmath464 being a rotation of angle @xmath465 in a plane orthogonal to the @xmath466-axis ) is such that @xmath467 .
_ we can also reexpress the antipodal condition of proposition [ prantipod ] under the following equivalent form , which asserts that the spectral measure @xmath468 has exponentially damped matrix elements between states localized in the mutually antipodal regions @xmath417 and @xmath426 : ( antipodal spectral condition ) . + the following relation holds for every pair of vector states @xmath440 , @xmath441 @xmath469 @xmath470 @xmath471 @xmath472 being in @xmath454 .
[ prantipod2 ] * proof .
* the proof is obtained by writing that the fourier transform of @xmath473 ( given by eq .
( [ antipod2 ] ) ) is equal to the fourier transform of @xmath474 ( given by eq .
( [ kmsvec1 ] ) ) multiplied by @xmath475 .
+ a sufficient condition for a generalized free field satisfying the postulates a),b),c ) to enjoy the full analytic structure entailed by the normal analyticity postulate d ) is that it satisfy either + e ) the geometrical kms condition together with the antipodal condition , + or + e ) the geodesic spectral condition together with the antipodal spectral condition .
[ prpostulates ] * proof .
* it is sufficient to show that either eqs .
( [ gms ] ) or eq .
( [ byby ] ) expressing equivalently the kms condition or the geodesic spectral condition for arbitrary one - component vector states @xmath476 and @xmath477 in @xmath478 imply the analyticity properties of @xmath201 described in 2.3 . according to the analysis given in the proof of proposition [ kms ] eqs .
( [ gms ] ) exactly express the fact that on the open set @xmath479 the distributions @xmath480 and @xmath481= @xmath482 are the boundary values of an analytic function of _ one _ complex variable ( depending on @xmath483 real variables ) @xmath484 defined in the union of the domains of the complex hyperbolae @xmath420 ( with @xmath485 varying in @xmath343 ) represented by @xmath486 ( see eq .
( [ ccut ] ) ) .
it is then clear that this information is sufficient ( and even redundant ) for constructing an invariant perikernel @xmath201 through its reduced form w(@xmath300 ) ( see 2.3 ) .
in fact by considering the special hyperbola @xmath487 ( i.e. @xmath488 , @xmath489 we obtain the following identification @xmath490 which defines the analytic function w(@xmath300 ) in the cut - plane @xmath491 for arbitrary points @xmath485 in @xmath343 represented by the parametrization ( [ parahor ] ) one would obtain more generally @xmath492 the consistency of this set of relations being ensured by the @xmath52-covariance of the boundary values of @xmath484 ( for @xmath493 ) , namely @xmath494=@xmath495 with the specification @xmath496 .
the reduced form w(@xmath300 ) thus obtained defines a unique invariant perikernel @xmath201=@xmath497 , whose boundary values from the normal analyticity domains coincide with the given distributions @xmath480 and @xmath481 on @xmath479 ( by construction , in view of the proof of proposition [ kms ] ) .
the fact that the coincidence relations extend to the whole space @xmath112 necessitates the use of the antipodal condition ( or of the equivalent antipodal spectral condition ) since the latter allows one to identify the ( coinciding ) boundary values of @xmath201 on @xmath498 with the given distribution @xmath480 = @xmath481 in that region .
in fact one notes that @xmath380
if @xmath499 @xmath500 while @xmath501 if @xmath499 @xmath502 . therefore , the right identification of the boundary values of w on @xmath479 _ and _ @xmath498 determines by @xmath52-covariance its right identification on the whole space @xmath112 , since these two subsets fix the right boundary values of the reduced form w(@xmath300 ) of w on the whole real @xmath300-axis ( from both half - planes @xmath503 and @xmath504 and with the coincidence region @xmath505 ) , i.e. on the image in @xmath506 of the whole product @xmath112 .
+ _ the fact that the perikernel @xmath201 is entirely determined by the knowledge of @xmath480 and @xmath481 in the region @xmath479 ( the antipodal condition appearing as a self - consistency condition ) suggests the following alternative presentation of proposition [ prpostulates ] .
_ let us consider a generalized free field theory whose wightman distributions are given _ only _ on the classically accessible region @xmath343 of an observer living on the geodesic @xmath339 of @xmath18 and satisfy the postulates a),b),c ) ( restricted to @xmath343 ) together with the corresponding geometrical kms condition ( or equivalently the geodesical spectral condition ) .
then this theory can be extended in a unique way to a generalized free field theory on the whole de sitter space - time which enjoys the full analytic structure entailed by the postulate d ) , and satisfies ( as a by - product ) the antipodal condition .
_ in this connection we mention the interesting approach to two - dimensional constructive de sitter qft s given in @xcite , where wightman functions are introduced ( with space - time cutoffs ) only in the region @xmath343 .
the previous statements clarify the situation concerning the extension of those theories satisfying suitable properties to the whole de sitter space - time .
in particular , in the treatment of klein - gordon fields of @xcite the use of space cut - offs which breaks the de sitter invariance leaves somewhat implicit the @xmath507-invariant limiting theory .
our results show that these limiting theories fit within the whole analytic structure described in this section , and can therefore be identified with the klein - gordon fields of 4.2 .
_ _ all the features that have been discussed in this section are also naturally interpreted in terms of the existence of an antiunitary involution @xmath508 relating the algebras @xmath509 and @xmath510 and the validity of the corresponding bisognano - wichmann duality theorem for the von neumann algebras @xmath511 and @xmath512 @xcite .
we will not further dwell on these facts here .
a complete discussion of this class of results will be given in a following paper @xcite in which we will deal with the general properties of ( free or interacting ) de sitter quantum field theories . _
the previously described analytic structure of two - point functions in complex de sitter space - time ( see 2.3 ) is the exact analogue of the standard analytic structure of general two - point functions in complex minkowski space - time implied by the wightman axioms . as in the latter case ,
the complex - space time function w@xmath319 presents two important aspects in agreement with the common wisdom of qft , namely the euclidean propagator s ( here a kernel on the sphere @xmath19 ) and the retarded propagator @xmath289 ( called in @xcite a `` volterra '' or `` causal kernel '' on @xmath18 ) . any general two - point function in the complexified de sitter space - time
is thus characterized in a suggestive way as a triplet ( w , s,@xmath289 ) .
the function @xmath201 ( or `` perikernel '' ) is analytic in the cut domain @xmath513 ( see proposition 2.2 ) ; @xmath514 is the restriction of w to @xmath515 extended to the set of coinciding points @xmath516 ; @xmath279 is a kernel on @xmath18 with support @xmath517 obtained by taking the discontinuity of w across the real trace @xmath518 of @xmath320 and restricting the latter to the future component @xmath519 .
the algebraic properties of the class of triplets ( w , s,@xmath289 ) which have been derived in particular in @xcite ( see also @xcite ) can then immediately be interpreted in the usual language of feynman diagrams and provide in fact a substitute to the standard wick - rotation procedure when one deals with de sitterian propagators .
we shall now describe these results .
= 12.cm we shall consider a class of self - energy diagrams @xmath520 whose vertices
are labelled by de sitter space - time variables @xmath521 and whose lines @xmath522 represent general two - point functions @xmath523 ( namely also the associated triplets ( @xmath524 , @xmath525 , @xmath526 ) ) ; this class is the one generated by the two operations described below ( bubble insertion and vertex convolution , see fig.[fig1]a ) .
our results , obtained by an iterated application of the properties listed below , can be stated as follows : with every diagram @xmath520 of this class , it is possible to associate a two - point function @xmath527 namely a triplet ( @xmath527 , @xmath528 , @xmath529 ) whose euclidean propagator @xmath528 , is obtained in terms of the various euclidean propagators @xmath525 involved in the diagram by the same integration rule as in the minkowskian space ( the sphere @xmath19 playing now the role of the euclidean space ) . + i ) _ bubble insertion _ : with any diagram @xmath530 ( represented on fig.[fig1]b ) is associated the two - point function @xmath531 whose analyticity domain is @xmath209 and whose corresponding triplet belong to the considered class .
+ ii ) _ vertex convolution _ :
this operation , represented by the elementary diagram of fig [ fig1]c ) , corresponds to the @xmath532-composition product of two perikernels introduced in @xcite .
let ( @xmath533 @xmath534 , @xmath535 ) , @xmath359 , be the triplets associated with two given two - point functions @xmath536 , @xmath537 ; there exists a two - point function w , denoted by @xmath538 , whose associated triplet ( @xmath205 @xmath539 , @xmath289 ) , enjoys the following properties : 1 .
@xmath540 is the convolution of euclidean propagators , defined as the standard composition of kernels on the sphere @xmath19 .
@xmath541 2 .
@xmath542 is the convolution of retarded propagators , introduced in @xcite as the composition of volterra kernels on @xmath18 by the following formula : @xmath543 where @xmath544 is the `` double cone '' with vertices @xmath545 and @xmath53 is the volume form on @xmath18 defined by eq .
( [ leray ] ) .
+ 3 . for @xmath319
varying in ( a dense subdomain of ) @xmath209 , there exists a compact integration cycle @xmath546 , obtained from @xmath19 by continuous distortion , such that @xmath547 in eq .
( [ cp1 ] ) and ( [ cp3 ] ) @xmath548 denotes the complexified form of @xmath53 , i.e. a @xmath549-invariant @xmath0-form on @xmath6
. the following comment can be added concerning this result .
the geometrical argument of contour distortion ( given in its complete form in @xcite and sketched in @xcite ) is of a different nature from the standard wick - rotation argument of the flat case ; the fact that only _ compact _ cycles are used in the present case ( namely in eqs .
( [ cp1])-([cp3 ] ) ) indicates that the chronological propagators can not be involved in the argument , the wick rotation being now unapplicable between the non - compact real space - time @xmath18 and the compact euclidean sphere @xmath19 . on the contrary , the connection between the euclidean and the retarded convolutions @xmath550 and @xmath551 relies on a more general distortion procedure , which remains also valid in the flat minkowskian limit ( up to the non - compactness of the euclidean integration in that case ) . _
the analyticity domains @xmath552 of the perikernels considered in @xcite involved the cut @xmath553 with @xmath554 , thus giving a weaker version of the previous results since @xmath555 .
the case @xmath556 which is relevant for us necessitates some distribution - theoretical refinement due to the singularities at coinciding points but can be treated along the lines indicated in @xcite ; these technical aspects do not spoil the purely geometrical essence of the argument .
_ the causality and de sitter covariance properties of the retarded propagators @xmath279 ( see 2.3 ) allow the introduction of a laplace - type transform of the latter enjoying analyticity properties with respect to the relevant conjugate variable @xmath557 , proportional to a complex mass .
we shall obtain in this way an analogue of the analyticity domain of the minkowskian propagators in the mass variable ; however , as in that case it is necessary to assume that the propagators satisfy suitable growth conditions at infinity on @xmath18 ( i.e. on their support @xmath558 , which will be specified below .
the retarded propagators are invariant volterra kernels to which the methods of @xcite can be applied . according to the latter ,
the laplace transform @xmath559 of @xmath279 can be defined by the following formula , @xmath560 being the laplace transform of @xmath561 introduced in @xcite .
] : @xmath562 where @xmath563 and @xmath564 ; @xmath565 has its support contained in @xmath566 .
@xmath567 denotes a generalized legendre function of the second kind @xcite , ( proportional to a gegenbauer function of the second kind @xcite admitting the following representation : a ) for @xmath0 even : @xmath568 @xmath569^{-\frac{1}{2}}dw \label{gege2}\ ] ] b ) for @xmath0 odd : @xmath570 in these expressions @xmath571 denotes the following polynomial : @xmath572 eq .
( [ gege2 ] ) ( resp . ( [ gege3 ] ) ) defines a function @xmath573 analytic in the half plane @xmath574 ( resp .
the whole plane ) .
it can then be shown that , if the behaviour of @xmath575 at infinity is governed by @xmath576 then the product @xmath577 is analytic in the half - plane @xmath578 and admits a continuous boundary value on the line @xmath579 . as a matter of fact
, the expression ( [ 34 ] ) has been obtained a posteriori in @xcite ; at first , the laplace transformation was introduced by a more geometrical method ( initiated in @xcite ) which exhibits the function @xmath559 as the one - dimensional laplace transform of a radon - type transform @xmath580 of @xmath581 . in this transformation ,
the integration is taken on the following family of plane sections of @xmath582 called `` horocycles '' : @xmath583 .
horocycles are parametrized in the following way : @xmath584 for each @xmath341 real , @xmath585 is a paraboloid ( parametrized by the vector @xmath586 ) whose intersection with the support of @xmath285 , namely the `` cone '' @xmath274 , is either empty ( if @xmath587 ) or equal to the ball @xmath588 represented by @xmath589 one then defines : @xmath590 as proven in @xcite ( eq . (
3.20 ) , corollary of proposition 13 and eq .
( 4.4 ) ) there holds the following \i ) @xmath591 can be expressed by the following abel - type tranformation @xmath592^{\frac{d-3}{2 } } \
\sinh v dv .
\label{llap1}\ ] ] ii ) if @xmath593 with @xmath594 then @xmath595 .
\iii ) under the assumption of ii ) the transform @xmath559 of @xmath464 is defined in the half - plane @xmath596 as the following one - dimensional laplace transform of @xmath597 : @xmath598 the similarity of the previous geometrical definition of @xmath559 with that of the laplace transform of the minkowskian retarded propagator as a function of the complex mass variable can be exhibited as follows .
if @xmath599 denotes the minkowski propagator in dimension @xmath0 ( with @xmath600 , @xmath601 ) , its laplace transform can be written as a function of the mass variable as follows : @xmath602 or , by introducing the family of space - like hyperplanes @xmath603 : @xmath604 where @xmath605 denotes the following radon transform of @xmath606 : @xmath607 formulae ( [ lap2 ] ) and ( [ rad1 ] ) are the respective analogues of ( [ llap2 ] ) and ( [ lapla]),([llap1 ] ) , and they define @xmath608 ( under the usual temperateness assumption on @xmath609 ) as an analytic function in the half plane @xmath610 .
moreover , @xmath611 is expressed in terms of @xmath464 by the following integral transformation , ( obtained by inserting the r.h.s . of eq .
( [ rad1 ] ) in the integral of eq .
( [ lap2 ] ) and inverting the order of the integrations ) : @xmath612 in this formula , which is the analogous of eq .
( [ 34 ] ) , the kernel @xmath613 is a hankel - type function defined for @xmath614 by the integral representation @xmath615 @xmath616 being the function introduced in 6.15 of @xcite we shall now state the de sitterian analogue of the following well - known property of the minkowskian propagator in momentum - space : the restriction of the latter to imaginary values of the masses is the fourier transform of the euclidean propagator @xmath617 , with @xmath618 , namely @xmath619 since in the de sitter case , @xmath620 is defined on the euclidean sphere @xmath621,one can only associate with it a _ discrete _ fourier transform , namely the sequence of coefficients @xmath622 of its fourier - legendre expansion : @xmath623 given by : @xmath624 under the assumptions of proposition 3.1 the laplace transform @xmath625 of the retarded propagator @xmath626 and the fourier - legendre coefficients @xmath627 of the euclidean propagator @xmath628 are related by the following relations : integer the case @xmath629 requires some additional regularity assumption on @xmath630 on the line @xmath631 ( theorem 3 of @xcite ) ] @xmath632 moreover @xmath625 is the ( unique ) carlsonian interpolation of the sequence of values ( [ huhuhu ] ) in the complex half - plane @xmath633 the proof of this result given in @xcite consists in first showing the property for the one - dimensional fourier case ( i.e. @xmath634 by a simple contour - distortion argument , and then proving that the general @xmath635-dimensional case reduces to the previous one by applying the radon - abel transformation @xmath636 described in 3.2 . _ one can also give a direct proof based on a contour distortion argument in the @xmath637 plane relating the formulae ( [ 34 ] ) and ( [ huhuhu ] ) thanks to the properties of the functions @xmath638 and @xmath639 ; an argument of this type was in fact produced ( for @xmath640 in @xcite and @xcite in the context of complex angular momentum analysis of scattering amplitudes in s - matrix theory , involving a somewhat similar mathematical situation ; this is why we shall call `` property fg '' the property stated in proposition 3.2 .
_ _ as it will be suggested by the detailed study of the de sitterian linear fields given in section 4 , the most significant assumptions on @xmath641 should correspond to the choice @xmath642 for the behaviour at infinity and to the inclusion of a distribution with singular support at @xmath643 and maximal order @xmath644 if @xmath645 therefore , strictly speaking , it is only in the case @xmath646 that such requirements allow to apply the statements of propositions 3.1 and 3.2 in a fully satisfactory way .
however , it is possible to extend the proofs of these statements given in @xmath647}$ ] , with a small amount of technicalities , so as to include the relevant requirements for @xmath648 this will be presented elsewhere .
_ the kms - type interpretation of the analyticity of the two - point functions in complex temporal geodesics ( see 2.4 ) is accompanied ( as in the usual matsubara formalism : see e.g. @xcite and references therein ) by the phenomenon of quantization of the associated imaginary energies ( this is the so - called imaginary - time - formalism ( itf ) ) .
the fg property is actually a phenomenon of similar mathematical nature ( in fact , _ the same _
phenomenon for @xmath649 see e.g. @xcite ) . however , in contrast with the previous geodesical thermal property , the fg property is _ global _
( i.e. coordinate independent ) on de sitter space - time ; in fact , as it will be seen below , the imaginary mass variable which is quantized in the euclidean formalism is actually a de sitter - invariant quantity . in order to complete the description of the double analytic structure of the de sitter two - point functions ( namely , analyticity in space - time variables on @xmath650 and analyticity in the mass variable @xmath651 for the relevant laplace transform )
, we must also give the inversion formulae which allow one to represent all the elements of a given triplet @xmath652 in terms of the corresponding laplace transform @xmath653 the following statement has been established in @xmath654}$ ] for triplets of moderate growth @xmath655 whose associated laplace transform @xmath656 is analytic be analytic in this half - plane . ] in the half - plane @xmath657 and satisfies appropriate regularity conditions on the boundary line @xmath658 the following inversion formulae hold : @xmath659 @xmath660 d\nu , \label{a320}\ ] ] @xmath661@xmath662 the functions @xmath663 are generalized legendre functions of the first kind @xcite ( proportional to gegenbauer functions of the first kind @xcite ) given by the following integral representation ( valid for @xmath664-\infty , -1]$ ] ) : @xmath665 [ 2(\cos\tau-\cos\theta)]^{\frac{d-3}{2}}d\tau \label{legendre1}\ ] ] formulae ( [ a320 ] ) and ( [ a321 ] ) apply respectively to the following ranges : @xmath666 and @xmath667 ; moreover , for @xmath668 ( i.e. @xmath669 real ) , eq .
( [ a321 ] ) yields the schwinger function @xmath670 , equivalently expressed by the fourier - legendre expansion ( [ a318 ] ) , with @xmath671 for @xmath672 . these formulae can be considered as _ complex mass representations _
valid for all two - point functions @xmath673 ( resp .
retarded propagators @xmath674 of de sitter qft s satisfying the general analytic structure of perikernels and moderate growth conditions ( in the sense of proposition 3.1 ) .
since the functions @xmath675 represent a family of special perikernels @xcite on @xmath18 , eq .
( [ a321 ] ) appears as a decomposition of general two - point functions on a basis of elementary two - point functions on @xmath18 admitting the same type of growth ( governed by @xmath676 ) . by now making use of the hermiticity of the underlying fields
, one can show the following result .
if the laplace transform @xmath677 of the retarded propagator @xmath678 of a de sitter qft is analytic in the half - plane @xmath679 then @xmath678 and the associated complex space - time two - point function @xmath680 satisfy the following kllen - lehmann - type representations : @xmath681 @xmath682 { d } \nu , \label{a324a}\ ] ] @xmath683 * proof .
* by writing eq.([a320 ] for the case @xmath642 and by taking into account that @xmath684 is an odd function of @xmath557 ( see eqs .
( [ gege4 ] ) and ( [ legendre1 ] ) ) , one readily obtains directly eq .
( [ a325 ] ) , with @xmath685 \label{a326}\ ] ] eq.([a324a ] ) is obtained similarly from eq.([a320 ] ) or from eq .
( [ a325 ] ) by using a discontinuity formula for @xmath686 across the line @xmath687 ( @xcite , eqs.(3.104 ) , ( 3.105 ) ) , since @xmath688 .
as in the minkowskian case , the hermiticity of the field @xmath689 implies that @xmath690\omega \right\rangle $ ] is a real - valued function ; the property of reality of @xmath691 then follows from eqs .
( [ 34])-([gege4 ] ) , and ( [ a326 ] ) . proving the positivity of @xmath692 from the positive - definiteness of the wightman functions
requires a longer argument which relies on the introduction of a full fourier - laplace transformation .
this will be given in 5.1 .
it is common to call free ( bosonic , scalar ) quantum field on a given riemannian manifold a distributional solution of the klein - gordon equation adapted to the given manifold .
these theories are also called linear ( as opposed to `` free '' ) because there is an `` interaction '' between the field and the ( unquantized ) metric which gives rise to linear field equations .
given the metric tensor @xmath125 , the corresponding laplace - beltrami is obtained by the well - known formula @xmath693 . in our case an easy way to deal with this operator is to consider the dalembert operator in the ambient space : @xmath694 the laplace - beltrami @xmath695 associated with @xmath696 is seen to be the trace of the dalembertian on @xmath18 .
this is constructed as follows : let @xmath697 .
we associate with it a function @xmath698 defined on the open set @xmath699 , homogeneous of degree zero and such that @xmath700 ; it follows that @xmath701 the de sitter - klein - gordon field equation is then written @xmath702 where @xmath703 is a mass parameter which also includes the coupling to the gravitational field , and we look for an operator - valued distributional @xcite solution for this equation satisfying the properties specified in 2.2 and 2.3 .
there is an important class of solutions of the associated equation @xmath704 , where now @xmath705 is real or complex , which is going to play the same role of _ plane wave basis _ as the exponentials in the minkowski case , but with an important difference : in contrast with the minkowskian exponentials , these waves are singular on three - dimensional light - like manifolds and can at first be defined only on suitable halves of the hyperboloid .
we will need an appropriate @xmath9-prescription ( indicated below ) to obtain global waves .
here is the relevant definition @xcite : let @xmath706 and consider the function @xmath707 defined for those @xmath708 such that @xmath709 ( same definition for @xmath710 for those @xmath708 such that @xmath711 ) ; @xmath712 is an auxiliary mass parameter introduced here for dimensional reasons , but it will also have a minkowskian physical interpretation .
the associated homogeneous function is given by @xmath713 and it is defined in a corresponding open subset of @xmath714 ; it follows that @xmath715 and therefore @xmath716 physical values of the parameter @xmath717 are given by @xmath718 corresponding in the first case to @xmath719 and in the second case to @xmath720 .
this will correspond to the existence of two qualitatively different families of klein - gordon fields , characterized by a mass @xmath703 greater or lower than the `` geometrical '' mass @xmath721 .
correspondingly , the plane waves have an oscillatory or power law behaviour . borrowing a terminology used in the context of the @xmath4 representation theory
we may call _ principal _ plane waves ( and consequently _ principal _ quantum fields ) the plane waves belonging to the first class and _ complementary _ plane waves those belonging to the second class .
we will further comment on this distinction . for the time being the following argument , based on the large @xmath1 behaviour ,
makes it clear that principal waves are more closely similar to minkowskian plane waves .
in fact , parameterizing @xmath722 by the wave - vector of a ( minkowskian ) particle of mass @xmath712 , i.e. @xmath723 $ ] , with @xmath724 , yields @xmath725 in this equation , points in the de sitter universe must be described using the minkowskian space - time variable @xmath726 measured in units of the de sitter radius @xmath1 ; for instance we can choose the global parametrization @xmath727 , \label{para1}\ ] ] or equivalently the one given by eq.([tytyty ] ) . from the other side , important field theories , like the massless field conformally coupled to the de sitter metric belong to the complementary class .
let us for the moment restrict our attention to principal klein - gordon fields . in section 3
, we have seen that the two - point function @xmath728 of any de sitter gff satisfying the double analytic structure and suitable growth conditions admits the kllen - lehmann - type representation ( [ a325 ] ) in terms of the elementary perikernels @xmath729 , ( with @xmath730 according to the notations of theorem 3.1 ) .
we shall now enhance this analogy with the wightman qft s in minkowski space - time by proving the following results .
for each positive number @xmath731 the function @xmath732 can be identified up to a constant factor with the two - point function @xmath733 of a _ linear _ de sitter qft satisfying eq.([kg ] ) with the mass parameter @xmath734 = \left({d-1 \over 2r } \right)^2+m^2 ; $ ] this constant factor is then determined unambiguously by imposing the canonical commutation relations ( ccr ) or the local hadamard behaviour ( see @xcite and references therein ) with the relevant coefficient of the dominant term ( i.e. the value of the corresponding quantity in the minkowskian case ) .
moreover , the function @xmath735 will be introduced by an integral representation in terms of principal de sitter plane waves which is completely analogous to the standard fourier representation of the two - point function of the minkowskian free field .
writing such a plane - wave representation immediately sets the problem of extending the previous plane waves @xmath736 of 4.1 to the whole ( real ) hyperboloid .
in particular , we need to specify the correct phase factors to glue together @xmath737 and @xmath738 .
following @xcite we solve this problem by making use of the properties of the complex hyperboloid @xmath143 which have been described in 2.3 ; in particular the tuboids @xmath148 and @xmath149 play here a crucial role . indeed , when @xmath739 is fixed and @xmath740 varies in @xmath148 or in @xmath149 then the plane waves @xmath741 are globally defined because @xmath742 has a fixed sign .
the phase is chosen such that @xmath743 with @xmath744 for @xmath745 real .
these global waves reproduce the analyticity properties of the minkowskian plane - waves @xmath746 in the corresponding tubes @xmath747 and @xmath748 ( for @xmath749 in @xmath750 let us now introduce for convenience the notion of an orbital basis @xmath751 of @xmath752 with respect to a subgroup @xmath753 of @xmath754 which is the stabilizer of a unit vector @xmath755 @xmath756 in @xmath757 @xmath300 ) if @xmath758 @xmath751 is the section of @xmath752 by a hyperplane of the form @xmath759 @xmath760 namely an orbit ( of spherical type ) of the corresponding subgroup @xmath761 @xmath762 ) if @xmath763 @xmath751 is the union of the sections of @xmath752 by two hyperplanes of the form @xmath764 @xmath765 namely the union of two hyperboloid sheets which are orbits of the corresponding subgroup @xmath766 we shall make a special use of the following bases @xmath767 and @xmath768 of respective types @xmath300 and @xmath762 .
@xmath769 all the orbital bases @xmath770 of either type @xmath771 or @xmath772 , equipped with a suitable orientation , belong to the same homology class @xmath773 in @xmath774 , where the latter denotes a homology group associated with a suitable family @xmath775 of supports , sufficiently regular at infinity @xcite . for example , the compact cycle @xmath776 and the non - compact cycle @xmath777are easily seen to be homologous , as enclosing the region of the cone @xmath778 described as follows : @xmath779 we can now define in a consistent way on each orbital basis @xmath770 a measure @xmath780 which is invariant under the corresponding subgroup @xmath781 by putting @xmath782 ; here @xmath783 denotes the ( d-1)-form defined on @xmath778 as the contraction of the vector field @xmath784 ( where @xmath784 is the restriction to @xmath778 of the vector field whose cartesian components are @xmath785 ) and the volume form @xmath786 .
we shall now rely on the following lemma , which can be directly checked . for every function @xmath787 on @xmath778 which is homogeneous of degree @xmath788
, the @xmath63-form @xmath789 is closed .
we can then state : [ maintheorem ] for each @xmath790 there exists a de sitter gff satisfying all the properties a ) , b ) , c ) , d ) described in section 2 which is a solution of the linear field equation ( [ kg ] ) with @xmath791 the corresponding two - point function @xmath792 labelled by the dimensionless parameter @xmath793 is given by the integral representation : @xmath794 in which @xmath751 denotes _ any _ orbital basis of @xmath752 and @xmath795 is a positive constant .
* from the analyticity properties of the plane - waves ( [ waves ] ) , it is clear that the function @xmath796 ( depending _ a priori _ on @xmath751 ) defined by eq.([wig1 ] ) is analytic in the tuboid @xmath797 of @xmath798 .
moreover , @xmath799 is of moderate growth when @xmath319 tends to the real from the tuboid @xmath199 , and therefore it admits a distributional boundary value @xmath800 on @xmath801 ( theorem a.2 . ) .
now , fixing @xmath802 and putting @xmath803 we can write @xmath804 since lemma 4.1 obviously applies to @xmath805 , it follows that the value of the integral ( [ wig111 ] ) is independent of the orbital basis @xmath806 , the integrability at infinity on non - compact bases of type @xmath772 being also ensured by the homogeneity properties of @xmath807 . to see this explicitly ,
consider for example the basis @xmath777 in eq .
( [ 4.8 ] ) which may be parametrized as follows : @xmath808 the corresponding invariant measures are written @xmath809 and @xmath810 = m\,d\vec{k}/{k^0}$ ] . in these variables eq .
( [ wig1 ] ) is rewritten as the following absolutely convergent integral : @xmath811 we conclude that for all @xmath770 , eq . ( [ wig1 ] ) defines the same analytic function @xmath812 in @xmath199 .
let us now check that its boundary value @xmath813 satisfies the positivity condition .
we consider for each test function @xmath814 the integral @xmath815 where the point @xmath816 varies in @xmath148 according to the following definition ( see eq .
( [ delta ] ) ) : @xmath817 , @xmath818 being a time - like vector in the meridian plane of @xmath18 determined by the conditions @xmath819 , @xmath820 , @xmath821 . plugging the expression ( [ wig1 ] ) of @xmath822 into ( [ piopo ] ) shows that @xmath823 ( for all @xmath787 and @xmath824 ) .
condition ( [ positivity ] ) then follows by taking the limit @xmath825 .
the hermiticity property , implied by positivity , is also obtained by taking boundary values from @xmath199 of the identity @xmath826 , easily checked on eq .
( [ wig1 ] ) .
since all the complex waves @xmath827 satisfy in their domain the complex version of the desitter - klein - gordon equation , it follows that eq .
( [ kg ] ) ( with @xmath828 ) is satisfied in both variables @xmath829 by the function @xmath830 ; correspondingly , by passing to the boundary values , one sees that the kernel @xmath831 is a distribution solution on @xmath112 of the equations : @xmath832 in order to introduce a bona fide de sitter field @xmath833 with two - point function @xmath834 , it is necessary to show that @xmath834 , not only satisfies the positivity and normal analyticity conditions a ) and d ) checked above , but also the locality , covariance and hermiticity properties properties .
once this is done , one can easily check that , in view of eqs .
( [ wigg ] ) , the field @xmath835 constructed in terms of @xmath813 by the formulae ( [ crea ] ) and ( [ azzo ] ) satisfies the equation @xmath836 as an operator - valued distribution on the relevant hilbert space @xmath837 ( in this theory , the subspace @xmath121 is characterized as a set of classes of test functions @xmath127 on @xmath18 , modulo the functions which are of the form @xmath838 ; all the subspaces @xmath839 are characterized similarly ) . to prove the covariance property c ) we first of all recall that in view of eq .
( [ tubi1 ] ) @xmath163 and @xmath164 are invariant under the action of the real de sitter group , like @xmath161 , @xmath162 and @xmath6 .
therefore , if @xmath840 denotes the point obtained by applying a real de sitter transformation @xmath125 to the point @xmath841 , the expression @xmath842 remains meaningful ; let us now show that @xmath843 let us consider any subgroup @xmath781 of @xmath4 and write eq .
( [ wig1 ] ) with an orbital basis @xmath770 invariant with respect to @xmath781 . in view of the invariance of the measure @xmath844 and of @xmath770 under the @xmath781
, it follows that eq .
( [ covariance ] ) holds for @xmath845 .
since any element of @xmath52 may be decomposed as the product of transformations belonging to the subgroups @xmath781 we have displayed , it follows that eq .
( [ covariance ] ) is true for any @xmath846 .
taking into account the fact that @xmath735 is analytic in @xmath847 it follows ( as in the proof of proposition 2.2 ) that the function @xmath735 extends to an invariant perikernel on @xmath650 ( analytic in @xmath209 ) which can be identified with a function @xmath848 of the single invariant variable @xmath849 in order to prove the locality condition b ) , let us also introduce the invariant perikernel @xmath850 , satisfying @xmath851 and such that @xmath852 .
we will then show that @xmath735 and @xmath853 can both be identified ( up to a factor ) with the generalized legendre function @xmath854 this is most easily shown by choosing the following points @xmath855 such that @xmath856 @xmath857 at these points , eq.([wig1 ] ) is written as follows : @xmath858 where we have chosen to integrate over @xmath776 ; by changing the integration variable @xmath859 such that @xmath860 allows us to recognize the integral representation ( [ legendre1 ] ) of the function @xmath861 and thus yields : @xmath862 , with @xmath863 .
this shows that @xmath864 is real - valued for all @xmath865 and therefore that one also has : @xmath866 we note that since each real pair @xmath867 in the locality region @xmath678 is on the same orbit of the complex de sitter group as the corresponding pairs @xmath868 and @xmath869 eq.([opq ] ) is also interpretable ( in view of the invariance property c ) for @xmath822 and @xmath870 as the locality relation @xmath871 at any such pair , and implies the identity of @xmath822 and @xmath870 .
the proof of the theorem is complete .
+ the two - point function @xmath735 of the de sitter - klein - gordon field is @xmath872 .
@xmath873 is fixed by imposing the ccr s or local hadamard behavior with the relevant coefficient of the dominant term and is given by @xmath874 * proof .
* the singular behaviour of @xmath875 at coinciding points is given by the behaviour of @xmath876 near its singular point @xmath877 .
using the integral representation ( [ legendre1 ] ) near @xmath878 for @xmath879 and @xmath880 spacelike separated and @xmath881 one gets : @xmath882 @xmath883^{\frac{2-d}{2}}. \label{4.21}\ ] ] the corresponding dominant term for the two - point function of the minkowski klein - gordon field is given by @xmath884^{\frac{d-2}{2}}. \label{4.22}\ ] ] since at short distances the de sitter and the minkowski distances ( in a tangent plane ) are asymptotically equal , we obtain that the two - point function satisfies the hadamard condition @xcite .
the constant @xmath873 may now be fixed by imposing canonical normalisation of the minkowskian case according to eq .
( [ 4.22 ] ) .
this is equivalent to requiring the ccr s . by equating coefficients in eqs .
( [ 4.21 ] ) and ( [ 4.22 ] ) we get eq .
( [ a420 ] ) . the constant appearing in eq .
( [ wig1 ] ) is then given by @xmath885 this result also covers the case @xmath886 , where the corresponding short - distance behaviour of @xmath875 is proportional to @xmath887 .
_ theorem 4.1 and corollary 4.1 can be extended to cover also field theories relative to real values of @xmath888 ( see eq .
( 4.4 ) ) .
the various points of the proof of these statements remain unchanged , except for the proof of the positive - definiteness .
we do not give here the details .
_ we can then state : [ maintheorem2 ] for each @xmath889 with @xmath890 there exists a de sitter gff satisfying all the properties a ) , b ) , c ) , d ) described in section 2 which is a solution of the linear field equation ( [ kg ] ) with @xmath891 the corresponding two - point function @xmath892 labelled by the dimensionless parameter @xmath793 is given by the following integral representation : @xmath893 in which @xmath751 denotes _ any _ orbital basis of @xmath752 and @xmath894 is given in corollary 4.1 .
[ gyu ] _ the existence of these two classes of klein - gordon fields on de sitter space - time shows that the kllen - lehmann decomposition ( [ a325 ] ) given in section 3 is not the most general one for two - point functions on de sitter space - time .
indeed theories covered by that decomposition are superpositions involving only the two - point functions of the principal series
. we will refer to this fact by calling from now on eq .
( [ a325 ] ) the _ kllen - lehmann principal decomposition_. one expects that in the most general case , obtained _ by relaxing the growth condition corresponding to the analyticity properties used in theorem 3.1 _ , a two - point function satisfying all the properties a ) , b ) , c ) , d ) will admit a kllen - lehmann representation of the form @xmath895 _ _ one meets decompositions similar to eq .
( [ aiu ] ) in the context of @xmath4 representation theory ( see e.g. @xcite ) .
however in general these decompositions have a supplementary term corresponding to what is called _ the discrete series of representations_. these representations should be discarded here because of the locality requirement .
_ _ eq . ( [ wig1 ] ) and ( [ wig1com ] ) gives a new integral representation of the indicated class of gegenbauer functions of the first kind with a clear geometrical interpretation in terms of two points ( i.e. as a kernel ) on the @xmath0-dimensional one - sheeted hyperboloid , summarizing several known integral representations in one single formula .
more precisely this means the following : different integral representations of the gegenbauer function @xmath896 ( @xmath557 complex ) are obtained by choosing a integration path on the cone and two suitable points @xmath879 and @xmath880 of the complex hyperboloid in such a way that @xmath897 .
_ _ the boundary value of eqs .
( [ wig1 ] ) and ( [ wig1com ] ) gives us the following representation for the two - point wightman function of the de sitter - klein - gordon field : _
@xmath898 @xmath899 ( similarly for the complementary case ) where we adopted the notation introduced by gelfand for the distributions @xmath900 @xcite
. the important point of this formula is that it allows a factorization of the two - point function in terms of global plane waves on the hyperboloid @xmath18 which is completely analogous to the corresponding fourier representation for two - point function of the minkowski free field of mass @xmath712 .
the latter will be obtained as the limit of eq .
( [ wig4 ] ) for @xmath901 ( see 5.2 )
. an important consequence of the representation ( [ wig1 ] ) is the introduction of a natural fourier transform on the hyperboloid ; this transform is crucial for what follows and , in our opinion , it will play a central role in any question connected with the harmonic analysis on the hyperboloid . given a function @xmath814
we define its fourier transform as the following pair of homogeneuous complex functions on the cone @xmath778 : @xmath902@xmath903 more generally , we may introduce a fourier transform depending on a complex parameter @xmath717 : @xmath904 the kernels in these fourier transforms are obtained by taking the boundary values of the plane waves @xmath905^s$ ] from the corresponding tubes @xmath906 .
this implies that there is a relationship between the analyticity properties of the test functions and the vanishing of the corresponding fourier transforms .
a detailed presentation of the mathematical properties of this fourier transform will be given somewhere else .
we end this section by giving an explicit @xmath907 representation of the scalar product ( [ hjk ] ) for @xmath908 belonging to the principal class , which is analogous to the momentum - space representation of one - particle states in minkowskian qft .
indeed , with the help of the fourier transform ( [ fourier ] ) we may rewrite the scalar product ( [ hjk ] ) as follows : @xmath909 which directly shows that @xmath910 .
this entails that the `` one - particle '' hilbert space of the theory @xmath121 may be realized as the space @xmath911 of complex - valued functions defined on the orbital basis @xmath912 and square integrable w.r . to the corresponding measure @xmath913 ( this assertion will be fully justified in the next subsection
; here we can only conclude that the one - particle hilbert space of the theory is a closed subspace of @xmath911 ) .
again , the full hilbert space @xmath107 of the representation can be described as the hilbertian sum @xmath119 $ ] . finally , we obtain the following fourier - type representation of the field operator : @xmath914 @xmath915 in this subsection we briefly discuss the representations of the de sitter group which are associated with the free de sitter field .
we will treat the `` one - particle '' representation ( in @xmath121 ) , the full representation being constructed by the fock procedure . since representation theory is not our principal aim , many mathematical details will be skipped and we will keep the discussion at a formal level ; moreover we will give explicit formulae only for the two dimensional case , which is notationally much simpler than the general one ; however the results are completely general and mathematically rigorous .
let us start by showing the following the representation of @xmath52 carried by the free de sitter - klein - gordon field of mass @xmath828 is the unitary irreducible representation of the principal series characterized by the value @xmath916 of the casimir operator .
* let us consider the representation @xmath86 of @xmath52 that we defined in 2.2 , restricted to the one - particle space of the borchers algebra @xmath917 : @xmath918 . by taking the fourier
transform ( [ fourier ] ) of this equation we obtain a representation of @xmath52 in the space @xmath919 of complex valued functions defined on @xmath920 , homogeneuous of degree @xmath717 ( @xmath921 ) , which is given by @xmath922 associated with each @xmath923 there is a @xmath52-invariant sesquilinear form @xmath924 given by @xmath925 in this equation @xmath770 is any orbital basis of @xmath920 and @xmath926 is the ( corresponding ) measure assuring @xmath52 invariance .
( [ sesqui ] ) gives actually a pre - hilbert product only for certain values of the parameter @xmath717 ; in particular for @xmath927 we get by theorem [ maintheorem ] that the sesquilinear form ( [ sesqui ] ) coincides with the pre - hilbert product associated with the two - point function @xmath813 .
since the measure @xmath926 lives on the submanifold @xmath770 of @xmath920 , one can naturally introduce the operation @xmath928 @xcite which consists in taking the restriction to @xmath770 of a function @xmath929 : @xmath930 this induces a representation @xmath931 of @xmath52 on a space of ( suitably smooth ) complex - valued functions defined on the manifold @xmath770 ; this representation is obtained as follows : @xmath932 concrete realizations are then obtained by specifying a choice of @xmath770 . in particular , the choice @xmath933 allows us to show ( by explicit construction ) that the representation @xmath934 extends to the unitary irreducible representation @xmath935 of the principal series , corresponding to the value @xmath936 of the casimir operator and realized in the hilbert space @xmath937 , where @xmath938 is the rotation invariant measure on @xmath776 .
+ _ as one can see , the procedure we used for constructing the representation is very similar to the one outlined in @xcite .
the further information we have added consists in the fact that here the whole construction is obtained naturally as an output of the corresponding quantum field theory .
indeed , we simply started by considering ( smooth ) functions on the hyperboloid @xmath18 , together with the left regular representation of the corresponding relativity group . all the further steps , including the introduction of the relevant spaces of homogeneous complex valued functions defined on the cone @xmath920 and the restriction of those functions to a suitable submanifold of @xmath939 , are imposed by the two - point function of the quantum field theory . _ in an analogous way we can state the representation of @xmath52 carried by the free de sitter - klein - gordon field of mass @xmath940 is the unitary irreducible representation of the complemantary series characterized by the value @xmath916 of the casimir operator .
we now specialize the discussion to the case @xmath886 and we briefly show how to construct the principal representations and take corresponding contraction .
the generators of the relativity group , which in this case is @xmath941 , are given by @xmath942 \;\;\;\;\;\;\ ; l_{02}=\left [ \begin{array}{ccc } 0 & 0 & 1 \\ 0 & 0 & 0 \\ 1 & 0 & 0 \end{array } \right]\;\;\;\;\;\;\ ; l_{12}=\left [ \begin{array}{ccc } 0 & 0 & 0 \\ 0 & 0 & 1 \\ 0 & -1 & 0 \end{array } \right ] \label{generators}\ ] ] a geodesical obsever passing through the origin of our de sitter space - time sees @xmath943 as the generator of lorentz transformations , @xmath944 as the generator of time translations and @xmath945 as the generator of space translations
. a generic element of @xmath946 may be parametrized in the following way : @xmath947 it follows from eqs .
( [ generators ] ) and ( [ factori ] ) that @xmath948 has the following action on @xmath920 : @xmath949 @xmath950 a concrete realization of the representation ( [ repgamma ] ) is obtained by choosing a ( complete ) curve @xmath770 . to compare our results with the representations of @xmath948 as they are customarily given in literature
we choose the curve @xmath951 ; the associate mapping @xmath952 has the following form : @xmath953 the inverse map ( defined on the image of @xmath954 ) is easily obtained as @xmath955 as in eq .
( [ repgamma ] ) the use of the map @xmath954 gives a representation @xmath931 of @xmath946 on @xmath956 . here
are explicit formulae for the three subgroups associated to the generators ( [ generators ] ) : @xmath957 @xmath958 @xmath959 @xmath960 @xmath961 one sees from these formulae that the choice of @xmath776 as integration manifold produces a representation of @xmath948 on a functional space in which rotations ( the subgroup @xmath962 ) are realized in a natural way .
this is consequence of the rotation invariance of the manifold @xmath776 .
one can also show directly that when @xmath963 the representation can be extended to the unitary irreducible representation of the principal series with casimir @xmath964 realized on the hilbert space @xmath965 .
we now turn to the discussion of the contraction of these representations . to this aim , it is better to realize them by using another manifold , namely the curve @xmath966 . as we have seen in 4.2 , this manifold has a more direct physical interpretation , and furthermore , this choice allows a contraction of the principal series of representations of the de sitter group ( onto the representations of the poincar group ) which is very direct ( compare our formalism with the one presented in @xcite ) since it involves only the representations of the poincar group @xmath967 : constructed ( as usual in qft ) _ la _ wigner and requires no use of group - theoretical structures . to this end
we define the map @xmath968 : @xmath969 = ( q_{\gamma_2}\tilde{f})(k ) = \left [ \begin{array}{c } \tilde{f}(\xi^{(0)},\,\xi^{(1)}=k,\,\xi^{(2)}=m)\\ \tilde{f}(\xi^{(0)},\,\xi^{(1)}=-k,\,\xi^{(2)}=-m ) \end{array } \right]\ ] ] where obviously @xmath970 . the inverse map ( defined on the image of @xmath971 ) is obtained as @xmath972 the map @xmath971 gives an alternative realization of the representation @xmath931 of @xmath946 , which lives on the space @xmath973 ; as before , when @xmath974 , @xmath931 extends to a unitary irreducible representation , which is now carried by the hilbert space @xmath975 .
the three previous subgroups of @xmath948 are now represented as follows : @xmath976@xmath977 , \label{sptr1}\ ] ] where @xmath978 and @xmath979 denotes the heaviside function ; @xmath980 @xmath981 \label{ttr1}\ ] ] where @xmath982 ; @xmath983\label{lorentz}\ ] ] _ one sees that now it is the representation of the lorentz subgroup @xmath984 which is realized naturally , and this is due to the lorentz invariance of the manifold @xmath985 . _ to discuss the contraction we also need the generators of @xmath967 : @xmath986,\;\;\;\ { \cal l}_{02}=\left [ \begin{array}{ccc } 0 & 0 & 0 \\ 1 & 0 & 0 \\ 0 & 0 & 0 \end{array } \right],\;\;\;\ ; { \cal l}_{12}=\left [ \begin{array}{ccc } 0 & 0 & 0 \\ 0 & 0 & 0 \\ 1 & 0 & 0 \end{array } \right].\;\;\;\;\ ] ] a generic element of @xmath987 may be parametrized as @xmath988 the group contraction map is defined in the following way : @xmath989 when the point @xmath990 belongs to a suitable ( @xmath824 depending ) neighbourhood of @xmath991 , we can inverte this map : @xmath992 @xmath967 is then a contraction of @xmath946 because @xmath993 for the reader s convenience we remind the definition of contraction of a family of representations .
let @xmath994 be a lie group and let @xmath995 its contraction .
consider a one - parameter family @xmath996 $ ] , of continuous and unitary representations of @xmath994 on hilbert spaces @xmath997 . furthermore consider a family of unitary maps @xmath998 such that for each @xmath999 there exists @xmath1000 for which if @xmath1001 then @xmath1002 . if the limit @xmath1003 exists for each @xmath999 and defines a continuous unitary representation of @xmath995 , then the representation @xmath1004 of @xmath995 is said a contraction of the representation @xmath1005 of @xmath994 . in our concrete case we have that @xmath1006 and @xmath1007 as defined in eq .
( [ function ] ) .
our aim is to prove that if @xmath1008 , @xmath1009 , the limit in eq .
( [ limit ] ) exists and that @xmath1010({\sl g})\psi.\ ] ] _ there are two things in our construction which render the contraction easier : first , the representations of @xmath948 and the representation @xmath1011 of the poincar group ( which will be obtained by contraction ) are realized in the same functional space .
secondly , one sees in eq .
( [ function ] ) that the lorentz subgroup stays untouched in the contraction .
our construction provides a representation of the lorentz subgroup which _ has the same functional form _ for the two groups .
_ it follows from the previous remark that for the lorentz subgroup there is actually no limit to take .
let us examine the space translations subgroup .
we prove pointwise convergence assuming all the continuity properties needed . by setting @xmath1012 in eq .
( [ sptr1 ] ) we then obtain that @xmath1013.\ ] ] which corresponds exactly to the desired representation of the minkowskian space translations . in exactly the same way
we may show that given a generic @xmath1014 we obtain the following result : @xmath1015 @xmath1016\ ; ( { a},a_{0},\chi)\;\underline{\tilde{f}})({k})\ ] ] thus , the unitary representations of the principal series of the de sitter group are contracted pointwise to a direct sum of wigner representations of the poincar group.furthermore , it can also be shown that the contraction is obtained also in the hilbert space topology , i.e. @xmath1017\ ; ( { a},a_{0},\chi)\;)\underline{\tilde{f } } \|_{{\cal h } } = 0\ ] ] but the complete proof will not be reproduced here ( see e.g. @xcite ) .
we begin this section by completing the proof of the kllen - lehmann principal representation we initiated in section 3 , by showing the positivity of the weight @xmath1018 for general two - point functions . for the sake of simplicity we carry out detailed computations in the two - dimensional case but
the general case can be easily recovered .
we use here the parametrization of @xmath1019 given in eq .
( [ para1 ] ) : @xmath1020 , ( @xmath1021 , @xmath1022 , @xmath1023 ) where we set @xmath1024 without loss of generality . in the following
, a special role will be played by the class of rotation invariant functions .
if we put @xmath1025 , rotation invariance of a function @xmath787 is expressed by @xmath1026 .
of course the fourier transform of @xmath1027 will also be rotation invariant .
indeed , by parameterizing the vector @xmath722 in the following way : @xmath722=(@xmath1028 , @xmath1029 , @xmath1030 ) one sees that @xmath1031^{s}{\underline{f}}(\psi)d\theta \cosh \psi d\psi \label{trasfoy}\ ] ] where the angle @xmath1032 has a small positive imaginary part , and one sees that the integral does not depend on @xmath770 ( analogously for @xmath1033 . to compute explicitly the fourier transform ( [ trasfoy ] ) we change the variables as follows : @xmath1034 this parametrization covers only one quarter of @xmath1035 .
the remaining parts are obtained similarly .
it follows that @xmath1036 thus the fourier transform ( [ trasfoy ] ) of a rotation invariant function is written in these variables as a weyl transform followed by an ordinary fourier transform .
let @xmath1037 be a test - function on @xmath1035 , with analyticity properties in @xmath1038 , namely @xmath1039 , where @xmath1040 is analytic in the strip is contained in @xmath1041 . ]
@xmath1042 with the consistency condition : @xmath1043 then @xmath1044 vanishes.[tu1 ] * proof . *
write the fourier transform @xmath1044 as an integral in @xmath1045 and @xmath1046 . at
fixed @xmath1047 the integration cycle in @xmath1038 can be distorted from @xmath1048 to @xmath1049 this integral is then identified ( in view of eq .
( [ ( 2 ) ] ) ) with _ minus _ the corresponding integral at @xmath1050 , which implies the vanishing of @xmath1044 and proves the statement . + thus if @xmath1051 is analytic in the strip @xmath1052 and satisfies the condition @xmath1053 then @xmath1054 .
this gives that @xmath1055^{1/2}}\ .
\label{5.5}\ ] ] we shall now make use of the legendre functions as analytic test - functions .
let @xmath1056 the function @xmath1057 is analytic in the strip @xmath1058 and moreover @xmath1059 ( since @xmath1060 is analytic for @xmath1061 and @xmath1062 @xmath1063 therefore satisfies the above property and therefore @xmath1064 .
( [ 5.5 ] ) applies and yields @xmath1065 @xmath1066^{1/2}}d \sinh \psi \label{tu2}\ ] ] to compute the last integral we use the following representation of @xmath1067 ( = @xmath1068 ) : @xmath1069^ { 1/2 } } { d } t \label{apdamp}\ ] ] eq .
( [ apdamp ] ) can thus be identified with the inversion of the weyl transform is written as @xmath1070^{1/2}}\ ] ] then @xmath1071^ { 1/2 } } \label{a466}\ ] ] ] occurring at the l.h.s .
( [ tu2 ] ) which allows us to compute the latter ; we thus obtain @xmath1072 @xmath1073 \label{aiuto}\ ] ] in view of the results of 4.2 , formula ( [ a325 ] ) can be interpreted as a kllen - lehmann representation expressing the two - point functions of a certain class of fields ( principal - type ) in terms of the two - point functions of all the massive klein - gordon fields @xmath813 , @xmath1074 we now complete the properties of this integral representation by proving assume that a de sitter quantum field has a two - point function of the form @xmath1075 and satisfying the positivity condition a ) , namely @xmath1076 then @xmath1077 is a positive measure . *
* let @xmath1078 be any even function in @xmath1079 , and let us choose the rotation invariant function @xmath1080 whose fourier transform is @xmath1081 eq .
( [ aiuto ] ) then gives , for @xmath1082 : @xmath1083 since @xmath1084 we then have in view of eq . ( [ consta ] ) : @xmath1085 therefore , @xmath1086 @xmath1087 since our freedom of choice on @xmath1088 allows @xmath1089 to vary in a set which is dense in the space of continuous functions with compact support in @xmath1090 , it follows that @xmath1091 is a positive measure .
+ _ _ _ the role of locality for general two - point functions on de sitter space - time .
_ it is known that in minkowski space the full analytic structure of general two - point functions can be obtained for hermitian scalar fields from poincar invariance , spectral condition and positivity , _ without using locality_. the argument goes through a proof of the kllen - lehmann formula based purely on a spectral decomposition of the wightman function ( in `` intermediate states '' ) with respect to the energy - momentum variables . in the present case of de sitter qft ,
there does not exist at the moment a general proof of the kllen - lehmann formula ( [ a325 ] ) ( when it holds ) on the basis of an @xmath4-decomposition of the representation carried by the hilbert space @xmath121 of the two - point function considered . at any rate ,
the use of locality in proposition 2.2 is necessary if one deals with fields whose behaviour at infinity only allows one to write a representation of the form ( [ a321 ] ) ( but not ( [ a325 ] ) ) for their two - point function , as it is the case for those considered below in 5.4 . in this paragraph
we show how to recover minkowski gff s by the knowledge of corresponding de sitter qft by taking the limit @xmath1092 ( contraction ) . to this end , only the principal series is relavant since the complementary series tend to disappear when @xmath464 is very large .
thus it is legitimate to use the kllen - lehmann representation ( [ a325 ] ) .
as a first step we show that the limit of the fourier representation given in eq .
( [ wig1 ] ) is the corresponding well - known fourier representation of two - point functions for klein - gordon fields .
indeed by taking as integration manifold @xmath1093 we see that when @xmath1094 is very large , the dominant term in ( [ wig4 ] ) is given by @xmath1095 the other terms being multiplied by the factor @xmath1096 ( see eqs .
( [ consta ] ) and ( [ wig4 ] ) ) . by inserting in the latter equation the parametrization ( [ para1 ] )
, we then obtain in view of eq .
( [ parcondicio ] ) that @xmath1097 @xmath1098 similarly one can obtain each minkowskian gff as the limit of a family of gff s @xmath1099 on the de sitter universes @xmath1100 , when the radius @xmath1101 of the latter tends to infinity .
we can state : for any kllen - lehmann weight @xmath1102 positive on @xmath1103 there exists a family of generalized free fields @xmath1104 each field @xmath1105 being defined on the corresponding de sitter universe @xmath1106 which satisfy the following properties : \a ) the two - point function @xmath1107 of the fields @xmath1108 is given by the following integral representation : @xmath1109 \b ) for each point @xmath1110 in @xmath1111 the following limit holds : @xmath1112 where @xmath1113 is the two - point function of the minkowskian generalized free field @xmath1114 on @xmath1115 characterized by the kllen - lehmann weight @xmath1116 namely : @xmath1117 * proof .
* in view of theorem 4.1 every two - point function of the form ( [ a328 ] ) with @xmath1118 satisfies conditions a ) , b ) , c ) , d ) , and therefore defines a unique de sitter gff on @xmath582 , which proves property a ) .
property b ) follows from eq .
( [ conttrac ] ) .
+ _ we notice that the unitary representation of the poincar group constructed by means of the distribution @xmath1119 is actually _
irreducible _ ; namely , it is the wigner representation of mass m and spin zero . on the other side , as we have seen in 4.4 , the representations obtained as contraction of the unitary representations of the de sitter group _ are not irreducible _ @xcite .
_ therefore the group - theoretical contraction is not a suitable procedure to draw conclusion on the structure of the limiting quantum field theory .
_ _ _ _ we can complement proposition 5.2 by applying the analysis of sections 2 and 3 to the two - point function @xmath1120 on @xmath582 , to the corresponding propagators @xmath1121 , and to their laplace transform @xmath1122 ( obtained from @xmath1123 via eq .
( [ 34 ] ) .
first of all , it can be checked that the weight function @xmath1124 of theorem 3.1 corresponding to the field @xmath1125 is given by @xmath1126 since , by putting @xmath1094 in eq .
( [ a325 ] ) , the latter can be identified with eq .
( [ a328 ] ) .
one can then see that there exists an @xmath1-independent _ normalized laplace transform _
@xmath1127 which is common to all the fields @xmath1125 ( including the minkowskian limit @xmath1128 , g being equal to the standard minkowskian laplace transform of the corresponding retarded propagator ) . to this purpose
, it is relevant to associate with the space time - variables @xmath7 the complex variable @xmath1129 which carries a _ mass dimension _ and to define the _ normalized laplace transform _
@xmath1130 of @xmath1131 , by the scaling equation : @xmath1132 it can be checked that the function @xmath1133 , analytic in the half - plane @xmath1134 , is uniquely determined from @xmath138 by the fact that its boundary value on @xmath1135 satisfies the following relation ( implied by eqs .
( [ a326]),([5.22 ] ) and ( [ 5.23 ] ) ) : @xmath1136\ ] ] as a matter of fact , for each mass @xmath712 the corresponding de sitter - klein - gordon field on @xmath582 ( solution of eq .
( [ kg ] ) , with @xmath1137 admits the following normalized laplace transform : @xmath1138 .
the kllen - lehmann formula written for the laplace transform of the propagators therefore admits the same interpretation as that of the minkowskian case , namely it is equivalent to the cauchy representation @xmath1139 .
one simple but surprising theorem in local quantum field theory is the reeh and schlieder property @xcite .
we will show that an analogous theorem still holds for de sitter generalized free fields . in a sense
this could be even more surprising , due to the presence of horizons in this model of space - time . in the following we give a detailed analysis of the situation for the principal free and generalized free fields .
a similar analysis could be performed for the complementary fields .
let @xmath78 be the free de sitter - klein - gordon quantum field of parameter @xmath557 .
the vector valued distribution @xmath1140 is the boundary value of a vector - valued function @xmath1141 strongly analytic in the tuboid @xmath148 .
[ pr9 ] * proof .
* let us consider a vector @xmath1142 where @xmath1143 is the dense subset of @xmath1144 obtained by applying the fields @xmath131 , @xmath1145 to the `` vacuum '' @xmath1146 ; we can then put @xmath1147 the analyticity properties of the two - point function then imply that the correlation function between @xmath1148 and the `` vacuum '' @xmath1149 is the boundary value of a function @xmath1150 which is analytic in the tuboid @xmath148 , i.e. @xmath1151 using the fourier representation ( [ wig1 ] ) for the two - point function we deduce the following majorization : @xmath1152 @xmath1153 since the set @xmath1154 is dense in @xmath1144 , the majorization ( [ maggiorazione ] ) and the riesz lemma imply that there exists a vector @xmath1155 such that @xmath1156 given now @xmath1157 , we may choose a sequence @xmath1158 , such that @xmath1159 in @xmath1144 .
associated with the sequence @xmath1160 , there is a sequence @xmath1161 of functions analytic in @xmath148 . by applying again the inequality ( [ maggiorazione ] )
we obtain that the functions @xmath1162 satisfy a uniform bound on each compact subset of @xmath148 ; moreover , the continuity of the hilbert product implies that , for each @xmath1163 @xmath1164 it then follows ( by vitali s theorem ) that the function @xmath1141 is weakly ( and therefore strongly ) analytic .
the remaining part of the lemma is then proven as in the flat case , because the constant @xmath1165 in eq .
( [ maggiorazione ] ) may be bounded as follows for any @xmath1166 : @xmath1167 and therefore ( in view of the results of appendix a ) the analytic function @xmath1168 ( with @xmath1169 ) admits a boundary value @xmath1170 on @xmath18 , in the sense of distributions such that @xmath1171 + _ in the @xmath722-space realization of the fock space that we have given in eq .
( [ azzolino ] ) the vector @xmath1172 which has as only nonzero component @xmath1141 may be represented ( for @xmath1163 ) by the vector @xmath1173 in the same way , if @xmath1174 the analytic function @xmath1175 is given in @xmath148 by its fourier - type representation @xmath1176 we could have used this more direct and simple argument to verify the strong analyticity of @xmath1177 but we went through the previous lemma because it promises to generalise to the interacting case . _ for every bounded open set @xmath1178 let @xmath1179 be the polynomial algebra of the free field of parameter @xmath557 affiliated with the region @xmath100 .
we then have : ( reeh - schlieder property for the free field ) for any open set @xmath100 of ( real ) de sitter space - time the `` vacuum '' is a cyclic vector for @xmath1179
. * proof . * the previous proposition 5.3 together with corollary a.1 imply that if the distribution @xmath1170 ( for @xmath1169 ) vanishes in an open set @xmath100 of @xmath18 then it vanishes on @xmath18 ; this in turn implies that @xmath1180 by construction of @xmath121 ( see 2.2 ) .
this proves that the set @xmath1181 is dense in @xmath1144 . the proposition is then proved by observing that the @xmath10-particle space @xmath1182 is the symmetrized tensor product of @xmath10 copies of @xmath1144 .
+ for every bounded open set @xmath1178 let @xmath1183 be the polynomial algebra of a generalized free field of weight @xmath1077 ( see proposition 5.2 ) affiliated with the region @xmath100 .
using now the kllen - lehmann representation of the corresponding two - point function , namely @xmath1184 , one could again justify estimates similar to ( [ maggiorazione ] ) and ( [ uffaulla ] ) , and therefore prove the analyticity and boundary value properties of the vector - valued function @xmath1141 for any such field ; the argument used for proving proposition 5.4 would then yield similarly : ( reeh - schlieder property for the generalized free fields ) for any open set @xmath100 of ( real ) de sitter space - time the `` vacuum '' is a cyclic vector for @xmath1183 . the idea that a natural class of de sitter qft s should be obtained by taking the restriction to @xmath1019 of qft s defined in the ambient minkowski space @xmath1185 has been proposed in @xcite .
the axiomatic approach of section 6 below , based on analyticity domains of the wightman functions which are the restrictions to @xmath1186of the standard axiomatic domains in @xmath1187 of the wightman functions of a @xmath1188-dimensional minkowskian qft gives a priori an additional support to this idea . in this connection , it is then useful at first to study the class of de sitter fields in @xmath1019 obtained from minkowskian free fields ( or gff ) in @xmath1189 by this procedure . our comments for this class of fields follow .
let @xmath1190 be the two - point function of a given free field ( or gff ) @xmath1191 in @xmath1192 @xmath1193 is analytic in the cut - domain @xmath1194 where @xmath1195 since the section of @xmath1196 by @xmath1197 is @xmath1198 the analytic function @xmath1199 satisfies the normal analyticity condition d ) ; it therefore admits distribution boundary values @xmath1200 and @xmath1201 from the respective tuboids @xmath847 and
@xmath1202 which define the restrictions to @xmath1019 of the corresponding wightman distributions of the field @xmath1191 namely : @xmath1203 the positivity condition a ) for @xmath680 then follows from the corresponding condition for @xmath1204 on @xmath1205 by taking sequences of test - functions on @xmath1189 which tend to @xmath1206 with @xmath1207 conditions b ) and c ) are trivially satisfied .
therefore in view of the results of section 2 , there exists a well - defined gff @xmath1208 on @xmath1019 whose @xmath1209-point functions are the restrictions to @xmath1210 of those of @xmath1211 and which satisfies the thermal interpretation of 2.4 .
however there does not exist a kllen - lehmann representation of the principal type for these fields .
in fact , one knows the asymptotic behaviour of the massive free field two - point function @xmath1212 with mass @xmath1213 in dimension @xmath1214 namely @xmath1215^{1/2 } } \over \left [ \left(x_1-x_2 \right)^2 \right]^{d/4}}\ , \ ] ] which , by restriction to @xmath801 yields : @xmath1216^{1/2 } } \over \left[2 \left ( { \rm cosh}{\delta \over r}-1 \right ) \right]^{d/4}}\ , \ ] ] for @xmath1217 this behaviour is strictly larger than the behaviour given by the klein - gordon de sitter two point functions of the principal series ( i.e. @xmath1218 which is such that @xmath1219 it follows that the laplace transform @xmath1220 of @xmath1221 is analytic ( in view of proposition 3.1 ) in the half - plane @xmath1222 but certainly has a singularity on the boundary of the latter .
therefore @xmath1223 admits a representation of the form ( [ a321 ] ) in terms of the legendre functions @xmath1224 which can not be put in the kllen - lehmann form ( [ a325 ] ) , due to the singularities of @xmath1225 ( in this respect see also remark ( [ gyu ] ) ) . on the contrary , the massless field @xmath1226 in @xmath3 ( considered in @xcite )
behaves at infinity as @xmath1227^{-\frac{d-1}{2}}$ ] and its restriction to @xmath18 actually has a principal - type behaviour .
in fact one can produce @xcite a kllen - lehmann decomposition of these fields in terms of the de sitter klein - gordon fields of the principal series .
we shall now propose a general characterization of what could be ( in this connection see also @xcite ) a satisfactory quantum field theory on de sitter space - time .
we shall adopt the approach presented in 2.2 , which is based on the gns - type construction associated with a positive linear functional @xmath1228 on the borchers - uhlmann algebra @xmath1229 on @xmath1230 in the general case of interacting fields , the functional @xmath1228 can be characterized by a sequence of wightman distributions @xmath1231 on @xmath1232^n $ ] satisfying an appropriate set of axioms ( as in the garding - wightman reconstruction theorem for minkowskian theories @xcite ) .
if three of these axioms ( listed under 1 , 2 and 3 below ) are straightforward adaptations to the de sitter space - time of the corresponding axioms of minkowskian q.f.t . , the fourth one which plays the role of the spectral condition must be given a completely new form .
in fact , due to the absence of a global energy - momentum interpretation on the curved space - time @xmath1233 our fourth axiom will be formulated in terms of analytic continuation properties of the distributions @xmath1234 in the complexified manifolds @xmath1235^n $ ] corresponding to @xmath1236^n .
$ ] of course , the choice of such global analyticity properties may certainly be done in a non - unique way , and the corresponding properties of the gns vacuum @xmath1237 of the considered theories will of course depend on the postulated analyticity properties .
this is in contrast with the gff case , which is determined unambiguously by the perikernel analytic structure of the two - point functions , and equivalently characterized by the thermal interpretation of the gns vacuum . in the general case
, the thermal interpretation of the theory might provide a criterion of selection between several _ a priori possible _ global analytic structures .
we do not address this kind of question here , but we would just like to propose , as an open suggestion , a rather natural form for the fourth axiom , which leads to an axiomatic approach of desitter qft very close to the wightman axiomatic approach by its analyticity properties and by the corresponding program which it suggests .
the set of distributions @xmath1238 ( @xmath104 ) is assumed to satisfy the following properties ( we limit ourselves here to the case of a boson field ) : 1 .
( covariance ) .
each @xmath1239 is de sitter invariant , i.e. @xmath1240 for all de sitter transformations @xmath125 , where @xmath1241 ( eq . ( [ cov ] ) implies that the one - point function is a constant , which can be put equal to zero by additive renormalization of the field , as in the minkowski case ) .
2 . ( locality ) @xmath1242 if @xmath1243 .
3 . ( positive definiteness ) . given @xmath1244^{k}),$ ]
then @xmath1245 where @xmath1246 we shall now give a substitute for the usual spectral condition of axiomatic field theory in minkowski space , which will be called `` weak spectral condition '' .
we propose under this name analyticity properties of the wightman functions which reproduce as closely as possible those implied by the usual spectral condition in minkowskian qft . in the latter
, it is known that the spectral condition can be equivalently expressed by the following analyticity properties of the wightman functions @xmath1239 @xcite resulting from the laplace transform theorem in @xmath1247 : for each @xmath1248 the distribution @xmath1249 is the boundary value of an analytic function @xmath1250 defined in the tube @xmath1251 when the minkowski space is replaced by the de sitter space @xmath18 embedded in @xmath27 , a natural substitute for this property can be proposed by replacing ( for each @xmath10 ) the tube @xmath1252 by the corresponding open subset @xmath1253^{n}\cap { \rm t}^{d+1}_{n-1}$ ] of @xmath1254^{n}$ ] . using the notions introduced in appendix
a and an argument similar to the proof of proposition 2.1 . one can show the following \i ) the set @xmath1255 is a domain of @xmath1254^{n}$ ] ; + ii ) @xmath1256 is a tuboid above @xmath1257^{n}$ ] , with profile @xmath1258^{n } } ( { \underline{x}},\lambda^{n}_{{\underline{x}}})\ ] ] where , for each @xmath1259^{n}$ ] , @xmath1260 is a non - empty open convex cone with apex at the origin in @xmath1261^{n})$ ] defined as follows : @xmath1262 [ pr1 ] proposition 6.1 .
allows us to state the following axiom : 1 .
( weak spectral condition ) . for each @xmath10
the distribution @xmath1263 is the boundary value of an analytic function @xmath1264 , defined in the tuboid @xmath1256 of the complex manifold @xmath1254^{n}$ ] by an application of theorem a.2 .
, this axiom implies that for every @xmath177 there is some local tube @xmath1265 around @xmath177 in any chosen system of local complex coordinates on @xmath1254^{n}$ ] whose image in @xmath1254^{n}$ ] is contained in @xmath1256 and has a profile very close to the profile of @xmath1256 ( restricted to a neighbourhood of @xmath177 ) , from which the boundary value equation @xmath1266 can be understood in the usual sense .
it implies equivalently that , in a complex neighbourhood of each point @xmath1267^{n}$ ] , the analytic function @xmath1264 is of moderate growth ( i.e. bounded by a power of @xmath1268 , where @xmath1268 denotes any local norm of @xmath1269 ) when the point @xmath1270 tends to the reals inside @xmath1256 .
_ use of locality together with weak spectral condition : the analytic n - point functions _ in view of the locality axiom the @xmath1271 permuted wightman functions @xmath1272 , respectively analytic in the permuted tuboids @xmath1273 have boundary values @xmath1274 which coincide on the region @xmath1275 of space - like configurations @xmath1276^{n};\ , ( x_i - x_j)^2 < 0,\ , i\leq i , j\leq n\}.\ ] ] by making use of the edge - of - the - wedge theorem for tuboids ( theorem a.3 ) , one sees that for each @xmath10 there is an analytic function @xmath1277 which is the common analytic continuation of all the functions @xmath1272 in the union of the corresponding tuboids @xmath1273 together with a complex neighbourhood of @xmath1275 .
other consequences of these axioms and the status of the associated field theories as regards the classical theorems ( pct , spin - statistics , bisognano - wichmann theorem , etc . , ) will be presented in a subsequent paper .
it is a pleasure to thank d.buchholz , j. faraut , k. fredenhagen j .-
p.gazeau r.l .
stuller and g.a .
viano for useful discussions .
u. m. is a fellow of the eec -human capital and mobility program
contract n. erbchbict930675 .
= 11 addtoresetequationsection
= 12 = 11 addtoresetpropositionsection
= 12 = 11 addtoresettheoremsection
= 12 = 11 addtoresetremarksection
= 12 = 11 addtoresetdefinitionsection
= 12 = 11 addtoresetdefinitionsection
let @xmath1278 be a real @xmath10-dimensional analytic manifold , @xmath1279 the tangent bundle to @xmath1278 and @xmath1280 a complexification of @xmath1278 .
if @xmath61 is any point in @xmath1278 , @xmath1281 and @xmath1282 will denote open neighbourhoods of @xmath61 , respectively in @xmath1278 and @xmath1280 such that @xmath1281 = @xmath1283 ; a corresponding neighbourhood of @xmath1284 with basis @xmath1281 in @xmath1285 will be denoted @xmath1286 .
we call admissible local diffeomorphism at a point @xmath61 any diffeomorphism @xmath183 which maps some neighbourhood @xmath1286 of @xmath1284 in @xmath1287 onto a corresponding neighbourhood @xmath1282 of @xmath61 in @xmath1280 ( considered as a @xmath1288-dimensional @xmath25 manifold ) in such a way that the following properties hold : the conditions of definition a.1 .
are obviously invariant under any real biholomorphic mapping describing a change of chart representation for @xmath1278 and @xmath1280 .
the simplest example of admissible local diffeomorphism is obtained in a chart representation of @xmath1281 ( resp .
@xmath1282 ) by open neighbourhoods of the origin @xmath1146 ( resp . @xmath1295 in @xmath1296 ( resp .
@xmath1297 ) , such that @xmath1146 = @xmath1298 ; the identity mapping @xmath1299 from @xmath1300 onto @xmath1301 defines a local diffeomorphism ( on @xmath1302 ) which satisfies conditions a ) and b ) .
more general admissible diffeomorphisms @xmath1303 ( which map some domain @xmath1304 of @xmath1305 onto the complex domain @xmath1306 ) will allow us to introduce the appropriate definition of _ local tube_. + let @xmath1307 be an open truncated cone in @xmath1296 of the form @xmath1308 where @xmath1309 denotes a domain in the sphere @xmath1310 . usually a local tube in @xmath1311 is defined as a domain of the form @xmath1312 . in order to extend this notion to complex manifolds we shall now call _ generalized local tube _ in @xmath1311 ( near the origin ) any domain of the form @xmath1313 , where @xmath1303 is any admissible local diffeomorphism at the origin .
if the basis @xmath1309 of @xmath1307 in @xmath1310 contains a given point @xmath1314 we shall say that @xmath1313 is a _ neighbouring local tube _ of the point @xmath1315 of the sphere bundle @xmath1316 .
[ loctub ] \i ) let us consider the action of the diffeomorphism @xmath1324 ( between neighbourhoods of the origin in @xmath1325 ) in the polar representation @xmath1326 of @xmath1297 ( namely @xmath1327 , with @xmath1118 and @xmath1328 ) : in view of def .
[ a1 ] one can check that @xmath1329 defines a homeomorphism @xmath1330 between neighbourhoods of the subset @xmath1331 in @xmath1332 ( in the topology induced on @xmath1332 by @xmath1333 ) such that @xmath1334=(x,\hat{y},0)$ ] .
now , the set @xmath1335 of @xmath1297 is represented in @xmath1332 by the open set @xmath1336 , which is a neighbourhood of the point @xmath1337 in @xmath1332 ; since the homeomorphism @xmath1330 preserves this point , @xmath1338 certainly contains a neighbourhood of @xmath1337 of the form @xmath1339 in x , which proves that correspondingly @xmath1340 .
\ii ) with every diffeomorphism @xmath1322 , the previous procedure associates a homeomorphism @xmath1341 which transforms the basis of neighbourhoods @xmath1342 of the point @xmath1343 ( in @xmath1332 ) into another basis of neighbourhoods of the same point ; one can therefore find such a set @xmath1344 such that @xmath1345 which implies correspondingly @xmath1346 .
+ returning to complex manifolds , one can call _ local tube _ in @xmath1280 any domain which admits a representation by a generalized local tube in @xmath1297 in a certain well - chosen chart .
however , the more general notion of _ tuboid _ introduced below is more flexible for considering analytic functions with moderate increase and corresponding distributional boundary values on @xmath1347 .
a tuboid can be described as a domain in @xmath1280 which is bordered by the real manifold @xmath1278 and whose `` shape '' near each point of @xmath1278 is ( in the space of @xmath1348 and for @xmath1349 ) very close to a given cone @xmath1350 of the tangent space @xmath1351 to @xmath1278 at the point @xmath7 .
the following more precise definitions are needed .
we call `` profile '' above @xmath1278 any open subset @xmath1352 of @xmath1287 which is of the form @xmath1353 , where each fiber @xmath1354 is a non - empty cone with apex at the origin in @xmath1355 ( @xmath1354 can be the full tangent space @xmath1355).[a2 ] it is convenient to introduce the `` projective representation '' @xmath1356 of @xmath1287 , namely @xmath1357 , with @xmath1358 = @xmath1359 . the image of each point @xmath1360 in @xmath1358 is @xmath1361 .
each profile @xmath1352 can then be represented by an open subset @xmath1362 of @xmath1358 ( each fiber @xmath1363 being now a relatively compact set ) .
we also introduce the complement of the closure of @xmath1364 in @xmath1356 , namely the open set @xmath1365 = @xmath1366= @xmath1367 ( note that @xmath1368 ) .
a domain @xmath1369 of @xmath1370 is called a tuboid with profile @xmath1352 above @xmath1278 if it satisfies the following property . for every point @xmath61 in @xmath1347
, there exists an admissible local diffeomorphism @xmath183 at @xmath61 such that : since it makes use only of the notion of local admissible diffeomorphism , this definition is invariant under the choice of the analytic chart representation of @xmath1278 and @xmath1280 . ; moreover , it is very flexible in view of the following if @xmath1369 is a tuboid with profile @xmath1352 above @xmath1347 , it satisfies the properties a ) and b ) of definition [ a3 ] for _ all _ admissible local diffeomorphisms @xmath183 at all points of @xmath1278 .
[ pra1 ] * proof .
* the set described in a ) ( or similarly in b ) ) actually contains a neighbouring local tube of the given point @xmath1371 whose representation in @xmath1297 ( in a given chart around @xmath61 ) could be described as in lemma [ lltt ] by a set of the form @xmath1381 , @xmath1303 being the representation of @xmath183 in the considered chart .
it then follows from part ii ) of lemma [ lltt ] that for any other admissible diffeomorphism @xmath1382 ( at @xmath61 ) there exists another neighbouring local tube of @xmath1371 represented in the chart by @xmath1383 and such that @xmath1384 . the set inclusion ( resp .
identity ) expressed by property a ) ( resp .
b ) ) will then be a fortiori satisfied by this new neighbouring local tube defined with the help of @xmath1382 .
+ a useful application of proposition [ pra1 ] is obtained by considering local diffeomorphisms @xmath183 which admit a privileged linear representation of the form @xmath1385 in a certain system of coordinates . in such variables ,
a tuboid @xmath1369 is represented _ locally _
( i.e. near a point @xmath61 ) as a tuboid @xmath1386 in @xmath1297 whose profile @xmath1387 ( in @xmath1305 ) satisfies the following properties
. in @xcite a slightly more restrictive definition of tuboids in @xmath1297 was given , in which condition ii ) was replaced by the following one ( non - invariant under real biholomorphic mappings ) : ii)@xmath1396@xmath1397 ( the condition i ) being valid as before ) .
we can however say that the domains introduced in @xcite form a `` basis of tuboids '' of @xmath1297 , since every tuboid @xmath1386 in @xmath1297 ( defined here ) contains a tuboid @xmath1398 of that special family having the same profile as @xmath1386 ( namely @xmath1399 ) .
properties of pseudoconvexity of tuboids which generalize in some sense the standard `` tube theorem '' ( see e.g. @xcite ) have been derived in @xcite ; in view of the previous remarks , these results can be readily applied to the general class of tuboids in @xmath1297 considered here ; they can be applied as well ( due to their coordinate - independent character ) to the more general class of tuboids on manifolds introduced above .
these properties can be summarized in the following \1 ) a necessary condition for a tuboid @xmath1369 with profile @xmath1352 above @xmath1347 to be a pseudoconvex domain ( or holomorphy domain ) in @xmath1280 is that each fiber @xmath1350 of @xmath1352 is an open _ convex _ cone in @xmath1351 ( possibly equal to the whole space @xmath1351 ) .
\2 ) if a tuboid @xmath1369 has a profile @xmath1352
whose all fibers @xmath1350 are convex , then , for each domain @xmath1400 of @xmath1278 described by a chart in @xmath1401 , there always exists a tuboid @xmath1402 above @xmath1400 contained in @xmath1369 with profile @xmath1403 which is a holomorphy domain in @xmath1280 we wish to give a generalization for tuboids on manifolds of the notion of distribution boundary value of holomorphic functions with moderate growth near the reals .
we first recall the following standard result for local tubes in @xmath1414 @xmath1415sw@xmath1416 @xmath1415w@xmath1416 .
there is an equivalence between the following two properties of a function @xmath1417 analytic in a given local tube @xmath1418 ( with e.g. @xmath1419 convex ) .
\i ) @xmath1420 is _ of moderate growth near the real _ in the following sense : in each local tube @xmath1421 such that @xmath1422 and @xmath1423 @xmath1424 @xmath1425 denoting the bases of @xmath1426 and @xmath1427 on the sphere @xmath1428 there exists an integer @xmath1429 and a constant @xmath1430 ( depending on @xmath1431 such that : @xmath1432 \ii ) the sequence of distributions @xmath1433 defined in @xmath1434 by the equation : @xmath1435 ( for all @xmath1436 in @xmath1437 is weakly convergent for @xmath1438 tending to 0 in @xmath1439 this defines a distribution @xmath1440 in @xmath1441 called the boundary value of @xmath1442 on @xmath1237 from the local tube @xmath1443 moreover , if the growth order @xmath1444 of @xmath1442 ( in i ) ) does not depend on @xmath1445 the boundary value @xmath1446 of @xmath1420 is a distribution of finite order ( linked to @xmath1447 in @xmath1448 and an alternative characterization of @xmath1442 ( which in fact makes the connection between i ) and ii ) ) is the following one : \iii ) there exists a holomorphic function @xmath1449 which is defined in @xmath1450 and uniformly bounded in the closure of each subdomain @xmath1451 ( defined in i ) ) , and an integer @xmath1452 such that @xmath1453 where @xmath1454 denotes a first order derivative operator with real constant coefficients ( note that a relevant property of the vector @xmath1455 is that it belongs to the basis @xmath1426 of @xmath1456 moreover , the corresponding relation @xmath1457 holds in the sense of distributions on @xmath1448 with @xmath1458 and @xmath1459 the latter being identified with the usual limit of @xmath1460 on @xmath1448 as a continuous function .
we now consider the general case of tuboids @xmath1461 with profile @xmath1462 in view of theorem a.1 , @xmath1463 can always be assumed to have _ exclusively convex fibers .
our generalization of the previous notions requires two steps : we consider successively tuboids in @xmath1414 and tuboids in manifolds . in view of definition a.3 and proposition
a.1 applied to tuboids in @xmath1468 ( with the choice @xmath1469 one can say that @xmath1470 where each domain @xmath1471 is a neighbouring local tube of _ the point _
@xmath1472 of @xmath1352 ( with @xmath1473 . for a holomorphic function @xmath1420 in @xmath1474
it is therefore meaningful to keep the previous definition of the _ property of moderate growth near the real _ ( see i ) ) , the uniform majorizations of the form ( a.2 ) being now required to hold in all the local tubes @xmath1475 ( thus implementing all the directions of the profile @xmath1463 of @xmath1456 similarly , we shall say that a holomorphic function in @xmath1461 _ admits a distribution boundary value on @xmath1465 from @xmath1461 _ if the restriction of @xmath1476 to each local tube @xmath1475 satisfies the previous property ii ) for functions @xmath1436 in the corresponding space @xmath1477 if this holds for all @xmath1478 it is clear that the resulting local distributions @xmath1479 satisfy consistency conditions in all the two - by - two intersections of the family of open sets @xmath1480 since this family forms a covering of @xmath1481 the set of distributions @xmath1482 therefore defines a unique distribution @xmath1446 on @xmath1465 which coincides with each of them on the corresponding open sets @xmath1483 \b ) since a tuboid in a manifold @xmath1484 can be represented as an atlas of tuboids in @xmath1414 satisfying consistency conditions in the two - by - two intersections of charts of @xmath1485 the theory of distribution boundary values of holomorphic functions from a tuboid in @xmath1484 should reduce to the theory in @xmath1414 ( given above in a ) ) provided the conservation of the previous notions under ( real ) biholomorphic mappings be established .
we shall now examine this point .
it is sufficient to consider in @xmath1414 a biholomorphic mapping @xmath1486 which transforms a tuboid @xmath1461 above @xmath1481 with profile @xmath1463 onto a tuboid @xmath1487 above @xmath1488 with profile @xmath1489 ; @xmath1465 and @xmath1490 are _ bounded _ domains of @xmath1491 such that @xmath1492 and ( in the corresponding tangent spaces ) @xmath1493 let @xmath1442 be a holomorphic function in @xmath1474 its image @xmath1494 defined by @xmath1495 being holomorphic in @xmath1496 at first , it is easy to check the invariance of property i ) under the mapping @xmath1497 namely , @xmath1420 is of moderate growth near the real in @xmath1461 iff the same property holds for @xmath1498 in @xmath1499 moreover , if there is a uniform growth order @xmath1500 it is preserved by @xmath1501 and @xmath1502 this follows from two simple facts : the first one is that near any real pair @xmath1503 in @xmath1504 any complex pair @xmath1505 is such that @xmath1506 with @xmath1507 on the bounded set @xmath1508 therefore inequalities of the form ( a.2 ) are preserved under the mappings @xmath1501 and @xmath1502 the second fact is that ( in view of proposition a.1 ) any local tube @xmath1509 in @xmath1487 ( neighbouring a profile element @xmath1510 has an inverse image @xmath1511 $ ] in @xmath1461 which contains a neighbouring local tube @xmath1475 of the profile element @xmath1512 @xmath1513 @xmath1514 @xmath1515 in view of the equivalence of properties i ) and ii ) for tuboids in @xmath1468 ( see a ) ) , we can then say that the moderate growth property of @xmath1420 and @xmath1517 in their respective tuboids @xmath1461 and @xmath1487 is equivalent to the fact that both @xmath1420 and @xmath1498 admit distribution boundary values @xmath1518 and @xmath1519 respectively in @xmath1520 , @xmath1521 from the tuboids @xmath1461 and @xmath1522 the remaining point to establish is that @xmath1523 is the relevant transform @xmath1524 of the distribution @xmath1446 under the ( real @xmath1525-mapping @xmath1526 according to what is needed in the theory of distributions on @xmath1527-manifolds ( see e.g. @xcite ) ; this transformation law is : @xmath1528 ( the transformation @xmath1529 being a continuous mapping from @xmath1530 to @xmath1531 this law coincides with the usual one when @xmath1446 is a continuous function @xmath1532 in order to show that @xmath1533 we shall make use of the property iii ) of the function @xmath1498 which we apply in any given local tube @xmath1534 in @xmath1499 let us assume for simplicity that @xmath1535 so that @xmath1536 where @xmath1537 is holomorphic in @xmath1509 and continuous in the closure of this set ; correspondingly , there holds the following relation for the boundary values in @xmath1538 @xmath1539 where @xmath1540 is continuous on @xmath1541 by taking the boundary values of both sides of eq.(a.6 ) on @xmath1548 one obtains the corresponding relation in @xmath1549 @xmath1550 where @xmath1551 is continuous on @xmath1552 and such that @xmath1553 now , since the partial derivatives of distributions are transformed by the usual rule under the mapping @xmath1501 ( see e.g. @xcite ) , the image by @xmath1554 of the r.h.s . of eq.(a.7 )
is equal to the r.h.s . of eq.(a.5 ) and therefore @xmath1555 since this holds for every pair of elementary local tubes @xmath1556 @xmath1557 in @xmath1558 this proves the desired identity @xmath1559 in @xmath1560 in the general case , the use of property iii ) for the function @xmath1498 will involve a derivation order @xmath1452 @xmath1561 the same argument as above can then be extended in an obvious way by recursion over @xmath1562 let @xmath1420 be a holomorphic function in a tuboid @xmath1461 of a complex @xmath1209-dimensional manifold @xmath1563 the following properties are equivalent and independent of the local systems of complex variables , chosen to represent any given local part of @xmath1461 as a tuboid in @xmath1564 \iii ) in each local chart ( of a sufficiently refined atlas ) , @xmath1420 admits a ( non - unique ) representation of the form @xmath1567 with @xmath1460 holomorphic in the restriction of @xmath1369 to this local chart and continuous on the real .
close relations exist between the local order @xmath1444 of moderate growth of @xmath1568 in i ) , the local order of the distribution @xmath1446 in ii ) and the derivation order @xmath1452 in the local representation of @xmath1420 in iii ) .
the previous notions lead to the following form of the edge - of - the - wedge property for tuboids on complex manifolds and more generally to the theory of singular spectrum ( or essential support ) of distributions on analytic manifolds .
( edge - of - the - wedge for tuboids ) .
let @xmath1569 @xmath1570 be two tuboids above @xmath1571 in @xmath1485 with respective profiles @xmath1572 @xmath1573 in @xmath1574 then there exists a tuboid @xmath1461 containing @xmath1575 whose profile @xmath1576 is such that each fiber @xmath1577 is the convex hull of @xmath1578 and which enjoys the following property . for every pair of functions @xmath1579 @xmath1580 respectively
holomorphic in @xmath1569 @xmath1570 and admitting distribution boundary values on @xmath1571 such that @xmath1581 there exists a unique holomorphic function @xmath1420 such that @xmath1582 @xmath1583 which is defined in @xmath1461 and such that @xmath1584 ( from @xmath1585 ) .
the proof of the latter theorem is a direct application of the edge - of - the - wedge theorem in the general situation of oblique local tubes ( @xmath1415e@xmath1416 , @xmath1415b.i@xmath1416 ) which one applies to any pair of neighbouring local tubes @xmath1586 of arbitrary profile elements @xmath1587 ( represented appropriately in arbitrarily chosen local charts of @xmath1588 99 allen , b. : phys .
d * 32 * , 3136 , ( 1985 ) .
araki , h. : j. math .
phys . * 5 * , 1 ( 1964 ) .
artin , e. : geometric algebra , interscience , new york ( 1957 ) .
bateman , h. : higher transcendental functions , vol .
i , mc graw - hill , new york ( 1954 ) .
bertola , m. , gorini , v. : private communication .
birrell , n.d . ,
davies , p.c.w . : quantum fields in curved space , cambridge university press , cambridge 1982 .
bisognano , j.j . ,
wichmann , e.h . : j. math
* 16 * , 985 ( 1975 ) borchers , h.j .
: nuovo cimento * 24 * , 214 ( 1963 ) . bros j. : nucl.phys .
( proc . suppl . )
* 18 b * , 22 ( 1990 ) .
bros , j. , buchholz , d. : fields at finite temperature : a general study of two - point functions , in preparation .
bros j. , gazeau j .-
, moschella u. : phys .
lett . , * 73 * , 1746 ( 1994 ) .
bros , j. , iagolnitzer , d. : ann .
fourier * 26 * , 49 ( 1976 )
. bros , j. , iagolnitzer , d. : ann .
poincar * 18 * , 147 ( 1973 ) ; iagolnitzer , d. : analytic structure of distribution and essential support theory , in `` structural analysis of collision amplitudes '' balian , r. and iagolnitzer , d. eds . , north - holland , amsterdam ( 1976 ) .
bros j. , moschella u. : an axiomatic approach to de sitter quantum field theory , in preparation .
bros j. , moschella u. : fourier transform on a one - sheeted hyperboloid , in preparation .
bros j. , viano g.a . :
connection between the harmonic analysis on the sphere and the harmonic analysis on the one - sheeted hyperboloid : an analytic continuation viewpoint .
saclay preprint ( 1993 ) .
forum mathematicum ( to appear ) .
bros j. , viano g.a . : bull .
france * 120 * , 169 ( 1992 ) .
bunch t.s . , davies p.c.w .
a * 360 * , 117 ( 1978 ) .
choquet - bruhat , y. , de witt - morette c. , dillard - bleick , m. : analysis , manifolds and physics , revised edition , elsevier , new york 1982 .
epstein , h. : j. math .
* 1 * , 524 , ( 1960 ) .
epstein h. , glaser , v. : ann .
h. poincar * 19 * , 211 ( 1973 ) .
faraut , j. : sminaire danalyse harmonique de tunis , expos 29 ( 1981 ) . faraut j. : noyaux sphriques sur un hyperbolode une nappe in lect .
notes in math . * 497 * , springer - verlag , berlin ( 1975 ) .
faraut j. , viano g.a . : j. math .
* 27 * , 840 ( 1986 ) .
figari , r. , hegh - krohn , r. , nappi , c.r .
phys . * 44 * , 265 ( 1975 ) .
fredenhagen , k. : quantum field theory on non - trivial space - times , in mathematical physics towards the 21@xmath1594 century , r.n . sen and a.
gersten eds .
ben - gurion univ .
negev press , beer - sheva ( 1993 ) .
froissart , m. : phys . rev . * 123 * , 1053 ( 1961 ) .
gelfand , i.m . , shilov ,
e.g. : generalized functions vol .
i , academic press , new york ( 1964 ) . gibbons g.w . , hawking s.w .
d * 10 * , 2378 ( 1977 ) .
gribov , v.n . : j. exp .
* 14 * , 1395 ( 1962 ) .
haag , r. : local quantum physics , fields , particles , algebras , springer , berlin 1992 .
hrmander , l. : acta math .
* 127 * ( 1971 ) , 79 .
duistermaat , j.j . , and hrmander , l. : acta math .
* 128 * ( 1972 ) , 183 .
kllen , g. , wightman , a.s .
danske videnskab .
selskab , mat .- fys .
skrifter 1 , no .
6 . kay , b.s . , wald , r. : phys .
* 207 * , 49 ( 1991 ) .
jost , r. : the general theory of quantized fields a.m.s . , providence 1965 .
isham c.j . : in proceedings of the bonn conference on differential geometrical methods in mathematical physics , eds .
k. bleuler , h.r . petry and a. reetz , springer verlag , new york 1978 .
landsman , n.p . , van veert , t.g .
: phys . rep . *
145 * , 142 ( 1987 ) .
linde , a. : particle physics and inflationary cosmology , harwood academic publishers , chur ( 1990 ) .
mickelssonn j. , niederle j. : commun .
* 27 * , 167 ( 1972 ) .
molchanov s. : math .
* 7 * , 1553 ( 1966 ) .
moschella u. : quantization curvature and temperature : the de sitter space - time , in quantization and infinite - dimensional systems , j - p .
antoine , s. twareque ali , w. lisiecki , i.m .
mladenov , a. odzijewicz eds . ,
p. 183 plenum , new york ( 1994 ) .
pham , f. : introduction ltude topologique des singularits de landau , mmorial des sciences mathmatiques * 164 * , gauthier - villars , paris ( 1967 ) .
radzikowski , m .- j .
, _ the hadamard condition and kay s conjecture in ( axiomatic ) quantum field theory on curved space - time _ , ph.d .
thesis , princeton university , october 1992 .
redmount , i.h . ,
takagi , s. : phys .
d 37 * , 1443 ( 1988 ) .
schomblond , c. , spindel , p. : ann .
h. poincar * 25a * , 67 ( 1976 ) .
sewell , g. l. : ann .
phys . * 141 * , 201 ( 1982 ) .
streater , r.f . ,
wightman , a.s .
: pct , spin and statistics , and all that w.a .
benjamin , new york 1964 .
unruh , w.g .
d * 14 * , 870 ( 1976 ) .
verch , r. : lett .
phys * 29 * , 143 ( 1993 ) . vilenkin n. ja . : fonctions spciales et thorie de la reprsentation des groupes , dunod , paris ( 1969 ) .
watson : a treatise on the theory of bessel functions , cambridge university press , cambridge ( 1958 ) .
wightman , a.s . : introduction to some aspects of quantized fields , in lecture notes , cargse summer school , gordon and breach , new york ( 1964 ) .
wightman , a.s . : analytic functions of several complex variables in relations de dispersion et particules lmentaires , c. de witt and r. omnes eds .
hermann , paris ( 1960 ) .
wyrozumski , t. : class .
. grav . * 5 * , 1607 ( 1988 ) . | we present a theory of general two - point functions and of generalized free fields in @xmath0-dimensional de sitter space - time which closely parallels the corresponding minkowskian theory .
the usual spectral condition is now replaced by a certain _ geodesic spectral condition _ , equivalent to a precise thermal characterization of the corresponding `` vacuum''states .
our method is based on the geometry of the complex de sitter space - time and on the introduction of a class of holomorphic functions on this manifold , called _ perikernels _ , which reproduce mutatis mutandis the structural properties of the two - point correlation functions of the minkowskian quantum field theory .
the theory contains as basic elementary case the linear massive field models in their `` preferred '' representation .
the latter are described by the introduction of _ de sitter plane waves _ in their tube domains which lead to a new integral representation of the two - point functions and to a fourier - laplace type transformation on the hyperboloid .
the hilbert space structure of these theories is then analysed by using this transformation . in particular
we show the reeh - schlieder property . for general two - point functions , a substitute to the wick rotation
is defined both in complex space - time and in the complex mass variable , and substantial results concerning the derivation of kllen - lehmann type representation are obtained . -30pt
= eufm7 at 12pt = eufm7 at 9pt = 11 addtoresetequationsection = 12 = 11 addtoresetpropositionsection
= 12 = 11 addtoresettheoremsection = 12 = 11 addtoresetremarksection
= 12 = 11 addtoresetdefinitionsection
= 12 = 11 addtoresetlemmasection
= 12 = 11 addtoresetcorollarysection
= 12 epsf mssymb |
a wealth of experimental data on hadronic charmless @xmath0 decays has become available from the babar and belle experiments .
these studies of the numerous @xmath0 decay channels are designed to test the cabbibo - kobayashi - maskawa ( ckm ) explanation of @xmath1 violation in the standard model as represented by the unitarity triangle condition @xmath2 on the ckm mixing parameters .
such tests have been facilitated by the recent significant progress in the theoretical understanding of hadronic decay amplitudes based upon qcd factorization which allows the amplitudes to be expressed in terms of numerous soft qcd parameters , such as meson decay constants and transition form factors , and a set of calculable coefficients . in this paper
we present an analysis of recent data , based upon qcd factorization .
we also investigate the potential contribution to the decay amplitudes of @xmath3 quark annihilation and so - called charming penguins .
the data that we attempt to fit includes decays to pseudoscalar ( @xmath4 and @xmath5 ) and vector ( @xmath6 and @xmath7 ) mesons .
this is an extension of an earlier study @xcite that was based upon simplified formulae derived from the heavy quark limit of qcd factorization .
the starting point for the calculation of all @xmath0 meson decay amplitudes is the effective hamiltonian @xmath8 + \text{other terms}\ ] ] where @xmath9 is a product of ckm matrix elements ,
@xmath10 and the local @xmath11 four - quark operators are @xmath12 where @xmath13 , @xmath14 and @xmath15 are color indices , @xmath16 for @xmath17-type quarks and we use the notation , for example , @xmath18 \sum_{q^{\prime}}\left[\bar{q}^{\prime}_{\beta}\gamma^{\mu } ( 1+\gamma_{5 } ) q^{\prime}_{\beta}\right].\ ] ] @xmath19 are the current - current tree operators , @xmath20 are qcd penguin operators and @xmath21 are electroweak penguin operators .
the other terms indicated in ( [ heff ] ) include the electromagnetic and chromomagnetic dipole transition operators . in the standard model the contributions from the electroweak penguin operators and magnetic dipole operators are generally small .
the exception is the electroweak penguin operator coefficient @xmath22 which is larger than the qcd penguin coefficients @xmath23 and @xmath24 .
it has been shown by beneke _
_ @xcite that , in the heavy quark limit , the hadronic matrix elements of the four - quark operators in the amplitudes for non - leptonic @xmath0 decays into two light mesons @xmath25 have the form @xmath26\ ] ] and can be calculated from first principles , including non - factorizable strong interaction corrections .
qcd factorization extends naive factorization by separating matrix elements into short distance contributions at scale @xmath27 that are perturbatively calculable and long distance contributions @xmath28 that are parameterized .
@xmath0 meson decay can also be initiated by @xmath3 quark annihilation with its partner .
although the annihilation contributions to the decay amplitude are formally of @xmath29 and power suppressed , they violate qcd factorization because of end point divergences . however these weak annihilation contributions can be included into the decay amplitudes by treating the end point divergences as phenomenological parameters .
analyses based upon qcd factorization with inclusion of weak annihilation have been undertaken for @xmath30 @xcite and @xmath31 @xcite .
although general agreement with experiment was found , some branching ratios for @xmath32 decays were only marginally consistent and all predictions were plagued by the large uncertainties associated with poorly determined parameters within the theory
. a recent global analysis of @xmath33 and @xmath32 decays @xcite found that qcd factorization plus weak annihilation could fit many decay channels but yielded results too low for the @xmath34 decays .
more recently , aleksan _
@xcite have undertaken a global analysis of @xmath32 decays and have concluded that qcd factorization can not fit the experimental data when decay channels involving @xmath35 mesons are included , and suggest that this failure is due to some larger than expected nonperturbative contribution .
motivated by the concept of so - called charming penguins , that is non - perturbative @xmath29 corrections from enhanced @xmath36-loop penguins , first introduced by ciuchini _
@xcite , they introduce additional long - distance contributions to the decay amplitudes and include the two complex parameters from these additional amplitudes in their global fit .
they obtain a slightly better fit but their best - fit parameters are at the limits of the allowed domain .
a recent study @xcite , limited to @xmath37 decays , has used qcd factorization to place bounds on the @xmath38 form factor .
a detailed study of qcd factorization applied to @xmath33 and @xmath32 decays of @xmath0 mesons has just been completed by beneke and neubert @xcite .
predicted branching ratios and @xmath1 asymmetries for a large number of @xmath33 and @xmath32 channels are given for default values of input parameters and detailed estimates of the theoretical uncertainties in these predictions determined for various scenarios of input parameters .
beneke and neubert find that there is a scenario for which there is general global agreement between the results of qcd factorization and measurement except for @xmath39 and the group of @xmath34 decays .
in this paper we undertake a global analysis of 18 @xmath33 , @xmath32 and @xmath40 channels using two theoretical models , ( i ) qcd factorization with inclusion of weak annihilation and ( ii ) this model supplemented with charming penguins .
our study is similar in spirit to that of aleksan _ et al . _
@xcite for @xmath32 decays but we extend the global fit to include @xmath33 and some @xmath40 decays . this paper is organized as follows . in sec .
ii we review the structure of the decay amplitude within qcd factorization and discuss the various parameters that occur in this amplitude .
inclusion of weak annihilation and charming penguins is discussed in sec .
iii and sec .
iv respectively .
the method and results of our best fit for our two models to current experimental data is presented in sec v , and sec .
vi contains our discussion and conclusions .
in qcd factorization , the amplitude for @xmath0 decay into two light hadrons ( mesons ) @xmath25 has the form , neglecting weak annihilation processes , @xmath41 \nonumber \\ * & & + \sum_{p = u , c}\sum_{i=3,\ldots , 6,9 } \lambda_{p}a^{p}_{i } [ t_{i}(m_{1},m_{2 } ) + t_{i}(m_{2},m_{1 } ) ] \}\end{aligned}\ ] ] where @xmath42 the two - quark matrix elements are the well determined electroweak decay constants @xmath43 e.t.c . and the @xmath0 transition form factors @xmath44 e.t.c . in principle transition form factors
are independently measurable in @xmath0 semi - leptonic decays but to date they are only loosely constrained by measurements and model estimations .
the coefficients @xmath45 have to be calculated from the wilson coefficients @xcite .
we have calculated the wilson coefficients at several scales @xmath46 of @xmath47 using the next - to - leading - order ( nlo ) renormalization group equations @xmath48 we follow the beneke _
@xcite prescription of ( i ) dropping terms of @xmath49 , @xmath50 and @xmath51 in ( [ rges ] ) , ( ii ) neglecting the effect of the electromagnetic penguins @xmath52 on the evolution of the qcd penguin coefficients @xmath53 and , ( iii ) in @xmath54 , splitting the @xmath55 electroweak penguin terms into those enhanced by large @xmath56 or @xmath57 , which are treated as leading order ( lo ) , and treating the remainder , together with the @xmath58 terms , as nlo .
our calculated values are shown in tables [ clo ] and [ cnlo ] and are very similar to those of beneke _ et al .
_ @xcite and du _ et al .
_ @xcite .
ldddd scale ( gev ) & 4.20 & 2.10 & 1.45 & 1.02 + @xmath59 & 1.1174 & 1.1848 & 1.2392 & 1.3120 + @xmath60 & -0.2678 & -0.3873 & -0.4755 & -0.5862 + @xmath23 & 0.0121 & 0.0185 & 0.0235 & 0.0299 + @xmath61 & -0.0274 & -0.0383 & -0.0459 & -0.0551 + @xmath24 & 0.0080 & 0.0105 & 0.0120 & 0.0136 + @xmath62 & -0.0341 & -0.0526 & -0.0677 & -0.0883 + @xmath63 & -0.0140 & -0.0282 & -0.0314 & -0.0303 + @xmath64 & 0.0288 & 0.0432 & 0.0555 & 0.0734 + @xmath65 & -1.2913 & -1.3785 & -1.4408 & -1.5190 + @xmath66 & 0.2888 & 0.4213 & 0.5194 & 0.6424 + ldddd scale ( gev ) & 4.20 & 2.10 & 1.45 & 1.02 + @xmath59 & 1.0813 & 1.1374 & 1.1820 & 1.2405 + @xmath60 & -0.1903 & -0.2948 & -0.3700 & -0.4619 + @xmath23 & 0.0137 & 0.0212 & 0.0274 & 0.0358 + @xmath61 & -0.0357 & -0.0506 & -0.0618 & -0.0762 + @xmath24 & 0.0087 & 0.0102 & 0.0105 & 0.0096 + @xmath62 & -0.0419 & -0.0653 & -0.0854 & -0.1146 + @xmath63 & -0.0026 & -0.0139 & -0.0147 & -0.0100 + @xmath64 & 0.0618 & 0.0986 & 0.1317 & 0.1813 + @xmath65 & -1.2423 & -1.3181 & -1.3526 & -1.4149 + @xmath66 & 0.2283 & 0.3388 & 0.4015 & 0.4991 + to lowest order in the strong coupling constant @xmath67 , the @xmath45 coefficients are the same as in naive factorization , that is @xmath68 where @xmath69 and @xmath70 is the number of quark colors .
the higher order corrections include lowest order gluon exchange between the quarks in the basic tree amplitudes which are calculated and folded into the light cone distribution functions @xmath71 of the participating mesons ( see , for example , @xcite ) .
this results in the @xmath45 coefficients having the form @xmath72 where @xmath73 is the recoil meson containing the spectator ( anti ) quark and @xmath74 is the emitted meson .
the complex quantities @xmath75 describe the formation of @xmath74 , including nonfactorizable corrections from hard gluon exchange or light quark loops in penguins .
they do not involve the hard gluon exchanges with the spectator quark , these are described by the ( possibly ) complex quantities @xmath76 . in the correction terms the leading - twist light cone distribution functions for both pseudoscalar and
vector mesons are expanded in the first few terms of a gegenbauer expansion @xmath77.\ ] ] the asymptotic limit @xmath78 is valid for the mass scale @xmath79 .
the parameters @xmath80 are anticipated to be small but they are not well established . to economize in the number of fitting parameters in the initial fits to data presented here they are taken to be zero .
with this simplification all light mesons included in our analysis have the same spatial wavefunction and all coefficients @xmath75 except @xmath81 are the same for all decays .
formulae for the evaluation of the @xmath75 can be found in @xcite . although beneke and neubert @xcite obtain a different expression for @xmath82 to that of du _
@xcite , this is not important here as @xmath83 does not occur in the decay amplitudes for @xmath84 since @xmath85 . the @xmath58 corrections to @xmath86 include contributions from one - loop vertex corrections @xmath87 to all @xmath75 and from penguin corrections @xmath88 , involving the quark mass ratio @xmath89 , to @xmath90 and @xmath81 .
we neglect the small electroweak penguin corrections to @xmath91 .
typical parton off - shellness in the loop diagrams contributing to @xmath92 and @xmath91 is @xmath47 , suggesting that the scale @xmath93 should be used in evaluating @xmath75 .
the results of our calculations of @xmath75 coefficients without light cone corrections are given in table [ aii0 ] .
results including the light cone corrections @xmath94 taken from @xcite are given in table [ aii1 ] .
the coefficients most affected are @xmath95 and @xmath90 .
cdd @xmath46(gev ) & 4.20 & 2.10 + @xmath96 & 1.0572 + 0.0200i & 1.0791 + 0.0369i + @xmath97 & 0.0060 - 0.0836i & -0.0377 - 0.1130i + @xmath98 & 0.0058 + 0.0021i & 0.0083 + 0.0037i + @xmath99 & -0.0312 - 0.0161i & -0.0338 - 0.0205i + @xmath100 & -0.0369 - 0.0068i & -0.0415 - 0.0079i + @xmath101 & -0.0070 - 0.0026i & -0.0106 - 0.0050i + @xmath102 & -0.0433 - 0.0152i & -0.0586 - 0.0188i + @xmath103 & -0.0465 - 0.0056i & -0.0630 - 0.0056i + @xmath104 & -0.0075 - 0.0007i & -0.0094 - 0.0013i + @xmath105 & 0.0009 - 0.0115i & 0.0019 + 0.0152i + @xmath106 & -0.0094 - 0.0002i & -0.0097 - 0.0003i + cdddd & & & & + @xmath96 & 1.0809 + 0.0369i & 1.0762 + 0.0432i & 1.0798 + 0.0369i & 1.0753 + 0.0439i + @xmath97 & -0.0432 - 0.1130i & -0.0290 - 0.1322i & -0.0400 - 0.1130i & -0.0261 - 0.1344i + @xmath98 & 0.0085 + 0.0037i & 0.0080 + 0.0043i & 0.0084 + 0.0037i & 0.0079 + 0.0043i + @xmath99 & -0.0306 - 0.0206i & -0.0313 - 0.0210i & -0.0325 - 0.0206i & -0.0321 - 0.0210i + @xmath100 & -0.0343 - 0.0072i & -0.0364 - 0.0061i & -0.0386 - 0.0076i & -0.0383 - 0.0061i + @xmath101 & -0.0108 - 0.0050i & -0.0102 - 0.0059i & -0.0107 - 0.0050i & -0.0101 - 0.0060i + @xmath107 & -0.0586 - 0.0188i & -0.0586 - 0.0188i & -0.0097 - 0.0012i & -0.0102 - 0.0013i + @xmath108 & -0.0630 - 0.0058i & -0.0630 - 0.0058i & 0.0022 + 0.0147i & 0.0032 + 0.0152i + @xmath106 & -0.0097 - 0.0003i & -0.0097 - 0.0004i & -0.0097 - 0.0003i & -0.0097 - 0.0004i + the coefficients @xmath76 are not universal even when light cone corrections are neglected .
they contain not only the low energy parameters ( decay constants and form factors ) from the lowest order calculations , in common with naive factorization , but low energy contributions to the folding integrals involving another non - perturbative complex parameter @xmath109 which is only loosely constrained by model estimations . to discuss the form of the @xmath76
, we first note that , from ( [ heff ] ) , the contributions of the @xmath45 coefficients to the decay amplitude for @xmath110 are of the form @xmath111a_{i},\ ] ] where @xmath112 and @xmath113 are products of clebsch - gordan coefficients tablulated in , for example , @xcite . using the formulae of @xcite
, we can write , with @xmath70 , @xmath114 where @xmath115 , and @xmath116 .\end{aligned}\ ] ] the chiral factors @xmath117 are zero for @xmath73 a vector meson and are @xmath118 for the pseudoscalar mesons .
it should be noted that these @xmath76 contributions to the decay amplitudes are independent of the @xmath0 transition form factors .
however , they do involve the poorly determined parameter @xmath119 where @xmath120 is the @xmath0 leptonic decay constant and @xmath121 gev is related to the @xmath0 light cone distribution function .
the parameter @xmath122 is the contribution of a logarithmic end - point divergence in the integration over the @xmath73 light cone distribution function @xmath123 these functions take no account of the internal quark transverse momenta which , if included , would make the integrals finite but not calculable within perturbative qcd .
@xmath122 is parameterized as @xmath124 where @xmath125 is not expected to be larger than 3 .
we take @xmath126 .
the energies involved in the calculation of @xmath76 imply that the appropriate scale is not that of the scale @xmath46 used in calculating the @xmath75 but @xmath127 where @xcite @xmath128 gev .
we use this with @xmath129 in our fitting so that @xmath130 .
we note that substantial light cone corrections @xmath94 can significantly enhance the @xmath76 coefficients .
because of the heavy @xmath3 quark mass it is expected that perturbative qcd calculations will give a reliable estimate of the annihilation contribution to the decay amplitude . in these calculations the basic perturbative quark amplitudes are again folded into the participating meson light cone distribution functions . apart from the low energy regions of the folding integrals the only low energy parameters that appear in the lowest order calculations are the participating meson electroweak decay constants @xmath131 e.t.c . again the low energy contributions to the integral introduce another non - perturbative complex parameter @xmath132 , which is only loosely constrained by the model estimations .
detailed formulae are to be found in @xcite . in this paper
we follow the more extensive calculations of du _
@xcite but express the annihilation contribution to the decay amplitude in the form @xmath133a^{i}_{1 } -\lambda_{t}\{d_{3}[c_{3}a^{i}_{1}+c_{5}a^{i}_{3 } ] \nonumber \\ * & & + [ d_{4}c_{4}+d_{6}c_{6}]a^{i}_{1 } + d_{5}[c_{5}+n_{c}c_{6}]a^{f}_{3}\}\end{aligned}\ ] ] where @xmath134 and @xmath135 .
the quantities @xmath136 , where the superscript @xmath137 denotes gluon emission from initial ( final ) state quarks , result from folding the quark amplitudes into the meson distribution functions .
if the asymptotic form of ( [ gegen ] ) is used then @xcite @xmath138 , \nonumber \\ a^{i}_{3}(p_{1},p_{2})&= & 6 \pi \alpha_{s}(r^{p_{1}}_{\chi}- r^{p_{2}}_{\chi})(x^{2}_{a}-2x_{a}+\frac{\pi^{2}}{3 } ) , \nonumber \\ a^{f}_{3}(p_{1},p_{2 } ) & = & 6 \pi \alpha_{s } ( r^{p_{1}}_{\chi}+r^{p_{2}}_{\chi})(2x_{a}^{2}-x_{a } ) , \nonumber \\
a^{i}_{1}(p , v)&= & \pi \alpha_{s } \left[18\left(x_{a}-4+\frac{\pi^{2}}{3}\right)\right ] , \nonumber \\
a^{i}_{3}(p , v ) & = & \pi \alpha_{s}r^{p}_{\chi}[2\pi^{2}+6(x_{a}^{2}-2x_{a } ) ] , \nonumber \\ a^{f}_{3}(p , v)&= & 6 \pi \alpha_{s } r^{p}_{\chi}(2x_{a}^{2}-x_{a } ) , \nonumber \\ a^{i}_{1}(v_{1},v_{2 } ) & = & a^{i}_{1}(p , v ) , \nonumber \\ a^{i}_{3}(v_{1},v_{2 } ) & = & a^{f}_{3}(v_{1},v_{2})=0.\end{aligned}\ ] ] the coefficients @xmath139 are clebsch - gordan type factors and are given in table [ dcofs ] for the particle sign conventions of @xcite .
note that , in using ( [ annih ] ) with table [ dcofs ] , there is no need to distinguish between @xmath32 and @xmath140 decays .
result for the unitarity triangle fit in the bbns model .
the shaded area shows the @xmath141 allowed region for the apex of the unitarity triangle .
the data point shows the fit result from the unitarity triangle fit from other measurements taken from _ _ ckmfitter__@xcite . ]
result for the unitarity triangle fit in the charming penguin model .
the shaded area shows the @xmath141 allowed region for the apex of the unitarity triangle .
the data point shows the fit result from the unitarity triangle fit from other measurements taken from _
_ ckmfitter__@xcite . ]
1 w. n. cottingham , i. b. whittingham , n. de groot and f. wilson , j. phys .
g : nucl . part .
* 28 * , 2843 ( 2002 ) .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , nuc .
b * 606 * , 245 ( 2001 ) .
d. du , h. gong , y. sun , d. yang and g. zhu , phys .
d * 65 * , 074001 ( 2002 ) .
d. du , h. gong , y. sun , d. yang and g. zhu , phys .
d * 65 * , 094025 ( 2002 ) ; phys .
d * 66 * , 079904(e ) ( 2002 ) .
d. du , h. gong , y. sun , d. yang and g. zhu , phys .
d * 67 * , 014023 ( 2003 ) .
r. aleksan , p .- f .
giraud , v. mornas , o. pne and a. s. safir , phys .
d * 67 * , 094019 ( 2003 ) .
guo , o. m. a. leitner and a. w. thomas , hep - ph/0307201 m. beneke and m. neubert , hep - ph/0308039 . p. ball , j. high energy phys . ,
* 69 * , 005 ( 1998 ) .
l and m - z yang , hep - ph/0212373 .
a. ali and c. greub , phys .
d * 57 * , 2996 ( 1998 ) ; a. ali , g. kramer and c - d .
l , phys .
d * 58 * , 094009 ( 1998 ) .
m. ciuchini , e. franco , g. martinelli , m. pierini and l. silvestrini , phys .
b * 515 * , 33 ( 2001 ) ; hep - ph/0110022 ; hep - ph/0208048 .
a. j. buras and l. silvestrini , nuc . phys .
b * 569 * , 3 ( 2000 ) .
the babar collaboration ( b. aubert _ et al .
_ ) , hep - ex/0307026 .
heavy flavour averaging group , http://www.slac.stanford.edu/xorg/hfag/index.html .
f. james and m. roos , minuit , cern d506 , cern program library office , cern , ch-1211 geneva 23 , switzerland .
http://ckmfitter.in2p3.fr/ n. de groot , w. n. cottingham and i. b. whittingham , hep - ph/0305263 .
the babar collaboration ( b. aubert _ et al .
_ , hep - ex/0207068 . k. c. yang , hep - ph/0308095 | we present fits to charmless hadronic @xmath0 decay data from the babar , belle and cleo experiments using two theoretical models : ( i ) the qcd factorization model of beneke _ et al . _ and ( ii ) qcd factorization complemented with the so - called _ charming penguins _ of ciuchini _ et al .
_ when we include the data from pseudoscalar - vector decays the results favor the incorporation of the _ charming penguin _ terms .
we also present fit results for the unitarity triangle parameters and the cp - violating asymmetries . |
strong gravity affecting an accretion disc in vicinity of a black hole horizon governs three observationally significant phenomena enabling to determine the black hole mass @xmath0 and the dimensionless spin @xmath1 .
the phenomena are related to the timing effects , namely to the frequencies of the hf qpos and connected lf qpos , to the spectral continuum of the accretion disc , and the fe spectral lines profiled by the influence of the black hole spacetime .
the most interesting ( and precise ) information is connected to the twin hf qpos observed with the fixed frequency ratio @xmath3 in the microquasars grs 1915 + 105 , xte j1550 - 56 , and gro j1655 - 40 .
such twin hf qpos can be explained by the so called geodesic oscillation models using frequencies of the geodesic epicyclic motion in the field of kerr black holes , i.e. , the orbital frequency , @xmath8 , and the epicyclic radial , @xmath9 , and latitudinal , @xmath10 , frequencies .
however , the twin hf qpos in the three microquasars can not be explained by a fixed oscillation model , if we assume a kerr black hole @xcite .
a unique , epicyclic resonance model exists @xcite if we assume central kerr naked singularities that demonstrate special properties of the prograde circular motion @xcite . in the case of the grs 1915 + 105 microquasar ,
the @xmath3 hf qpos can be explained by the epicyclic resonance model @xcite , while for the xte j1550 - 56 and gro j1655 - 40 microquasars , the twin hf qpos and the related lf qpo can be explained by the relativistic precession model @xcite that is by definition combined with the relativistic nodal model of the lf qpos @xcite . in the case of the microquasar grs 1915 + 105 the limits on the black hole mass and spin implied by the models of the twin hf qpos are in agreement with the limits implied by the spectral measurements @xcite , and agreements of limits implied by models of the qpos and the models of spectral measurements has been demonstrated also in the case of the xte j1550 - 56 microquasar @xcite . on the other hand , in the case of the microquasar gro j1655 - 40 , the spin limits of the spectral ( continuum and fe - line ) measurements @xcite are contradicting each other , and , moreover , they are both contradicting the spin limits implied by the geodesic models of qpos @xcite , if the limits on the black hole mass determined by the dynamical studies based on optical measurements @xcite are considered . in the present paper
we consider the controversial restrictions on the gro j1655 - 40 black hole mass and spin .
keeping the relevance of the mass restrictions implied by the weak gravity dynamical models based on the optical measurements of the binary system that have high relevance and are quite independent of the timing and spectral measurements connected with the strong gravity near the black hole horizon , we first discuss the possibility to obtain an agreement of the twin hf qpo geodesic models and any of the spectral methods , and then we test possibility of agreement of the string loop oscillation model of twin hf qpos with the predictions of the spectral measurements .
the gro j1655 - 40 low - mass - x - ray - binary ( lmxb ) source is one of the extensively studied galactic microquasars , i.e. , the sources where the accreting central object is assumed to be a black hole . however , controversial estimates of the gro j1655 - 40 black hole gravitational mass @xmath0 and dimensionless spin @xmath1 have been reported recently @xcite .
we first summarize origin of this controversy related in the strong gravity regime to models of x - ray spectral measurements implying limits on the black hole spin , and the so called geodesic models of qpos implying in the strong gravity regime limits on the black hole mass and spin that are confronted with the limits on the black hole mass implied by dynamical studies in the weak gravity regime .
mass of the gro j1655 - 40 black hole is restricted by dynamical methods related to spectro - photometric optical measurements @xcite that are not related to the timing studies of qpos based on the x - ray measurements , and have high degree of credibility as they are related to the weak field gravity techniques .
the range of allowed values of the black hole mass implied by the optical measurements reads @xcite m_opt = ( 5.40.3 ) m_. [ mopt ] note that there exists an earlier and larger estimate of the gro j1655 - 40 black hole mass determined in @xcite that reads m_bh = ( 7.020.22 ) m_. [ mopt2 ] however , the later estimate presented in @xcite is in recent papers considered as the relevant one .
the x - ray spectroscopy techniques give in the case of the gro j1655 - 40 microquasar controversial results .
the spectral continuum measurements @xcite predict the black hole spin in the range 0.65 < a < 0.75 . [ a - cont ]
the measurements of the fe spectral lines profiled by the strong gravity of the black hole predict @xcite 0.94 < a < 0.98 .
[ a - line ] clearly , there is a strong discrepancy between the results of spectral measurements related to radiation processes in the strong gravity regime .
we could expect that the timing effects related to qpos occuring in the strong gravity regime could help to determine the correct spectral measurements due to the spin estimates .
the rossi xte observatory brings many timing measurements of the x - rays emitted by the gro j1655 - 40 source that are summarized in @xcite .
interesting are those related to the twin hf qpos and the lf qpos .
the lf qpos were observed between 0.1 hz and 30 hz , usually independently on the hf qpos .
however , most important is observation of twin hf qpos with frequencies stabilized at the lower frequency @xmath11 hz and the upper frequency @xmath12 hz that were observed simultaneously ; then it is natural to assume that the twin hf qpos are related to a common radius where both the observed oscillatory modes occur .
the twin hf qpos were reported for the first time in @xcite where simultaneously with the twin hf qpos a lf qpo was reported at @xmath13 hz . the magnitude of the observed frequencies and the character of the oscillations indicate strongly that the twin hf qpos should occur in close vicinity of the black hole horizon , being related to the orbital motion . the special simultaneous observation of the twin hf qpos and the lf qpo enables to obtain stringent restrictions on the mass and dimensionless spin of the central black hole , if we assume that all of these qpos arise at a given radius of the accretion disc @xcite .
the monte carlo technique applied to the observational data and the relativistic precession ( rp ) model , belonging to the geodesic models of twin hf qpos , with the frequency identification @xmath14 and @xmath15 , along with the related relativistic nodal precession model of the lf qpo with the frequency identification @xmath16 , implied for the black hole mass and spin the limits @xcite m = ( 5.310.07 ) m_,a = 0.2900.003 .
[ m - a - mot ] the mass limit is in very good agreement with the mass limit given by the optical measurements .
however , there is a clear discrepancy with the spin limits given by both the spectral measurements , as the rp model predicts spin @xmath17 .
recently , the standard nodal precession model of the lf qpos and a variety of the geodesic models of twin hf qpos , i.e. , models using frequencies of the oscillatory modes combined from the frequencies of the geodesic epicyclic motion , has been tested for matching to observational data of the simultaneously observed twin hf qpos and lf qpo in the microquasar gro j1655 - 40 ; models originally proposed to explain only the twin hf qpos were generalized to interpret also the lf qpo by the relativistic nodal precession @xcite . instead of the monte carlo technique ,
the method of frequency relations introduced in @xcite has been used @xcite , and the fitting has been done to the data of the twin hf qpos and simultaneous lf qpo presented as sample b1 in tab.2 . of @xcite .
the peak frequencies with measurement error of the peak frequencies ( centroid frequencies dominated by statistics of the frequencies ) were used @xcite & & _ l = 2984 , _ low = 17.30.1 , + & & _ u = 4412 .
[ fff ] for all the considered models of twin hf qpos it was assumed that both the twin hf qpos and the lf qpo arise at a common radius .
then the frequency relation technique enables to obtain a mass - spin relation @xmath18 for the twin hf qpos with frequency ratio parameter p = ( ) ^2 , and due to the assumption of the common radius a mass - spin relation @xmath19 can be determined for the lf qpo .
the two relations imply the limits on the black hole mass and spin related to the qpo measurements @xcite .
it has been found in @xcite that three geodesic models can predict the mass in agreement with the optical measurement limit . along with the rp models when the frequency relation method implies the limits
m = ( 5.30.1 ) m_,a = 0.2860.004 , [ m - a - rp ] in agreement with the estimates given by the monte carlo technique results in @xcite , only two other twin hf qpos models can fit the optical mass limit . first is the so called total precession model , with the frequency identification given by @xmath15 and @xmath20 , implying the limits m = ( 5.50.1 ) m_,a = 0.2760.003 .
[ m - a - rp1 ] second is the resonance epicyclic model demonstrating a beat frequency , with the frequency identification @xmath21 and @xmath22 , implying the limits m = ( 5.10.1 ) m_,a = 0.2740.003 .
[ m - a - re1 ] clearly , all the models predict the black hole spin @xmath17 . we conclude that
, unfortunately , none of the considered geodesic oscillatory models of twin hf qpos combined with the nodal precession model of lf qpos can predict values of the black hole spin that could be matched to the predictions of the x - ray spectral measurements , if we assume that the simultaneously measured qpos occur at the same radius being thus physically related .
( for completeness , we have tested also the fitting to the mass estimate @xmath23 then the models denoted in @xcite as epicyclic resonance model er3 and the tidal distortion model td meet this mass limit for the spin ranges @xmath24 and @xmath25 that are again in contradiction with the results of both the spin spectral estimates . )
the controversy of the gro j1655 - 40 black hole spin estimates could be solved by using models of twin hf qpos that are based on frequencies reflecting non - geodesic effects , including thus not only the kerr spacetime parameters , but also some additional parameters related to the non - geodesic phenomena .
we test in the present paper the string loop oscillation model @xcite potentially reflecting the influence of tension related to internal toroidal magnetic fields in accretion discs @xcite .
of course , it is worth to consider also possibilities keeping the assumption of the purely geodesic ( gravitational ) origin of the oscillatory modes giving rise to the twin hf qpos and the lf qpo .
we discuss this possibility in the following section .
in order to keep assumption of validity of purely geodesic models of twin hf qpos for explanation of observed qpos in the gro j1655 - 40 microquasar and enable possibility to match the black hole spin values predicted by the spectral measurements methods and the mass limit given by the optical measurements , we have to abandon the assumption that the dimensionless radius @xmath26 where the twin hf qpos occur coincides with the dimensionless radius @xmath27 where the simultaneously observed lf qpo occurs .
the hf qpos and the lf qpo are thus assumed to be physically independent . we still keep assumption that the lf qpo is determined by the relativistic nodal oscillations , but these lf oscillations are not correlated with the twin hf qpos .
then we can treat the three simultaneously observed qpos by the frequency relation method proposed in @xcite , but the version introduced in @xcite has to be slightly modified due to the fact that we do not assume @xmath28 .
of course , we can determine the ratio of these two radii in dependence on the twin hf qpos model , in the range of allowed values of the black hole spin .
we first introduce the geodesic models of qpo and then we use them to fit the qpo data observed in gro j1655 - 40 source . in the kerr spacetimes ,
circular geodesics exist only in the equatorial plane @xcite .
the radial epicyclic frequency @xmath29 , and the vertical epicyclic frequency @xmath10 of the near - circular epicyclic motion are given by the relations @xcite [ frequencies ] _
r^2 = _ _ ^2 , _
^2 = _ _ ^2 , where the orbital ( azimuthal ) frequency @xmath30 , sometimes called keplerian frequency , and the related dimensionless epicyclic frequencies are given by the formulae _ & = & ( ) , + _ r&=&1-+ - , + _ & = & 1-+ .
the dimensionless radius @xmath31 is introduced , where the gravitational radius of the black hole , @xmath32 .
the radial profiles of the epicyclic frequencies are illustrated in fig.1 .
since all the frequencies @xmath33 have the same mass scaling , it is clear that in the geodesic models of twin hf qpos containing only linear combinations of these frequencies , the frequency ratio of the lower and upper frequencies will be independent of the mass parameter @xmath0 , being dependent only on the spin parameter @xmath1 @xcite .
this fact enables an effective application of the frequency relation method in the case of the geodesic models of twin hf qpos .
+ the geodesic models of twin hf qpos can be separated into three classes : the hot spot models ( the rp model and its variants @xcite , the tidal precession model @xcite ) , resonance models @xcite and disc oscillation ( discoseismic ) models @xcite .
these models were applied to match the twin hf qpos and the lf qpo in the microquasar gro j1655 - 40 @xcite .
we have tested in the present study all the models considered in @xcite .
here we briefly summarize properties of the resonance models that were shown to be the only sucessful models in matching at least one of the spin limits predicted by the spectral measurements , along with the mass limits implied by the dynamical restrictions due to optical measurements .
the lf qpo remains related to the relativistic nodal ( lense - thirring ) precession with frequency @xmath34 .
the observations demonstrate that lf qpos given by the lense - thirring effect can occur independently on the twin hf qpos , and only exceptionally their simultaneous occurrence is observed @xcite .
therefore , we are approved to use the assumption of the twin hf qpos and the lf qpo simultaneously observed , but physically uncorrelated .
the epicyclic resonance ( er ) models @xcite consider resonance of axisymmetric oscillation modes of accretion discs .
the accretion discs can be geometrically thin , having the geodetical ( keplerian ) profile of angular velocity @xcite , or toroidal , geometrically thick , with angular velocity profile determined by pressure gradients @xcite .
frequencies of the disc oscillations are related to the orbital and epicyclic frequencies of the circular geodesics for both the keplerian discs @xcite and slender tori @xcite .
the resonance can be of two kinds .
the internal , parametric resonance of the radial and vertical epicyclic oscillatory modes , representing the basical resonance epicyclic model , governed by the mathieu equation , which predicts the strongest resonant phenomena for the frequency ratio @xmath3 @xcite .
the forced non - linear resonance admits presence of combinational ( beat ) frequencies in the resonant solutions @xcite . of course , the resonant phenomena could be relevant also in the framework of the hot spot models @xcite .
while in the er models the oscillatory modes of the accretion disc are assumed axisymmetric , in the warped disc ( wd ) oscillation model using the inertial - acoustic modes and the so called g - modes of thin discs oscillations the oscillations are assumed non - axisymmetric @xcite .
the parametric resonance admits slight scatter of the resonant frequencies , i.e. , this kind of resonance can occur while the oscillating modes in resonance have frequency ratio slightly different from the exact rational ratio ; width of the resonance scatter decreases with increasing order of the resonance @xcite . for forced resonances , scatter of frequency ratio from the rational ratio
is governed by non - linear effects @xcite .
therefore , we consider as relevant all the frequency ratios given by the measured hf qpo frequencies with their errors . + [ cols="<,^,^,^,^,^,^",options="header " , ] + dynamics of the axisymmetric string loops in the axisymmetric kerr geometry is governed by the parameters of the gravitational field , the energy parameter @xmath35 of the string loop , and the two parameters , @xmath36 , governing the combined effects of the string tension and its angular momentum @xcite .
small harmonic or quasi - harmonic oscillations of the string loops can occur around stable equilibrium positions in the equatorial plane of the kerr geometry .
for the radial and latitudinal ( vertical ) harmonic oscillatory string loop motion in the kerr spacetimes , the frequencies related to distant observers are given by @xcite _
r = _ r , _
= _ ; the dimensionless angular frequencies read ^2_(r ) & = & , + ^2_(r ) & = & , where j_e(ex ) ( r ) & & h ( ^2 + 1 ) ( r-1 ) ( 6 a^2 r -3 a^2 r^2 -6 a^2 -5 r^4 + 12 r^3 ) + & & + ( ^2 + 1 ) + 8 a ( r-1 ) ( a^2 + 3r^2)^2 + & & + 4 a ^-1/2 h [ jeex ] and h(r;a ) & = & a^2 ( r+2 ) + r^3 , + f(r;a ) & = & ( r-3 ) r^4 -2 a^4+a^2 r ( r^2 - 3 r+6 ) .
[ ffce ] the zero points of the function @xmath37 determine the marginally stable equilibrium positions of the string loops ; at the zero points the frequency of the radial oscillatory mode of the string loops vanishes - for details see @xcite .
the radial profiles of the frequencies @xmath10 and @xmath9 of the string loop harmonic oscillations are demonstrated in figure [ figloop ] .
we assume resonance phenomena in the string loop oscillatory motion , governed by the kolmogorov - arnold - moser theory @xcite , as the 3:2 frequency ratio is observed at the twin hf qpos observed in the gro j1655 - 40 microquasar .
we directly identify the observed frequencies @xmath38 with the @xmath39 or @xmath40 frequencies . in the case of the string loop oscillation model
the additional nodal frequency model can be attributed to a physically independent relativistic nodal precession of a hot spot related to the lf qpo .. the fitting of the string loop oscillation frequencies to the observed frequencies introduced in @xcite will be used here .
the string loop oscillation model implies a triangular limit on the spacetime parameters @xmath41 see figure [ figloop ] .
introduces no restriction on the black hole spin , only the stringy parameter is restricted in this case . ]
the limiting values of the black hole mass are presented in table 2 . the mass limit given by the optical observations @xcite introduces an additional restriction on the gro j1655 - 40 black hole spin and on the stringy parameter @xmath42 restricting them to negative values only , as given in table 2 .
we can see that due to the string loop oscillation model of the twin hf qpos , small values of the black hole spin are forbidden , while the bottom limit of @xmath43 implies possibility of fast - rotating black hole in the microquasar gro j1655 - 40 , in agreement with the x - ray spectra measurements .
we used a variety of the geodesic models of twin hf qpos to match the twin hf qpos simultaneously observed with a low frequency qpo in the microquasar gro 1655 - 40 , testing their ability to predict the gro j1655 - 40 black hole mass in agreement with the limits of the optical measurements , and its spin in agreement with the spectral measurements , solving thus the recently discussed controversy @xcite . we have shown that the controversy in the mass and spin estimates of the gro j1655 - 40 black hole can be overcome by the er model of twin hf qpos that can predict the black hole spin in agreement with the fe - line spectral measurements , excluding thus the relevance of the spectral continuum measurements .
however , the assumption of occurrence of the twin hf qpos and the simultaneously observed lf qpo at a common radius has to be abandoned ; the position of the source of the lf qpo relative to the position of the source of the twin hf qpos is then determined .
the er model introduces additional restrictions on the gro j1655 - 40 black hole mass ( @xmath44 ) and dimensionless spin ( @xmath45 ) .
the radius @xmath46 is relative distant from the black hole horizon .
the radius ratio @xmath47 is then relatively low .
acceptable is also the special variant er5 of the epicyclic resonant model allowing for observability of a beat frequency . in this case
only the precise values of the kerr spacetime parameters are allowed , namely , @xmath48 and @xmath49 .
the radius @xmath50 is closer to the black hole horizon in comparison to the case of the er model .
the radius ratio @xmath51 is then larger than in the case of the er model . for comparison
, we have tested also recently introduced string loop oscillation model using phenomena of non - geodesic origin that can reflect tension of toroidal magnetic fields in accretion discs @xcite .
we have demonstrated that restrictions of the string loop oscillation model on the gro j1655 - 40 black hole mass can be matched to the restrictions on the mass parameter implied by the optical measurements , if the black hole spin @xmath7 .
therefore , predictions of the string loop oscillation model can be in agreement with both the restrictions implied by the spectral measurements .
the matching of the twin hf qpos by the string loop model puts restrictions on the string - loop parameters , namely the parameter @xmath42 is restricted to negative values .
the lf qpo can be matched by the relativistic nodal precession model for any black hole parameters limited by the string loop model .
we can conclude that in order to select between the successful models of the timing effect related to the qpos observed in the gro j1655 - 40 microquasar , giving coherent limits on the black hole mass and spin , additional data from measurements of the spectral continuum and profiled spectral lines in the microquasar , and more precise data of the timing measurements of qpo , are necessary .
such precise measurements could finally exclude validity of the geodesic models for the microquasar gro j1655 - 40 .
we stress that confirmation or falsification of the presented qpo models can be expected due to observations of the qpos at the microquasar gro j1655 - 40 by the planned space x - ray observatory loft promising detection of the timing effects related to the qpos with precision by one order higher than those obtained recently by the rossi x - ray detector ; even the temporal evolution of qpos during the measurements is expected in @xcite .
for example , in the case of the lf qpos with frequencies around 10hz even lag effects could be estimated , while for the twin hf qpos observed in hundreds of hz the expected precision of the frequency measurements could enable to determine the character and details of the assumed resonant phenomena .
z.s . acknowledges the albert einstein centre for gravitation and astrophysics supported by the czech science foundation grant no .
14 - 37086 g . m.k . acknowledges the czech science foundation grant no .
16 - 03564y . | estimates of the black hole mass @xmath0 and dimensionless spin @xmath1 in the microquasar gro j1655 - 40 implied by strong gravity effects related to the timing and spectral measurements are controversial , if the mass restriction determined by the dynamics related to independent optical measurements , @xmath2 , are applied .
the timing measurements of twin high - frequency ( hf ) quasiperiodic oscillations ( qpos ) with frequency ratio @xmath3 and the simultaneously observed low - frequency ( lf ) qpo imply the spin in the range @xmath4 if models based on the frequencies of the geodesic epicyclic motion are used to fit the timing measurements , and correlated creation of the twin hf qpos and the lf qpo at a common radius is assumed . on the other hand , the spectral continuum method implies @xmath5 , and the fe - line - profile method implies @xmath6 .
this controversy can be cured , if we abandon the assumption of the occurrence of the twin hf qpos and the simultaneously observed lf qpo at a common radius .
we demonstrate that the epicyclic resonance model of the twin hf qpos is able to predict the spin in agreement with the fe - profile method , but no model based on the geodesic epicyclic frequencies can be in agreement with the spectral continuum method .
we also show that the non - geodesic string loop oscillation model of twin hf qpos predicts spin @xmath7 under the optical measurement limit on the black hole mass , in agreement with both the spectral continuum and fe - profile methods . |
it is likely that structures in the universe grow from small initially gaussian density perturbations that progressively detach from the general expansion , reach a maximum radius and then collapse to form bound objects .
larger haloes are formed hierarchically by mergers between smaller ones . + two different kinds of methods are widely used for the study of the structure formation .
the first one is n - body simulations that are able to follow the evolution of a large number of particles under the influence of the mutual gravity from initial conditions to the present epoch .
the second one is semi - analytical methods . among these ,
press - schechter ( ps ) approach and its extensions ( eps ) are of great interest , since they allow us to compute mass functions ( press & schechter 1974 ; bond et al 1991 ) , to approximate merging histories ( lacey & cole 1993 , lc93 hereafter , bower 1991 , sheth & lemson 1999b ) and to estimate the spatial clustering of dark matter haloes ( mo & white 1996 ; catelan et al 1998 , sheth & lemson 1999a ) .
+ in this paper , we present merger - trees based on monte carlo realizations .
this approach can give significant information regarding the process of the formation of haloes .
we focus on the mass growth histories of haloes and we compare these with the predictions of n - body simulations . + this paper is organized as follows : in sect.2 basic equations are summarized . in sect.3 the algorithm for the construction of merger - trees as well as tests regarding the reliability of this algorithm are presented .
mass - growth histories are presented in sect.4 , while the results are summarized and discussed in sect.5 .
in an expanding universe , a region collapses at time @xmath0 , if its overdensity at that time exceeds some threshold . the linear extrapolation of this threshold up to the present time is called a barrier , b. a likely form of this barrier is as follows : @xmath1\ ] ] in eq.(1 ) @xmath2 , @xmath3 and @xmath4 are constants , @xmath5 , where @xmath6 is the linear extrapolation up to the present day of the initial overdensity of a spherically symmetric region , that collapsed at time @xmath0 .
additionally , @xmath7 , where @xmath8 is the present day mass dispersion on comoving scale containing mass @xmath9 .
@xmath10 depends on the assumed power spectrum .
the spherical collapse model ( sc ) has a barrier that does not depend on the mass ( eg .
lc93 ) . for this model
the values of the parameters are @xmath11 and @xmath12 .
the ellipsoidal collapse model ( ec ) ( sheth & tormen 1999 , st99 hereafter ) has a barrier that depends on the mass ( moving barrier ) .
the values of the parameters are @xmath13 , @xmath14 , @xmath15 and are adopted either from the dynamics of ellipsoidal collapse or from fits to the results of n - body simulations .
additionally , the non - radial ( nr ) model of del popolo & gambera ( 1998 ) -that takes into account the tidal interaction with neighbors- corresponds to @xmath13 , @xmath16 and @xmath17 .
+ sheth & tormen ( 2002 ) connected the form of the barrier with the form of the multiplicity function .
they show that given a mass element -that is a part of a halo of mass @xmath18 at time @xmath19- the probability that at earlier time @xmath0 this mass element was a part of a smaller halo with mass @xmath9 is given by the equation : @xmath20\mathrm{d}s\ ] ] where @xmath21 and @xmath22 with @xmath23 .
+ the function @xmath24 is given by : @xmath25^n}{n ! } \frac{\mathrm{\partial ^n}}{\partial s^n}b(s , t).\ ] ] setting @xmath26 , and @xmath27 in eq . 3
, we can predict the unconditional mass probability @xmath28 , that is the probability that a mass element is a part of a halo of mass @xmath9 , at time @xmath0 . the quantity
@xmath29 is a function of the variable @xmath30 alone , where @xmath31 .
since @xmath32 and @xmath33 evolve with time in the same way , the quantity @xmath29 is independent of time . setting @xmath34 ,
one obtains the so - called multiplicity function @xmath35 .
the multiplicity function is the distribution of first crossings of a barrier @xmath36 by independent uncorrelated brownian random walks ( bond et al .
that is why the shape of the barrier influences the form of the multiplicity function .
the multiplicity function is related to the comoving number density of haloes of mass @xmath9 at time @xmath0 -@xmath37- by the relation , @xmath38 that results from the excursion set approach ( bond et al .
1991 ) . in eq.4
, @xmath39 is the density of the model of the universe at time @xmath0 .
+ using a barrier of the form of eq.1 in the unconditional mass probability , the following expression for @xmath35 is found : @xmath40\exp\left(-0.5a\nu^2[1+\beta(\alpha \nu^2)^{-\gamma}]^2\right)\ ] ] where @xmath41 in fig.1 we plot @xmath42 versus @xmath30 .
the solid line results from eq . 5 for the values of the parameters that correspond to the sc model .
the dotted line is derived from the values of ec model while the dashed line corresponds to the nr model . for matter of completeness
, we draw two more lines . the first line , the one with thick dashes- corresponds to the function of jenkins et al .
( 2001 , j01 hereafter ) .
this satisfies the following equation : @xmath43\mid^{3.8})\ ] ] in order to express the above relation as a function of @xmath30 , we substitute @xmath44 and we assume a constant value of @xmath32 , that of the einstein - de sitter universe , namely @xmath45 .
the above formula is valid for @xmath46 .
the second additional line , dot - dashed one , is the function proposed by st99 and is : @xmath47 where @xmath48 and the values of the constants are : @xmath49 , @xmath50 and @xmath51 .
+ we have noted that the comparison of the above curves with the results of n - body simulations is in a very good agreement , except for the sc model . such a comparison is presented by yahagi et al .
according to their results , the numerical multiplicity functions reside between the st99 and j01 multiplicity functions at @xmath52 and are below the st99 function at @xmath53 .
additionally , the numerical multiplicity functions have an apparent peak at @xmath54 -as those given by eq .
5- instead of the plateau that is seen in the j01 function .
let there be a number of haloes with the same present day mass @xmath18 .
the purpose of merger - trees realizations is to study the past of these haloes .
this is done by finding the distribution of their progenitors ( haloes that merged and formed the present day haloes ) at previous times . a way to do
this is by using eq.2 .
first of all , one has to use a model for the universe and a power spectrum .
in what follows we have assumed a flat model for the universe with present day density parameters @xmath55 and @xmath56 where @xmath57 is the cosmological constant and @xmath58 is the present day value of hubble s constant .
we used the value @xmath59 with @xmath60 and a system of units with @xmath61 , @xmath62 and a gravitational constant @xmath63 . at this system of units
@xmath64 + regarding the power spectrum- that defines the relation between @xmath10 and @xmath9 in eq.2- we employed the @xmath65 formula proposed by smith et al .
the power spectrum is smoothed using the top - hat window function and is normalized for @xmath66 . + for each one of the sc , ec and nr models , we studied four different cases .
these four cases differ to the present day mass of the haloes under study and are denoted as sc1 , sc2 , sc3 , sc4 for the sc model , ec1 , ec2 , ec3 and ec4 for the ec model and nr1 , nr2 , nr3 and nr4 for the nr model .
the first case of each model ( sc1 , ec1 and nr1 ) corresponds to haloes with mass 0.1 -measured in our system of units- while the second case ( sc2 , ec2 and nr2 ) corresponds to mass 1 .
sc3 , ec3 , nr3 correspond to haloes with mass 10 and sc4 , ec4 , nr4 correspond to masses 100 .
+ pioneered works for the construction of merger - trees are those of lc93 , somerville & kollat ( 1999 ) , st99 and van de bosch ( 2002 , vdb02 ) .
many of their ideas are used for the construction of our algorithm , that is described below .
+ first , let us define as @xmath67 , the number of realizations used .
this number is the total number of present day haloes of given mass @xmath18 .
a mass cutoff @xmath68 is used , that is a lower limit for the mass of progenitors .
no progenitors with mass less than @xmath68 exist in the merger history of a halo .
@xmath68 is set to a fixed fraction of @xmath18 . in what follows , @xmath69
( see also vdb02 ) . also , @xmath70 is the minimum value of the scale factor .
we set @xmath71 .
merging histories do not extend to values of @xmath72 .
a discussion about the choice of @xmath70 is given in section 4 . + useful matrices which have been used are the following : @xmath73 ( the masses of the parent haloes ) , @xmath74 ( the masses that are left to be resolved ) and @xmath75 ( the masses of the progenitors ) .
the argument @xmath76 takes the values from @xmath77 to @xmath78 , where @xmath78 is the number of haloes that are going to be resolved .
initially , @xmath79 .
the values of @xmath80 are used as input .
we discuss the choice of the value of @xmath81 in section 4 .
the scale factor @xmath82 is set to its initial value , ( that is the present day one ) , @xmath83 .
then , the matrices @xmath73 and @xmath74 take their values : @xmath84 , @xmath85 for @xmath86 .
it is convenient to set a counter for the number of time steps , so we set @xmath87 .
the rest of the procedure consists of the following steps : + * 1st step * : @xmath88 .
+ the equation @xmath89 is solved for @xmath90 , that is the new , ( current ) value of the scale factor . +
* 2nd step * : for @xmath91 + * 3rd step * a value of @xmath92 is chosen from the desired distribution . + * 4th step * : the mass @xmath93 of the progenitor is found , solving for @xmath93 the equation : @xmath94 .
if @xmath95 then , we return to the 3rd step . else , the halo with mass @xmath93 is a progenitor of the parent halo @xmath76 .
the mass to be resolved is now given by : @xmath96 .
if @xmath97 then @xmath93 is too small and is not considered as a real progenitor , therefore we return to the 3rd step .
else , + * 5th step * : the number of progenitors is increased by one , ( @xmath98 ) , and the new progenitor is stored to the list @xmath99 . +
* 6th step * : if the mass left to be resolved exceeds the minimum mass , ( @xmath100 ) , we go back to the 3rd step .
else , we return to the 2nd step , where the next value of @xmath76 is treated . + * 7th step * : we set the new number of haloes @xmath78 to be the number of progenitors , @xmath101 .
we store the progenitors , @xmath102 .
the masses of new parents are defined as the masses of progenitors by @xmath103 .
the value of the scale factor is updated @xmath104 , as well as the mass left to be resolved , @xmath105 . if @xmath106 we stop .
else , we return to the first step
. + output is the list of progenitors after a desired number of time steps . regarding the third step
, we have the following remarks . for the sc case
, the change of variables @xmath107 leads to a gaussian distribution with zero mean value and unit variance for the new variable @xmath108 .
so , we pick values @xmath108 from the above gaussian ( that is the well - known and very fast procedure press et al .
the values of @xmath109 are distributed according to the distribution described in eq .
2 . the distributions in the ec and nr cases
can not be expressed using a gaussian , so a more general method is used .
this method , described in press et al .
( 1990 ) is as follows : + let s suppose that we would like to pick numbers from a distribution function @xmath110 , where the variable @xmath108 takes positive values in the interval @xmath111 $ ] .
first , we calculate the integral @xmath112 .
we note that @xmath113 is not necessarily equal to unity .
then , a number , let say @xmath114 , between zero and @xmath113 is picked from a uniform distribution . the integral equation : @xmath115 is solved for @xmath108 .
the resulting points @xmath116 are then uniformly distributed in the area between the graph of @xmath117 , the x - axis and the lines @xmath118 and @xmath119 .
the values of @xmath108 have the desired distribution .
we have to note that this procedure is -as it is expected- more time demanding than the sc .
this is mainly due to the numerical solution of the integral equation .
however , it has the advantage of being general .
+ in eq.2 , @xmath120 becomes an one variable function for given values of @xmath121 and @xmath0 .
then , the above described general procedure is used in order to pick values of @xmath10 . for numerical reasons , it is convenient to use a new variable @xmath122 .
+ the distribution of the masses of progenitors is found using the following procedure : let @xmath123 be the number of progenitors of @xmath67 parent haloes at time @xmath0 and let @xmath124 be the minimum and the maximum mass of these progenitors , respectively .
we divide the interval @xmath125 $ ] to a set of @xmath126 intervals of length @xmath127 .
then , the number of progenitors , @xmath128 , in the interval @xmath129 $ ] for @xmath130 is found .
the distribution of progenitors masses is then given by @xmath131 .
this distribution has to be compared with the distribution that results from the analytical relations .
thus , multiplying eq.2 by @xmath132 , one finds the expected number of progenitors at @xmath0 that lie in the range of masses @xmath133 .
this number obeys the equation : @xmath134 which is known as the number - weighted probability for @xmath9 .
( redshift @xmath135 ) .
all curves correspond to a halo with present day mass @xmath136 . the smooth dotted and the crooked dotted lines give @xmath137 and @xmath138 versus @xmath139 for the sc1 model , respectively .
smooth solid and crooked solid lines give @xmath137 and @xmath138 for the nr1 model , respectively . in a similar way , the line with short thick dashes correspond to ec1 model.,width=529 ] in figs 2 and 3 , the distributions of progenitors are plotted for all models studied , for two different values of the scale factor .
the results of fig.2 correspond to scale factor @xmath140 and those of fig.3 to @xmath141 .
details are given in the captions of the figs .
large differences between the predictions of spherical and non - spherical models are shown .
these differences are very clear and they can be seen even if a relatively small sample of haloes is used ( we used @xmath142 ) .
on the other hand , the predictions for the two non - spherical models nr and ec are very close . + ( redshift @xmath143 ) .
all curves correspond to a halo with present day mass @xmath144 . the smooth dotted and the crooked dotted lines give @xmath137 and @xmath138 versus @xmath139 for the sc3 model , respectively .
smooth solid and crooked solid lines give @xmath137 and @xmath138 for the nr3 model , respectively .
lines with short thick dashes correspond to the results of ec3 model .
the smooth line gives @xmath137 while the crooked one gives @xmath138.,width=529 ] the time evolution of halo s mass is known as its mass growth history . however , a clear definition of the halo s mass at time @xmath0 is needed .
one can define as halo s mass at a step , the mass of its most massive progenitor at that step .
we call this procedure the _ most massive _ progenitor approximation .
a different approximation is the one followed by vdb02 .
we call this approximation the _ main _ progenitor approximation and it is as follows : at the first step , the most massive progenitor @xmath145 of a halo is found .
this is the halo s mass at that time . at the next step ,
the most massive progenitor of the halo with mass @xmath145 is found and is considered as the new halo s mass .
the procedure is repeated for the next steps .
according to the definition of lc93 , the scale factor at which the mass of the _ main _ progenitor equals half the present day mass of the halo is called the _ formation _ scale factor .
we denote it by @xmath146 .
a detailed description of the procedure follows : + * 1st step * : @xmath88 .
+ the equation @xmath89 is solved for @xmath90 that is the new ( current ) value of the scale factor . + * 2nd step*:for @xmath147 + the mass of the most massive progenitor of parent @xmath76 is set equal to zero ( @xmath148 ) . + * 3rd*. a value of @xmath92 from the desired distribution is chosen . + * 4th step * : a mass @xmath93 is found , solving for @xmath93 the equation : @xmath94 . if @xmath95 then , we go back to the 3rd step .
else , the halo with mass @xmath93 is a progenitor .
we set @xmath96 . if @xmath97 then @xmath93 is too small and is not considered as a real progenitor and so we return to the 3rd step .
else , + * 5th step * : the most massive progenitor at this time step is found .
+ @xmath149 . + * 6th step * : if the mass left to be resolved exceeds the minimum mass , ( @xmath100 ) , we go back to the third step .
else , we check if the mass of the most massive progenitor is for the first time smaller than half the initial mass of the halo . if this is correct , then the formation scale factor of this halo is defined by linear interpolation : @xmath150 otherwise , we proceed with the next @xmath76 . + * 7th step * : the list of the most massive progenitors is used , in order to find the mean mass at this time step .
the time of the scale factor is updated @xmath104 , as well as the mass left to be resolved , @xmath151 .
if @xmath152 we stop , else we go to the first step .
+ we have to note that the halo s mass at a step -derived by the above procedure- does not give the mass of its most massive progenitor at that step . for example , let a halo of mass @xmath18 that has at the first step two progenitors with masses @xmath153 and @xmath154 with @xmath155 .
then , the mass of the halo at the second step is defined as the mass of the most massive progenitor ( @xmath153 ) .
this progenitor is not necessarily the most massive of all progenitors at that step , since it could be less massive than one of the progenitors of @xmath154 .
the procedure of _ main _ progenitor described above has the advantage of being economic .
it does not require the construction of a complete set of progenitors in order to find the most massive one at a step .
+ we tested both the _ most massive _ and the _ main _ progenitor approximations and we found small differences only for massive haloes . in order to be able to compare our results with those of other authors
, we used the _
progenitor approximation . + a test for the reliability of merger - trees is the comparison of the distribution of formation scale factors with the one given by the analytical relations that follow .
the probability that the mass of the progenitor , at time @xmath0 , is larger than @xmath156 is given by : @xmath157 the formation time is defined ( lc93 ) as the epoch when the mass of the halo , as it grows by mergers with other haloes , crosses the half of its present day value .
then , @xmath158 is the probability the halo with present day mass @xmath18 had a progenitor heavier than @xmath156 at @xmath0 , which is equivalent to the probability that the halo is formed earlier than @xmath0 , @xmath159 .
thus , @xmath160 .
differentiating with respect to @xmath0 , we obtain : @xmath161\mathrm{d}m\ ] ] which gives the distribution of formation times .
+ the distribution of formation scale factors as it results from eq.13 is plotted in fig .
4 . the left hand side snapshot corresponds to haloes with present day mass 0.1 , while the right hand side one to haloes with present day mass 100 . in every snapshot
dotted , solid and dashed lines correspond to sc , nr and ec models , respectively . notice that the distributions that correspond to nr and ec models are shifted to the left relative to the sc model .
this means that structures formed earlier in the nr and ec models . as it can be seen in fig .
4 , the differences between spherical and non - spherical models are smaller for larger haloes .
as it is known in the literature(e.g yahagi et al , 2004 and references therein ) , the results of non - spherical models are in good agreement with the results of n - body simulations .
consequently , for massive haloes , the results of n - body simulations are close to the predictions of the spherical models , as well .
this disagrees with the results of vdb02 which predicts that the disagreement between spherical model and n - body simulations is larger for massive haloes .
+ fig.5 consists of four snapshots .
bars show the distributions of formation scale factor as they result from eq.11 .
wide bars are the predictions of sc model , while narrow filled bars are the results of ec or nr models .
solid lines are the predictions of eq . 13 for the sc model , while dashed lines are the predictions of the same eq .
for ec or nr models .
the two snapshots of the first row show the results of sc1 and ec1 ( left snapshot ) and sc1 and nr1 ( right snapshot ) while the two snapshots of the second row show the results of sc4 and ec4 ( left snapshots ) and sc4 and nr4 ( right snapshot ) .
it is shown that the predictions of eq .
13 are in very good agreement with the numerical distributions .
additionally , it is clear that the differences between the distributions of various models , resulting from eq .
13 , are successfully reproduced by numerical distributions .
it is also shown that differences between spherical and non - spherical models are more significant for haloes of small mass .
we define as mass - growth history , mgh , of a halo with present - day mass @xmath18 , the curve that shows the evolution of @xmath162 where @xmath163 is the mean mass at scale factor @xmath82 of haloes with present day mass @xmath18 .
two input parameters are used in the method of construction merger - trees in the procedure described in the previous section .
the first one is the time - step @xmath81 . for values of @xmath164 ,
it is shown ( vdb02 ) that mghs are not time - step dependent . according to that , we used a value @xmath165 .
the second parameter is @xmath70 .
our results are derived for @xmath71 . in order to justify this choice ,
let us discuss the following points regarding the definition of the mean mass .
let @xmath166 be the number of haloes at scale factor @xmath82 that have masses greater than @xmath68 . obviously , @xmath167 .
let also @xmath168 be the mass of the @xmath169 halo at scale factor @xmath82 .
we define by @xmath170 the sum @xmath171 , where for @xmath172 the sum is extended to all haloes that satisfy @xmath173 . for @xmath174 the sum
is extended to all haloes ( obviously in that case , the masses of a number of haloes are equal to zero since , by the construction of merger - trees , their mass histories are not followed beyond @xmath175 . finally , for @xmath176 the sum is over those haloes that satisfy @xmath177 .
then , we can define the mean mass at @xmath82 in three different ways : first @xmath178 , second @xmath179 and third @xmath180 . + as @xmath82 evolves from its present day value to smaller values , the number @xmath166 becomes smaller and the three mean values defined above become different .
it is clear that @xmath181 overestimates the mean value , while @xmath182 underestimates it . on the other hand
, @xmath183 requires a large number of @xmath67 so that the remaining sample of haloes @xmath184 to be large enough .
the choice of the value @xmath71 is justified by the fact that for @xmath185 the differences in the above three definitions are negligible .
+ the following results have been derived using a number of 10000 haloes .
we found that a number of haloes greater than 1000 is sufficient for producing smooth mghs that are in exact agreement with the ones of our results .
+ fig.6 consists of three snapshots .
the first snapshot shows the results of the sc model .
the other two snapshots correspond to the ec and to the nr model , respectively .
every snapshot contains four lines . from top
to bottom these lines correspond to the cases 1 , 2 , 3 and 4 , respectively .
it is clear that -at present time- the slopes that correspond to smaller masses are smaller .
this sequence of slopes is a sequence of formation times .
large haloes continue to increase their masses with a significant rate , even at the present epoch , while the mass growth of small haloes is negligible .
thus , small haloes have been formed earlier .
however , a _ formation _ scale factor can be defined by the condition that the growth rate of mass falls below some value .
bullock et al . ( 2001 , b01 ) used as a characteristic scale factor , the one satisfying the relation @xmath186 , where @xmath187 is a specific value , typically equal to 2 .
we used the value @xmath188 and we denote the solution of the equation @xmath189 .
additionally , we denote @xmath146 the solution of the eq . @xmath190 and we plot the results in fig.7 .
this figure consists of two snapshots . and @xmath9 .
from top to bottom , the solid lines are the plots of @xmath191 versus @xmath192 for the sc , nr and ec models , respectively . the dotted , short - dashed and long - dashed lines are the linear fits of the above lines , respectively . from top to bottom in the right snapshot ,
the solid lines are the plots of @xmath193 versus @xmath192 for the sc , nr and ec models , respectively .
the dotted line is the linear fit of the sc model , while the dashed line is a prediction of @xmath194 given by the toy - model of w02 ( see text for more details),width=529 ] left snapshot shows @xmath195 versus @xmath192 for the three models studied . from top to bottom ,
solid lines show the predictions of sc , nr and ec models , respectively .
these lines are well fitted by linear approximations that are shown by dotted , dashed and long - dashed lines , respectively .
thus @xmath146 and @xmath9 are related by a power law @xmath196 .
the values of @xmath197 are @xmath198 , @xmath199 and @xmath200 for sc , nr and ec models , respectively .
the solid lines in the right snapshot show @xmath201 versus @xmath192 .
the dotted line is the linear approximation of the results of sc model .
it can be seen that this approximation is very satisfactory .
linear approximations are not satisfactory for non - spherical models .
the dashed line in this snapshot shows the predictions of the toy - model of bullock et al .
( 2001 , b01 hereafter ) that calculates @xmath189 from the power spectrum using following procedure : let a halo with mass @xmath9 .
the scale factor at the epoch of its collapse @xmath189 , is found by the solution of equation .
: @xmath202=\sigma(fm)\ ] ] where , @xmath203 is the typical collapsing mass satisfying : @xmath204=1.686d(1)/d(a)$ ] , and @xmath205 is a constant .
@xmath206 is the growth factor predicted by linear theory ( peebles , 1980 ) .
the dashed line presented in the right snapshot of fig.7 is the prediction of the above procedure for @xmath207 and it is a good approximation of our results . + a less accurate but very practical approximation of @xmath189
is given ( see also suto 2003 ) by @xmath208 where , @xmath9 is expressed in our system of units .
a value of about @xmath209 for the exponent of the above relation is confirmed by our results too , since the best linear fits of the solid lines in the right snapshot of the fig.7 , are of the form @xmath210 with @xmath211 , @xmath212 and @xmath213 for the sc , nr and ec models , respectively .
+ in a complementary paper of that of b01 , wechsler et al .
( 2002 , w02 hereafter ) , based on the results of their n - body simulations , proposed for the growth of mass of dark matter haloes the exponential model : @xmath214\ ] ] in fig.8 , we compare our mghs with the predictions of the model of w02 .
this figure consists of twelve snapshots .
every row has four snapshots that correspond to the four cases of masses ( 0.1 , 1 , 10 and 100 ) studied for every model .
the predictions of sc model are presented in the first row .
the second row shows the predictions of nr and the third one the predictions of ec model .
the solid lines show our mghs while the dashed lines are the predictions of the model of w02 given by eq.16 .
the values of @xmath189 used are those shown by the solid lines of fig.7 that are predicted by the numerical solution of the equation @xmath215 thus , no other information ( such as the cosmology , power spectrum etc ) is used but only the mghs of haloes .
all snapshots refer to @xmath216 .
+ the conclusion from fig.8 is clear .
the predictions of eq.16 are very good approximations to the mghs predicted by merger - trees realizations .
versus @xmath217 for all the models studied .
the solid lines are the predictions of merger - trees realizations while the dashed ones are their fits by the model given in eq.16 ( see text for more details),width=529 ] the quality of fit depends on the mass of the halo and not on the model used . for example , the left column of fig .
8 shows a very good fit for all haloes with mass @xmath218 . a more careful look at this figure shows that the fit becomes less satisfactory for the earlier stages of the evolution of large haloes .
in fact , the predictions of merger - trees are steeper curves than those of the model of w02 .
this can be seen clearly at the 3thd and 4th column that show the predictions for haloes with masses @xmath219 and @xmath220 .
we note that the mass growth rate can be connected - under some assumptions- with the density profile of the halo ( e.g. nusser & sheth 1999 , manrique et al .
2002 , hiotelis 2003 ) .
according to the stable - clustering model used by the above authors , haloes grow inside - out .
this means that the accreted mass is deposited at an outer spherical shell without changing the inner density profile .
we recall that the radial extent of a halo is defined by its virial radius @xmath221 , that is the radius that contains a mass with mean density @xmath222 times the mean density of the universe @xmath223 .
@xmath222 is , in general , a function of the scale factor @xmath82 but in many applications it is set equal to a constant value of @xmath224 ( that is a value for @xmath222 for an einstein de sitter universe ) .
the mass @xmath225 contained inside @xmath221 at scale factor @xmath82 satisfies the two following equations : @xmath226 @xmath227 if the mass history of the halo is known , ( that is the evolution of its mass as a function of the scale factor ) , then the density profile can be calculated by differentiating eq .
17 with respect to @xmath82 , which gives : @xmath228 where @xmath229 is calculated by differentiating eq.18 .
finally eqs .
19 and 18 give , in a parametric form , the density profile .
it is clear that the density profile calculated from the above approximation depends crucially on the mass growth rate .
hiotelis ( 2003 ) used a model based on the above assumptions for haloes that grow by accretion of matter .
this means that the infalling matter is in a form of small haloes ( relative to the mass of the growing halo ) and the procedure can be approximated by a continuous infall . in that case
, it was shown that large growth rate of mass ( observed in heavy haloes ) leads to inner density profiles that are steeper than those of less massive haloes ( where the growth rate of mass is smaller ) .
+ we recall that the density profile of dark matter haloes is a very difficult problem contaminated by debates between the results of observations , the results of numerical simulations and the ones of analytical methods ( e.g. sand et .
al 2004 and references therein ) .
especially the inner density profile seems to follow a law of the form @xmath230 but the value of @xmath4 is not still known .
observations estimate a value of @xmath4 about @xmath231 ( sand et .
al 2004 ) , while the results of n - body simulations report different values as @xmath77 ( navarro et al .
1997 ) , or between @xmath232 and @xmath233 ( moore et al .
additional results from n - body simulations , as well as from analytical studies , connect the value of @xmath4 with the total mass of the halo ( ricotti 2002 , reed et al .
2005 , hiotelis 2003 ) . in any case , it is difficult to explain the differences appearing in the results of different approximations .
for example , regarding n - body simulations , some of the results are effects of force resolution . in a typical n - body code , such as treecode ( hernquist 1987 ) the force acting on a particle is given by the sum of two components : the short - range force that is due to the nearest neighbours and the long - range forces calculated by an expansion of the gravitational potential of the entire system .
as it can be shown , the value of the average stochastic force in the simulation is an order of magnitude greater than that obtained by the theory of stochastic forces .
consequently , small fluctuations induced by the small - scale substructure are not seen. this is the case of cold dark matter models in which the stochastic force generators are substructures at least three orders of magnitude smaller in size than the protoclusters in which they are embedded ( e.g. clusters of galaxies ) . on the other hand , regarding the differences between the results of various approximations , we notice that some of these differences could be due to the presence of the baryonic matter . especially in regions , as the central regions of haloes , the large density of baryonic matter may affect the distribution of dark matter .
+ since in this paper we describe in detail technical problems related to the construction of reliable merger - trees , we have to note that the inside - out model for the formation of haloes described above , requires calculations for very small values of @xmath82 , because the central regions are formed earlier .
therefore , in order to explore the center of the halo we have to go deep in the past of its history .
consequently , the problems relative to the calculation of mean mass that analyzed at the beginning of this section , become more difficult .
the density profiles predicted by the above described approximation are under study .
analytical models based on the extended ps formalism provide a very useful tool for studying the merging histories of dark matter haloes .
comparisons with the results of high resolution numerical simulations or observations show that the predictions of eps models are satisfactory ( e.g. lin et al .
2003 , yahagi et al .
2004 , cimatti et al .
2002 , fontana et .
al 2004 ) , although differences between various approximations always exist .
these differences show the need for further improvement of various methods .
+ in this paper we described in detail the numerical algorithm for the construction of merger - trees .
the construction is based on the mass - weighted probability relation that is connected to the model barrier .
the choice of the barrier is crucial for a good agreement between the results obtained by merger - trees and those of n - body simulations . as it is shown in yahagi et al .
( 2004 ) mass functions resulting from the sc model are far from the results of n - body simulations while those predicted by non - spherical models are in good agreement .
our results show that the distributions of formation times predicted by the non - spherical models are shifted to smaller values .
thus , the resulting formation times are closer to the results of n - body simulations than the formation times predicted by the sc model .
it should be noted that n - body simulations give , in general , smaller formation times than that of eps models ( e.g. lin et al .
it should also be noted that observations indicate a build - up of massive early - type galaxies in the early universe even faster than that expected from simulations ( cimatti , 2004 ) .
+ we have shown that a small sample of haloes ( @xmath234 ) is sufficient for the prediction of smooth mass growth histories .
these curves have a functional form that is in agreement with those proposed by the n - body results of other authors .
in particular the set of relations : @xmath235 and @xmath236\ ] ] describes very satisfactorily the results of non - spherical models ( nr and ec ) .
these relations give the mass @xmath9 of dark matter haloes at scale factor @xmath82 for the range of mass described in the text and for the particular cosmology used .
@xmath18 is the present day mass of the halo in units of @xmath237 .
+ finally , it should be noted that a large number of questions regarding the formation and evolution of galaxies remains open .
for example , a first step should be the improvement of the model barrier so that the results of merger - trees fit better both the mass function ( yahagi et al .
2004 ) and the collapse scale factor ( lin et al . 2003 ) .
it should be noted also that structures appear to form earlier in n - body simulations and even earlier in real universe .
additionally , predictions relative to the density profile are of interest .
some of these issues are currently under study .
we are grateful to the anonymous referee for helpful and constructive comments and suggestions .
n. hiotelis acknowledges the _ empirikion foundation _ for its financial support and dr m. vlachogiannis for assistance in manuscript preparation .
fontana , a. , pozzetti , l. , donnarumma , i. , renzini , a . ,
cimatti , a. , zamorani , g. , menci , n . ,
daddi , e. , giallongo , e. , mignoli , m . , perna , c. , salimbeni , s. , saracco , p . ,
broadhurst , t . ,
cristiani , s .
, dodorico , s . ,
gilmozzi , r . , 2004 , a&a , 424 , 23 suto , y. , 2003 , density profiles and clustering of dark halos and clusters of galaxies , in `` matter and energy in clusters of galaxies '' , asp conference series , vol.301 , 379 , 2003 , eds . : s. bowyer and c .- y.hwang | we have used merger trees realizations to study the formation of dark matter haloes .
the construction of merger - trees is based on three different pictures about the formation of structures in the universe .
these pictures include : the spherical collapse ( sc ) , the ellipsoidal collapse ( ec ) and the non - radial collapse ( nr ) .
the reliability of merger - trees has been examined comparing their predictions related to the distribution of the number of progenitors , as well as the distribution of formation times , with the predictions of analytical relations .
the comparison yields a very satisfactory agreement .
subsequently , the mass growth histories ( mgh ) of haloes have been studied and their formation scale factors have been derived .
this derivation has been based on two different definitions that are : ( a ) the scale factor when the halo reaches half its present day mass and ( b ) the scale factor when the mass growth rate falls below some specific value .
formation scale factors follow approximately power laws of mass .
it has also been shown that mghs are in good agreement with models proposed in the literature that are based on the results of n - body simulations .
the agreement is found to be excellent for small haloes but , at the early epochs of the formation of large haloes , mghs seem to be steeper than those predicted by the models based on n - body simulations .
this rapid growth of mass of heavy haloes is likely to be related to a steeper central density profile indicated by the results of some n - body simulations . and galaxies : halos formation structure , methods : numerical analytical , cosmology : dark matter 98.62.gq , 98.62.a , 95.35.+d |
cavity quantum electrodynamics ( cqed ) experiments provide insight into the fundamental concepts of quantum mechanics such as entanglement , decoherence and measurement @xcite .
additionally , optical cqed is currently growing into a key role in quantum information processing @xcite , where the transmission of quantum states between distant nodes is a central problem @xcite , and in entanglement - enhanced metrology @xcite , where the use of optical cavities has very recently lead to convincing demonstrations of metrologically useful spin squeezing capable of improving the stability of atomic clocks @xcite .
a fundamental figure of merit of the atom - cavity system is its cooperativity @xmath2 , which up to factors of order 1 is given by @xmath3 where @xmath4 it the cross - section and @xmath5 the finesse of the cavity mode , and @xmath6 is the ( effective ) scattering cross - section of the emitter(s ) placed in this mode . for example , the purcell effect arises when @xmath7 and the efficiency of an important class of single - photon sources scales as @xmath8 @xcite .
in addition to having a high ccoperativity , cavities used in quantum information applications should be miniaturized and fiber - coupled so that they can be used in scalable setups . this motivates considerable efforts to develop miniature high - finesse cavities @xcite .
however , no cavity type so far unites all desired properties in a single device .
therefore , progress in cqed and its applications hinges on the development of new , miniature cavities with high cooperativity . here
we describe a fiber - based fabry - perot cavity ( ffpc ) that combines tunability and high finesse with excellent and stable coupling to single - mode optical fibers , which is achieved without mode - matching optics .
the waist @xmath9 is about an order of magnitude smaller than in the macroscopic high - finesse cavities typically used in optical cqed experiments @xcite .
the cavity mode is located in free space , making it easy to couple to atomic emitters .
the cavity design is based on a new laser machining process where a single , focused co@xmath0 laser pulse creates a concave depression in the cleaved fiber surface . the first use of this ffpc was in an experiment that demonstrated strong - coupling cqed with bose - einstein condensates on atom chips @xcite , prompted by these results , the use of such cavities is now being explored with trapped neutral atoms and ions @xcite , but also with color centers in diamond @xcite , semiconductors @xcite and vibrating sub - wavelength objects @xcite . in all of these applications ,
the small size , ruggedness and built - in fiber coupling of the ffpc are advantageous , favouring its use in hostile environments such as cryostats .
moreover , we have already obtained single - atom cooperativities @xmath10 @xcite , and still higher values can be expected when using state - of - the art dielectric coatings .
this combines with the good fiber coupling , such that decisive progress can be expected from this ffpc for several applications : it should lead to single - photon sources @xcite with exceptional overall performance , where not only an atomic excitation is converted into a cavity photon with a probability close to one , but also this photon is efficiently extracted into a single - mode optical fiber . for similar reasons , the ffpc may improve the performance of quantum memories @xcite , where currently the overall conversion efficiency between the memory qubit and the desired optical mode is still low .
it will also enable single - atom state detection ( `` qubit readout '' ) with high quantum efficiency , and may enable optical detection with less than a single photon scattered by the emitter @xcite . in this article , we present a detailed theoretical and experimental investigation of this new type of fiber cavity .
the core of our cavity design is a concave , ultralow - roughness mirror surface fabricated directly on the end face of an optical fiber .
ffpcs with mirrors on fiber end faces have been built before @xcite , but until now had only moderate finesse ( up to a few 1000 ) , limited by the methods used to fabricate the concave surface on the fiber .
here we use a single co@xmath0 laser pulse to shape the concave mirror surface , which is then coated with a high - performance dielectric coating .
as we show below , this improves the finesse by more than an order of magnitude over the older methods and gives access to an interesting range of small radius of curvature ( roc ) . as in our earlier design using glued mirrors @xcite , a stable cavity is then constructed either from one mirror fiber tip facing a macroscopic , planar or concave mirror , or from two closely spaced fiber tips placed face to face ( fig .
[ fig : scheme ] ) . in this article
, we will be mainly interested in the second variant .
radii of curvature can be fabricated down to @xmath11 m and probably below .
because of the small mirror diameter @xmath12 ( smaller than the fiber diameter , which is typically @xmath13 m ) , the mirrors can approach each other very closely ( down to a few @xmath14 ) without touching .
the result is a very small mode waist @xmath9 between 1 and @xmath15 m , and a small mode volume down to a few @xmath16 .
the small @xmath9 combined with a length much shorter than the rayleigh range is the reason for the excellent sm fiber coupling efficiency with no need for mode - matching optics . as we discuss theoretically in sec .
[ sec : coupling ] and demonstrate experimentally in sec .
[ sec : couplingexp ] .
as will be described in more detail in @xcite , we have found a parameter regime where a single pulse of co@xmath0 laser radiation focussed onto the cleaved end face of an optical fiber produces a concave surface with extremely low roughness ( figure [ fig : surface ] ) . in this regime ,
thermal evaporation occurs , while melting is restricted to a thin surface region , avoiding global contraction into a convex shape .
this sets it apart from the regimes used in co@xmath0 laser - based fabrication of microspheres @xcite and transformation of microdiscs into high - q microtoroid resonators @xcite .
we can currently laser machine structures with roc between @xmath17 m and 2 mm , diameters between 10 and @xmath18 m , and a surface roughness of about 0.2 nm rms in the optical range .
these structures are obtained with co@xmath0 waist sizes between 18 and @xmath19 m , powers in the range @xmath20w and pulse lengths between 5 and 400ms .
a dichroic beam splitter and optical microscope are used to align the fiber with respect to the co@xmath0 laser focus . in our current setup
, we estimate the lateral alignment precision to be better than 2@xmath21 m .
we have laser machined a large number ( @xmath22 ) of fiber mirror structures .
initially , we used an afm to get both profile and roughness data . scanning the full surface to determine curvatures is a slow process however , so that we could only measure a small number of fibers with this method . to get approximate values for other fibers , we compared them to the well - characterized ones under an optical microscope .
more recently , we have used an interferometric microscope for measurements of the large - scale ( @xmath23 ) surface topography .
this method is fast enough to characterize each individual fiber in reasonable time .
figure [ fig : surface ] shows sem and afm images of representative fiber end faces after the application of a single laser pulse .
the section through the center in fig .
[ fig : profile](a ) shows a profile which is reasonably well approximated by a gaussian , @xmath24 over a wide parameter range .
( a limit is reached for long pulses and big waists , where a rim begins to form around the depression . )
because the profile is not spherical , the local roc varies with the transverse coordinate ( fig .
[ fig : profile](b ) ) .
close to the center of the profile , this variation is slow however , and the shape is well approximated by a circle ( fig .
[ fig : profile](c ) ) .
we use the central roc to define the mirror curvature @xmath25 ( fig .
[ fig : rddef ] ) .
the full width at @xmath26 of the gaussian profile gives an estimate of the useful mirror diameter @xmath12 ( cf .
[ sec : finesse ] ) .
, obtained from a tenth - order polynomial fit to the data in ( a ) to reduce the effect of measurement noise .
( c ) zoom of the data in ( a ) ( red solid line ) and a section of a circle with @xmath27 m ( the roc of the profile at @xmath28 ) for comparison ( black dot - dashed line).,scaledwidth=75.0% ] because of the approximately gaussian shape of the depression , @xmath25 , @xmath12 and the total structure depth @xmath29 are related by @xmath30 for example , a mirror structure with @xmath31 m and useful diameter @xmath32 m is only @xmath33 m deep . with the laser parameters given above ,
resulting structures are 0.01 to 4@xmath34 m deep and have diameters @xmath12 between 10 and @xmath18 m .
rocs measured at the bottom of the depression are between @xmath35 and @xmath36 m .
part of the processed surfaces show a slight ellipticity ( up to a few percent ) , which is probably due to an observed astigmatism of the co@xmath0 beam : when varying the fiber position along the co@xmath0 beam axis , the beam profile changes from circular to elliptical .
alignment error along this axis translates into ellipticity of the depression .
this could easily be improved in a future setup .
afm measurement areas were between @xmath37 and @xmath38 . to extract surface roughness
, we subtract a two - dimensional , fourth - order polynomial that accounts for the concave overall shape and then calculate the isotropic power spectral density ( psd ) . integrating the psd up to @xmath39 consistently yields @xmath40 nm .
( reference measurements on mica sheets ( neglegible roughness ) yield @xmath41 nm . ) a widely used estimate linking this roughness to the scatter loss is @xcite @xmath42 we thus obtain an estimate of @xmath43ppm for near - infrared light at @xmath44 nm , assuming a high - quality mirror coating that does not significantly increase this roughness .
such a coating also has very low absorption loss , @xmath45ppm being a realistic value @xcite . with a transmission equaling the losses , @xmath46
, we thus expect a maximum finesse @xmath47 for a cavity made from two fiber mirrors with identical coatings , and 170000 if the transmission of one mirror is reduced to 2ppm .
we consider cavities consisting of two mirrors , labeled @xmath48 , of intensity transmission @xmath49 , scatter and absorption losses @xmath50 , and reflectivity @xmath51 .
rocs are @xmath52 and ( effective ) diameters @xmath53 .
the optical distance between the mirrors is @xmath54 ( which is slightly larger than the geometric distance @xmath55 , cf .
[ sec : waist ] ) .
basic quantities characterizing the cavity are its free spectral range @xmath56 and the width of the tem@xmath57 cavity resonances , usually expressed as fwhm frequency @xmath58 in laser physics and as hwhm angular frequency @xmath59 in cqed . here
, @xmath60 is the round - trip loss ( @xmath61 if there are no additional losses such as clipping ) .
the cavity finesse is @xmath62 in contrast to the quality factor @xmath63 , the finesse depends only on the properties of the mirror coatings and not on the cavity length ( as long as the mirror diameters @xmath53 are large enough to neglect clipping loss
see below ) . in many applications ,
the foremost requirement is a small waist @xmath9 in order to optimize coupling to an emitter located inside the cavity . in the symmetric case @xmath64 , @xmath65 ( fig . [ fig : fceff ] , left ) . with macroscopic supermirrors ,
the interesting region @xmath66 m is only accessible in the near - concentric regime @xmath67 , which is difficult or impossible to exploit due to the extreme alignment sensitivity in this regime . existing macroscopic fp cavities ( fpcs ) rather have @xmath68 m .
the ffpc design makes it possible to enter the interesting regime of small @xmath54 and small @xmath25 simultaneously , which is inaccessible to macroscopic cavities .
typical @xmath25 values for our laser - machined fiber mirrors are in the @xmath69 m range , two to three orders of magnitude smaller than for traditional high - finesse fpcs . additionally , for a given @xmath25 , the length of the fiber cavity can be made much smaller than for its macroscopic counterpart , because of the smaller mirror diameter .
( the length limit is reached when the mirrors touch @xcite . )
both factors contribute to enable exceptionally small waists for an open cavity .
the full expression of @xmath9 for non - symmetric cavities can be found in textbooks such as @xcite . as with symmetric cavities ,
small waists are obtained for short cavities ( @xmath70 ) and close to the stability limits ( @xmath71 and @xmath72 if furthermore @xmath73 and @xmath74 are substantially different , the minimum waist size is determined by the smaller of the two .
for example , if @xmath75 , then @xmath76 this limit particularly applies to half - symmetric cavities ( @xmath77 ) .
interestingly , comparison with eq .
[ eq : waist ] shows that when replacing one mirror of a short symmetric cavity by a planar one ( leaving @xmath54 unchanged ) increases @xmath9 by less than @xmath78 .
to summarize : if we exclude cavities close to the stability limit , then a small - waist cavity should be as short as possible and have at least one strongly curved mirror . to determine the smallest possible @xmath9 within our fabrication limits
, we have to consider how small @xmath54 can be made .
we can currently machine mirror structures with diameters down to @xmath79 m . as we have seen in sec .
[ sec : surface ] , the mirror depth is about @xmath80 nm for an @xmath31 m fiber mirror with @xmath32 m , so that a geometric cavity length as short as @xmath81 nm can be realized even with such a strongly curved mirror .
( note that the @xmath32 m is still large enough to avoid clipping loss , cf .
[ sec : clip ] . )
the part of the field penetrating into the multilayer stack then contributes significantly to the effective cavity length .
we account for this effect by setting @xmath82 @xcite , where the coating materials give @xmath83 in our case dominating over the minimum geometric length . taking into account this effect and leaving a small gap to introduce atoms
, it should still be possible to achieve @xmath84 m for all @xmath25 down to our current fabrication limit @xmath31 m .
a symmetric cavity with @xmath31 m and @xmath85 m then has @xmath86 m at @xmath44 nm according to eq .
( [ eq : waist ] ) .
this is more than ten times smaller than the waist of macroscopic cqed cavities and slightly smaller than that of high - q micropillars @xcite . modeling this mode
more accurately would require taking into account light propagation inside the multilayer stack ( and not just the stack s effect on cavity length ) , and may also require going beyond the paraxial approximation @xcite .
minimizing @xmath54 and @xmath25 also minimizes the mode volume @xmath87 where the second form is for the symmetric case . with the parameters above , @xmath88 .
@xmath89 enters in cqed coupling rates , which will be discussed below in sec .
[ sec : cqedparams ] .
( red solid line ) and mode radius @xmath90 on the mirrors ( blue dashed line ) for a cavity with @xmath91 m and @xmath44 nm .
dot - dashed lines indicate the maximum @xmath90 for a mode - matching efficiency @xmath92 to a fiber with mode field radius @xmath93 m ( green line ) and for neglegible clipping loss up to @xmath94 assuming mirror diameter @xmath95 m ( orange line ) .
( b ) power coupling efficiency @xmath96 between a sm fiber ( mode field radius @xmath93 m ) with a mirror of curvature @xmath97 m and the cavity mode , as a function of the cavity mode radius @xmath98 on the mirror .
red , solid line : full model taking into account wavefront curvature and the lensing effect of the mirror surface ( eq . ( [ eq : mmfull ] ) ) .
blue , dashed line : wavefront curvature and the lensing effect are neglected ( [ eq : mmeff ] ) ) .
the approximation is very good for small @xmath98 , and remains remarkably accurate even for larger values .
the predicted maximum efficiency is @xmath99 for @xmath100 m in the approximation and @xmath101 for @xmath102 m in the full expression .
[ fig : fceff],scaledwidth=90.0% ] as the cavity mirrors are part of the incoupling and outcoupling fibers , coupling to and from the cavity is robust and stable over time .
this is one key advantage of the ffpc .
there are no mode - matching lenses , so coupling efficiency is given directly by the mode matching between the mode leaving the fiber and the cavity mode . if single - mode operation is not required such as for the output fiber in a cavity transmission measurement a multimode ( mm ) fiber can be used on the output side @xcite .
this virtually eliminates coupling loss and also makes the cavity more robust against various types of misalignment , such as centering errors between the mirror and the fiber core . but even for sm fibers , the coupling efficiency @xmath96 is typically very high , in spite of the strong mirror curvature .
for example , we have measured @xmath103 for a sm fiber and a mirror with @xmath104 m ( see sec . [
sec : couplingexp ] ) .
this high efficiency can be seen as a consequence of the small cavity mode radius , which leads to small phase difference across the relevant mode cross - sections even for strongly curved wavefronts . to get a feeling for the orders of magnitude , consider a spherical wavefront with roc @xmath105 .
this wavefront deviates from its tangential plane by @xmath14 when the transverse coordinate @xmath106 reaches @xmath107 , so even on an @xmath108 m mirror and @xmath109 m , the mode radius can be as large as @xmath79 m before it deviates from planar by @xmath14 .
typical mode field radii in sm fibers are much smaller than this .
this has several simplifying consequences : first , the lensing effect by the curved fiber surface can be neglected , and second , the power coupling efficiency @xmath96 between the fiber and cavity modes can be approximated simply by the overlap integral of the fiber and cavity mode intensity distributions , neglecting phase mismatch . for a sm fiber with its nearly gaussian transverse mode profile of radius @xmath110 , and assuming that the cavity mode is well aligned with the fiber axis , we thus obtain the simple estimate @xmath111 where @xmath98 is the cavity mode s radius on the mirror . a more complete expression including lensing effect and wavefront curvature is @xmath112 where @xmath113 is the refractive index of the fiber and @xmath25 the roc of the mirror .
this expression is derived in the appendix . what it still ignores is misalignment between the mirror and the fiber axis . in figure
[ fig : fceff](right ) , @xmath96 is plotted as a function of @xmath98 .
optimum coupling occurs around @xmath114 .
the remaining mismatch is due to mirror curvature . to minimize it
, @xmath25 should be chosen as large as possible .
@xmath99 is achieved on the planar side ( @xmath115 ) of a half - symmetric cavity where @xmath54 and the curvature of the second mirror are chosen such that @xmath116 .
a possible drawback of this configuration is its waist size , @xmath117 : @xmath118 is typically much larger than the minimum @xmath9 which ffpcs can achieve .
if a mm fiber is used on the outcoupling side , the coupling coefficient is determined by its numerical aperture @xmath119 .
@xmath120 is the acceptance angle of the fiber , which must be compared to the full divergence angle of the cavity mode .
the latter grows from 0 ( in the waist ) up to @xmath121 ( for @xmath122 ) . as an example , for a typical @xmath123 at @xmath44 nm , in the worst case of a long cavity ( @xmath124 ) , the waist can still be made as small as @xmath125 m before the mode divergence becomes larger than the acceptance angle of the fiber .
the large acceptance angle also makes the cavity more robust with respect to misalignment . in some applications ,
a larger mirror distance is required , for example to introduce an object such as a membrane @xcite ) , to minimize trap distortion and heating in an ion trap , or to reduce the fsr .
as @xmath54 is increased , the mode radius on at least one of the mirrors will also increase , and this limits the cavity length through two effects : first , fiber coupling efficiency decreases , which may or may not be a problem depending on the application .
then , cavity finesse decreases due to clipping loss . in this section
, we will consider these effects for just one of the fiber mirrors , calling @xmath12 the mirror diameter , @xmath98 the cavity mode radius on this mirror , and @xmath110 the mode field radius in its fiber .
we will first determine the maximum allowable @xmath98 as imposed by ( a ) fiber coupling and ( b ) clipping .
the maximum @xmath54 for a given @xmath98 then follows from the standard fpc formulas ( c ) .
\(a ) calling @xmath126 the target fiber coupling efficiency , and using the approximation of eq .
( [ eq : mmeff ] ) , @xmath98 must fulfill @xmath127 for example , if the fiber coupling efficiency is to be at least @xmath128 , the cavity mode radius must not be larger than @xmath129 . for the sm fibers used in our cavities ( @xmath93 m ) , this estimate yields @xmath130 m .
\(b ) to estimate clipping loss on the the finite - diameter fiber mirrors , we conservatively assume a gaussian cavity mode and consider its `` spillover '' loss upon reflection on a finite - diameter mirror . for a single reflection , @xmath131 where @xmath98 is the mode radius of the mode impinging on the mirror of diameter @xmath12 . if @xmath132 is to contribute less than 10% to the total loss , @xmath133 the mode radius on the mirror must verify @xmath134 in the finesse range of interest here , @xmath135 . with our current fabrication limit of @xmath136
m , we thus have @xmath137 m in the most demanding case ( @xmath138 ) .
this limitation is generally less restrictive than the one imposed by good fiber coupling .
\(c ) now let us calculate the maximum @xmath54 by requiring that the actual @xmath98 be smaller than these maximum values . from the elementary fpc formulas , we have for a symmetric cavity @xmath139 for a given pair of identical mirrors , the maximum length is @xmath140 with @xmath141 or @xmath142 depending on the case . obviously , if @xmath143 , this becomes much shorter than the stability limit .
if the application also imposes @xmath9 , the length limit follows from the first expression in eq .
( [ eq : w12sym ] ) : @xmath144 for example , with @xmath145 m ( low clipping loss ) and @xmath44 nm , we find that a cavity with @xmath146 m can be made up to @xmath147 m long .
the curvature of the cavity mirrors is then @xmath148 m . requiring instead @xmath149 m ( good mode matching ) leads to @xmath150 m and @xmath151 m .
if there are no restrictions on the mode waist , as might be the case for a filter cavity or a membrane cooling experiment for example , @xmath25 can be optimized to further increase @xmath54 while keeping @xmath90 small ( right expression in eq .
( [ eq : w12sym ] ) ) .
the length limit for the symmetric cavity case then becomes @xmath152 and is attained for a confocal cavity ( @xmath153 ) . in other words , for a given , long @xmath54 , the confocal configuration has the smallest mode radii on the mirrors , and therefore optimizes both mode matching and finesse .
taking again @xmath44 nm and @xmath145 m ( low clipping loss ) , we find @xmath154 mm , which is of the same order as the limit imposed by the stability criterion for our maximum roc .
thus , within this simple model , clipping loss in confocal cavities remains neglegible except for very long wavelength and the confocal cavity length is limited to @xmath155 m by the attainable roc .
note however that our simple model does not include other imperfections , such as misalignment between the fiber axis and the center of the mirror . in any case
, it should be possible to make still longer cavities from one fiber mirror and one macroscopic mirror with larger diameter and roc . the more restrictive requirement @xmath149 m ( good mode matching ) leads to @xmath156 m .
this is the longest cavity that can be made if the power transmission between fiber and cavity modes is to be at least @xmath128 . for comparison ,
the @xmath157 m confocal cavity has @xmath158 , which is still an acceptable value in many cases . for @xmath159
, @xmath160 still generally grows with @xmath54 except in the regions near @xmath161 and @xmath162 , where the resonator becomes unstable . in cavity qed problems ,
the cavity parameters enter in the form of the coherent single - photon coupling rate @xmath163 and the incoherent cavity decay rate @xmath164 .
it is instructive to consider these rates as functions of @xmath9 and @xmath54 ( choice of a gaussian mode ) , and alternatively as functions of @xmath25 and @xmath54 ( choice of a pair of mirrors ) , where we restrict ourselves to the symmetric case for simplicity . for a cavity mode with frequency @xmath165 and an atom or other emitter ( dipole matrix element @xmath34 ) located at maximum field intensity , the coherent coupling rate is @xmath166 where @xmath167 is the hwhm linewidth of the excited state , @xmath168 is the dielectric constant at the location of the emitter ( @xmath169 for free space ) .
the second form applies to a two - level atom in free space ; for simplicity we will use this form from now on .
@xmath170 depends on the cavity parameters through the mode volume ( ( eq .
[ eq : modevol ] ) ) .
[ fig : cqedparams ] shows an example .
we have already determined the minimum achievable mode volume , @xmath171 for @xmath44 nm ( sec .
[ sec : minlength ] ) .
this leads to a maximum coupling rate of @xmath172ghz for rb atoms , about four times the most optimistic projected limit @xcite of macroscopic fpcs .
note that the limit of @xcite assumes significant future improvements in polishing technology , while our value is calculated with the surface quality and roc that have already been fabricated and measured .
maximizing @xmath170 requires minimizing @xmath89 , so reducing the mirror spacing ( and adapting the roc accordingly ) maximizes @xmath170 for a given @xmath9 .
however , doing so also increases the cavity decay rate ( eq . [ eq : kappa ] ) .
if the goal is to enter as far as possible into the `` strong coupling regime '' ( i.e. , @xmath173 , or equivalently , resolved coupled - system resonances ) , the coupling - to - dissipation ratio @xmath174 may be used as a figure of merit . if the cavity is short more precisely , if @xmath175 , then @xmath176 , so that this ratio is optimized by maximizing @xmath177 . with @xmath178 and @xmath179mhz , @xmath180 m , which means that typically @xmath176 over the full stability range . expressing this ratio as a function of the mode and mirror parameters : @xmath181 fig .
[ fig : cqedparams ] shows an example . as expected , choosing a small @xmath9 improves this ratio .
somewhat surprisingly , the ratio also improves with growing cavity length , despite the growing mode volume .
this is because @xmath164 depends on @xmath54 more strongly than does @xmath170 . for long cavities ,
the ratio is determined by @xmath182 , and decreases again . by contrast , in a large class of applications which notably contains single - photon sources , a more important figure of merit is the single - atom cooperativity @xmath183 , @xmath184 @xmath183 determines the purcell factor ; the probability for a spontaneous photon to be emitted into the cavity mode is @xmath185 .
the inverse of @xmath183 is called the critical atom number .
in contrast to @xmath170 , @xmath183 does not depend on the mode volume , but only on the mode waist @xmath9 .
another way of seeing this is by realizing that @xmath183 is effectively the optical density of the emitter , which depends on the ratio of its absorption cross - section to the cross - section of the mode , but is independent of the cavity length ( as long as @xmath54 is larger than the emitter size ) .
the conclusion is that the choice of a particular gaussian mode fixes the value of @xmath183 , no matter where we decide to place the mirrors that confine this mode .
the strong curvature of our fiber mirrors enables very small waist size , we have already seen that @xmath86 m is a realistic value .
combined with the maximum finesse @xmath138 ( sec .
[ sec : surface ] ) , eq .
( [ eq : c0_2lvl ] ) predicts @xmath186 for the maximum cooperativity that can be achieved at @xmath44 nm . to conclude this section , fig .
[ fig : cqedparams ] shows cqed parameters as a function of cavity length @xmath54 , taking into account clipping loss , for the following parameters : @xmath187 and @xmath188ppm per mirror . , coupling to rb atoms ( @xmath189 ) .
( a ) finesse ( blue solid line ) and q factor ( red dot - dashed line ) .
the decay for @xmath190 m is caused by clipping , and is responsible for the increase in @xmath164 and decrease in @xmath2 at these lengths .
( b ) @xmath170 without clipping ( red , dot - dashed ) , @xmath164 without clipping ( magenta , dashed ) and with clipping loss ( solid , blue ) . throughout the length range , @xmath164 is larger than @xmath167 ( orange , dotted ) .
( c ) @xmath2 without clipping ( magenta , dashed ) and with clipping ( blue , solid ) . even without clipping loss , the high cooperativity near @xmath191 m would be difficult to reach in practice , as it occurs at the limit of the stability range .
( d ) the ratio @xmath177 without ( magenta , dashed ) and with clipping ( blue , solid ) .
as @xmath176 , this ratio quantifies the strong coupling .
in contrast to @xmath183 , it increases for longer cavities as long as clipping is neglegible .
[ fig : cqedparams],title="fig:",scaledwidth=45.0% ] , coupling to rb atoms ( @xmath189 ) . ( a ) finesse ( blue solid line ) and q factor ( red dot - dashed line ) .
the decay for @xmath190 m is caused by clipping , and is responsible for the increase in @xmath164 and decrease in @xmath2 at these lengths .
( b ) @xmath170 without clipping ( red , dot - dashed ) , @xmath164 without clipping ( magenta , dashed ) and with clipping loss ( solid , blue ) . throughout the length range , @xmath164 is larger than @xmath167 ( orange , dotted ) .
( c ) @xmath2 without clipping ( magenta , dashed ) and with clipping ( blue , solid ) . even without clipping loss , the high cooperativity near @xmath191 m would be difficult to reach in practice , as it occurs at the limit of the stability range .
( d ) the ratio @xmath177 without ( magenta , dashed ) and with clipping ( blue , solid ) .
as @xmath176 , this ratio quantifies the strong coupling .
in contrast to @xmath183 , it increases for longer cavities as long as clipping is neglegible .
[ fig : cqedparams],title="fig:",scaledwidth=45.0% ] , coupling to rb atoms ( @xmath189 ) .
( a ) finesse ( blue solid line ) and q factor ( red dot - dashed line ) .
the decay for @xmath190 m is caused by clipping , and is responsible for the increase in @xmath164 and decrease in @xmath2 at these lengths .
( b ) @xmath170 without clipping ( red , dot - dashed ) , @xmath164 without clipping ( magenta , dashed ) and with clipping loss ( solid , blue ) . throughout the length range , @xmath164 is larger than @xmath167 ( orange , dotted ) .
( c ) @xmath2 without clipping ( magenta , dashed ) and with clipping ( blue , solid ) . even without clipping loss , the high cooperativity near @xmath191 m would be difficult to reach in practice , as it occurs at the limit of the stability range .
( d ) the ratio @xmath177 without ( magenta , dashed ) and with clipping ( blue , solid ) .
as @xmath176 , this ratio quantifies the strong coupling .
in contrast to @xmath183 , it increases for longer cavities as long as clipping is neglegible .
[ fig : cqedparams],title="fig:",scaledwidth=45.0% ] , coupling to rb atoms ( @xmath189 ) . ( a ) finesse ( blue solid line ) and q factor ( red dot - dashed line ) .
the decay for @xmath190 m is caused by clipping , and is responsible for the increase in @xmath164 and decrease in @xmath2 at these lengths .
( b ) @xmath170 without clipping ( red , dot - dashed ) , @xmath164 without clipping ( magenta , dashed ) and with clipping loss ( solid , blue ) . throughout the length range ,
@xmath164 is larger than @xmath167 ( orange , dotted ) .
( c ) @xmath2 without clipping ( magenta , dashed ) and with clipping ( blue , solid ) . even without clipping loss , the high cooperativity near @xmath191 m would be difficult to reach in practice , as it occurs at the limit of the stability range .
( d ) the ratio @xmath177 without ( magenta , dashed ) and with clipping ( blue , solid ) .
as @xmath176 , this ratio quantifies the strong coupling .
in contrast to @xmath183 , it increases for longer cavities as long as clipping is neglegible .
[ fig : cqedparams],title="fig:",scaledwidth=45.0% ]
we have measured transmission and reflection spectra as well as insertion loss for a variety of ffpcs with mirrors of various curvatures fabricated on sm and mm fibers .
all of them had coatings fabricated by laserzentrum hannover ( lzh ) , designed for maximum reflectivity at 780 nm .
note that these are not the best commercially available coatings .
we have chosen this supplier as a compromise between coating quality and delay at the time we needed the coatings ; the measurements below indicate that still better cavity performance can be achieved by coating the same fibers with `` supermirror '' coatings , as in @xcite for example .
the most comprehensive measurements were done on two cavities , ffp1 and ffp2 , which are part of an atom chip experiment in paris in which we investigate cqed effects with @xmath192rb bose - einstein condensates @xcite and detection of trapped single atoms .
these cavities have been in vacuum for more than 2 years now ; periodic remeasurements show no degradation of their performance in spite of their exposition to low - pressure rb vapor .
while measuring the transmission and reflection spectra is fairly straightforward , other measurments such as insertion loss and mirror losses tend to be more difficult than with macroscopic cavities due to the inherent fiber coupling . in the following subsections
we describe our methods and show results for the various parameters .
we have laser - machined a batch of sm and mm gradient index fibers with @xmath13 m cladding diameter and @xmath193 m mode field diameter / @xmath11 m core diameter , respectively .
the fibers are metal coated ( cu / cu alloy ) , which makes them suitable for ultra - high vacuum use and also shields stray light .
structures with roc from @xmath194 m to @xmath195 m and mirror diameters from @xmath196 m to @xmath17 m were coated in an ion beam sputtering process by lzh .
each fiber was cleaned in aquaeus solution of hcl ( h@xmath0o : hcl 5:1 ) for 2 min in an ultrasonic bath ( usb ) , then rinsed for 2 min ( usb ) in h@xmath0o and finally for 2 min ( usb ) in acetone . after cleaning ,
the surfaces were individually inspected for contaminations and then inserted into a purpose - built holding plate .
this procedure was carried out immediately before coating , but outside the cleanroom in which the coating took place .
the dielectric coating has 14 layers of sio@xmath0 with refractive index @xmath197 and 15 layers ta@xmath0o@xmath198 with @xmath199 .
the calculated transmission of the layer stack is @xmath200ppm at @xmath201 nm .
reference substrates with @xmath202 nm rms roughness were coated in the same run .
a microscope inspection of several coated fibers showed that 30 - 50% of the end faces contained contaminations , some of them making the fibers unemployable .
the most likely explanation is that dust remaining on the holding plate contaminated the fibers when they were inserted . to determine the quality of the coatings , we have done various measurements both with reference substrates and with fibers .
the lzh carried out a calorimetric absorption loss measurement at 1064 nm which yielded @xmath203ppm and an ulbricht sphere total scatter measurement at 633 nm which gave @xmath204ppm .
our afm measurements of the coated reference substrates show a roughness of @xmath205 nm rms , an increase of at least 0.1 nm over the value before coating and corresponding to scatter loss @xmath43ppm .
our direct measurement of the transmission of reference substrates yielded @xmath206ppm , close to the calculated value .
we also built macroscopic cavities from the reference substrates and measured their decay time constant @xmath207 using a fast scan ringdown method @xcite .
the resulting value @xmath208s obtained for @xmath209 mm corresponds to a finesse @xmath210 and total loss per mirror @xmath211ppm .
table [ tab : mirrorlosses ] summarizes these results .
the deviations between the different measurement methods suggest isolated defects , possibly the dust particles mentioned above .
quantity&method¯oscopic substrates&fibers + absorption loss @xmath212&calorimetric @ 1064nm&@xmath213ppm & + & bistability&&@xmath214ppm + scatter loss @xmath215&ulbricht sphere @633nm&@xmath216ppm & + & afm roughness&10ppm&@xmath217ppm + transmission @xmath218&direct&@xmath219ppm & + @xmath220&finesse ( fast ringdown)&101ppm & + & finesse ( cavity scan)&&@xmath221ppm + the total losses for the fiber mirrors were obtained from finesse measurements with short ffpcs , which are discussed in more detail below .
measurements on several cavities yield @xmath222 , corresponding to total loss @xmath223ppm . using the value @xmath206ppm obtained for the macroscopic substrates
, we deduce @xmath224ppm .
an independent estimate of @xmath215 comes from the afm measurements . like the reference substrates , the coated fibers exhibit an increased surface roughness , in this case to @xmath225 nm rms .
using eq .
[ eq : scattloss ] , this corresponds to @xmath226ppm .
absorption loss can be estimated from absorption - induced bistability of the cavity ( see section [ sec : bistability ] ) ; we obtained @xmath227ppm .
the sum of these individual estimates is @xmath228ppm .
( considering the errors of the individual measurements , the very good agreement with @xmath229 is partly fortuitous . )
these measurements consistently indicate that the losses of the ffpcs are dominated by the coatings .
in particular , finesse values obtained with macroscopic and fiber mirrors are very similar , in spite of the significantly lower roughness of the macroscopic substrates before coating .
the afm measurements reveal that , indeed , the coatings significantly increased the surface roughness .
( for state - of - the art `` supermirror '' coatings , it is known that this does not occur @xcite . )
there is no indication that the small roc of the fiber mirrors would cause any significant reduction of the coating quality as compared to the reference substrates .
using state - of - the art coatings should make it possible to fully exploit the surface quality of the laser - machined fibers .
fibers were angle - cleaved at their coupling ports and mounted either in a macroscopic mount based on translation stages ( `` test mount '' ) , or in a dedicated `` miniature mount '' involving v - grooves and shear piezos on a ceramic baseplate made from macor ( figure [ fig : mount ] ) .
the miniature mount has better passive stability , as expected from its low profile and small overall size .
, scaledwidth=49.0% ] in the test mount , fibers are clamped into side - loaded ferrules that are fixed on two micropositioning stages .
together they provide three axes of translation , including fine tunability by integrated piezos , and two angular degrees of freedom . for the miniature mount
, we use a similar micropositioning setup to align the fibers during assembly .
after glueing , the only adjustable parameter is cavity length within the displacement range of the shear piezos , which is several fsrs .
all other alignment is done during assembly , where we apply and cure epoxy glues while monitoring cavity transmission ( `` active alignment '' ) .
we use epo - tek 353nd and epo - tek 301 to glue the components together .
both are uhv compatible down to @xmath230 mbar .
cavity alignment is observed visually with a stereo microscope ( magnification up to 63x ) and magnifying video cameras from two axes . after a geometrical preadjustment
, the alignment is optimized by maximizing the overall transmission of the fiber cavity , which is scanned over at least one fsr continuously .
we did not observe any degradation of the alignment of the glued cavity over time , nor did baking to @xmath231c leave a measurable irreversible effect once the system had cooled down .
as a first method to characterize different ffpcs over the whole usable cavity length , we measure transmission spectra at fixed laser frequency , using a piezo actuator to scan @xmath54 ( fig .
[ fig : cavscan ] ) . this can be done with both the test mount and the miniature mount . to obtain a calibration of the length scale ,
we have used various methods to simultaneously couple several laser frequencies with a well - known difference . to obtain a difference in the ghz range , we use two lasers , typically @xmath232 nm , locked on the @xmath192rb d2 line , @xmath233 nm measured with a 6 digit burleigh wavemeter , @xmath234ghz ) . with this wavelength calibration , the fsr and the distance @xmath235 can be determined with an uncertainty below @xmath236 nm . to measure the cavity linewidth ,
a smaller frequency difference is useful , and can be obtained by modulating a single laser at a known frequency in the rf range ( fig .
[ fig : cavscan ] , left ) .
@xmath5 is obtained as the ratio of the fsr and linewidth results .
for a set of three different fiber pairs , we have performed finesse measurements over the whole length range of the cavities ( figure [ fig : fvsl ] ) .
all cavities had sm fibers on the input side and mm fibers at the output , and roc was large on the sm side and small on the mm side : @xmath73 between 300 and @xmath237 m , @xmath74 between 60 and @xmath238 m , @xmath239 between 20 and @xmath17 m , and @xmath240 between 30 and @xmath17 m . for small mirror distance , we obtain a finesse of @xmath241 .
the large scatter comes from variation in the alignment quality and from the limited mechanical stability of the test mount .
the finesse drops significantly for @xmath242 m and the resonances vanish in the noise level above @xmath243 m .
this behaviour is well reproduced by including clipping loss ( eq . [ eq : cliploss ] ) into the calculated finesse : the solid curve in fig .
[ fig : fvsl ] is calculated for @xmath244 m , @xmath245 m and @xmath246 m .
note that the finesse drops sharply over a small distance range .
the `` cutoff '' length at which this occurs depends strongly on the effective mirror diameters . for the cavities in fig .
[ fig : fvsl ] , where the waist is located close to mirror 1 , it is dominated by @xmath240 .
it will be interesting to repeat such @xmath247 measurements for mirrors with different diameters , as this will lead to more accurate predictions of the @xmath12 values required to reach a desired cavity length .
mode is more than 20 times stronger than all higher order modes .
right : _ linewidth of the same cavity , measured in transmission .
the laser is modulated with sidebands at @xmath248mhz for frequency calibration .
the measurement yields a fwhm linewidth of 100.4mhz .
[ fig : cavscan],title="fig:",scaledwidth=49.0% ] mode is more than 20 times stronger than all higher order modes . _ right : _ linewidth of the same cavity , measured in transmission .
the laser is modulated with sidebands at @xmath248mhz for frequency calibration .
the measurement yields a fwhm linewidth of 100.4mhz .
[ fig : cavscan],title="fig:",scaledwidth=49.0% ] .
the solid curve is the calculated finesse including clipping loss for @xmath244 m , @xmath245 m and @xmath246 m . [ fig : fvsl],scaledwidth=40.0% ] for cavities ffp1 and ffp2 , mounted in the miniature mount , we used an improved version of the above protocol .
both lasers were locked to rb resonances , with a frequency difference of 212mhz . in this setup , the length could be determined to better than @xmath14 so that the resonance order is exactly known : @xmath249 m , @xmath250 m . with a frequency calibration via amplitude modulation of the laser ,
we obtain a value for the resonance linewidth of @xmath251mhz and @xmath252mhz ( fig .
[ fig : cavscan ] ) . combining these measurements yields
@xmath253 , @xmath254 . for the ffp1 cavity
, we have also measured the dependence on the polarization of the incoming beam .
we observe a splitting of @xmath255mhz between two orthogonal , linear input polarizations .
similar birefringence has been observed in macroscopic high - finesse cavities . in our case
, it may be related to the ellipticity observed on some of the mirrors .
we have measured transmitted and reflected powers for several cavities in the test and miniature mounts .
our main goal was to obtain information about the coupling efficiency @xmath96 between a sm fiber and the cavity .
we have investigated this question with cavities having a sm fiber at the input and a mm fiber on the output side .
the on - resonance total transmission ( from the free - space beam before the first fiber to the beam leaving the second fiber ) is @xmath256 the analysis is more complicated than for macroscopic fpcs because neither the coupling efficiency from the free - space beam into the incoupling fiber , @xmath257 , nor the mode - matching efficiencies between the fibers and the resonator mode , @xmath258 , can be determined directly in a non - destructive way .
( it would be possible to measure @xmath257 by breaking the fiber . ) the measured intensity transmission from before the input of the sm input fiber to after the output of the mm output fiber is 0.094 on resonance for ffp1 .
( similar results were obtained for other cavities . ) from the mirror transmission measured on the macroscopic substrates , @xmath259ppm , and @xmath260ppm , the on - resonance transmission of the perfectly mode - matched cavity would be @xmath261 .
( note again that this value is limited by the sub - optimum coatings . )
we infer that @xmath262 .
a lower bound for @xmath263 can be obtained by attributing all this loss to the input mode - matching , so @xmath264 . in reality , @xmath257 is responsible for a significant part of the losses .
we have estimated @xmath257 by replacing the cavity input fiber with an `` open '' fiber ( no mirror on the output side ) of the same type . in this case
, the fiber coupling efficiency is easily measured , and never exceeded @xmath265 in our setup .
a more realistic value is therefore @xmath266 .
the theoretical maximum value for perfect alignment ( eq . [ eq : mmfull ] ) is @xmath267 .
some additional information can be obtained from a reflection measurement .
let us call @xmath268 the power incident on the fiber input port and @xmath269 the power that emanates from the same port in the reverse direction ( which we can measure using a beam splitter ) .
off resonance or with the second fiber removed , we have @xmath270 here we assume that all light from the input beam which is not coupled into the fiber is lost and contributes neglegibly to the reflected beam .
the parameter @xmath271 takes into account that some of the light reflected back from the mirror is not guided by the fiber because the mirror is not planar and may be misaligned .
( note that in general @xmath272 because the reflected mode is not identical to the cavity mode . )
a measurement of @xmath269 yields the product @xmath273 . for ffp1
, we obtained @xmath274 .
measuring @xmath257 with an open fiber gave @xmath275 in this case , so we obtain @xmath276 . for various test cavities , we obtained @xmath277 . on resonance ,
a symmetric cavity excited by a perfectly mode - matched power @xmath278 reflects @xmath279 again , we can not measure @xmath278 and @xmath280 directly .
@xmath269 has two contributions @xcite on resonance : the cavity , which is now excited by the mode - matched power @xmath281 , reflects according to eq . [ eq : mmref ] .
the mode - matched part of this reflection , @xmath282 , is guided back by the fiber .
second , the non - modematched component of the input light , @xmath283 is reflected on the mirror and a fraction @xmath284 of it guided back . unfortunately , @xmath285 because the two constants relate to different incident modes .
( we expect @xmath286 . ) in total , we have @xmath287 because of the additional unknown constant @xmath284 , this formula is difficult to exploit .
strong coupling between an emitter and a resonant cavity gives rise to a split resonance , displaced by @xmath288 with respect to the empty cavity resonance , where @xmath163 is the coherent coupling rate ( `` vacuum rabi splitting '' ) .
measuring the resonance frequencies of the coupled system can thus be used to determine @xmath163 .
the coupling rate @xmath170 in sec .
[ sec : cqedparams ] is the maximum coupling calculated for a point - like particle localized in an antinode of the cavity field .
it is straightforward to extend to the position - dependent coupling @xmath289 .
furthermore , for @xmath290 independent atoms with a density distribution @xmath291 , one finds : @xmath292 in @xcite , we have measured @xmath293 for bose - einstein condensates with variable @xmath290 .
the atoms were strongly confined and well localized in the region of maximum coupling of cavity ffp1 described above .
@xmath290 was determined independently by absorption imaging .
an estimated uncertainty of about a factor 2 in the @xmath290 measurement dominates the precision to which we can determine @xmath170 from this measurement .
@xmath294 depends on @xmath290 and could in principle be taken into account precisely . however , within the uncertainty imposed by the @xmath290 measurement , the distribution can be considered point - like for @xmath295 .
@xmath170 can be estimated simply as the prefactor of a @xmath296 function fitted to the data for small @xmath290 .
this estimate gives @xmath297mhz , in agreement with the calculated value .
when scanning the cavity length using an optical power above a certain threshold , we find strong optical bistability induced by absorption in the mirror : the small fraction of the intracavity power that is absorbed in the mirror coating causes local heating and thermal expansion of the mirror and the fiber .
this changes the effective cavity length and therefore the cavity field , leading to bistability .
we can observe strongly broadened resonance lines when the cavity is scanned towards increasing length , showing an almost linear increase of transmitted power followed by a rapid drop .
this is explained by the expansion of the substrate compensating the length change of the scanning piezo . in the other direction ( towards shorter cavity length ) , we see a narrowed line , corresponding to the cavity being pushed across the resonance by the expansion .
figure [ fig : opticalbistability](a ) shows the characteristic lineshape for both scan directions .
1/s and @xmath298mw .
( b ) the lineshift @xmath299 as a function of incident power for @xmath3001/s . a linear fit to the data yields @xmath3011/mw.,title="fig:",scaledwidth=49.0% ] 1/s and @xmath298mw .
( b ) the lineshift @xmath299 as a function of incident power for @xmath3001/s
. a linear fit to the data yields @xmath3011/mw.,title="fig:",scaledwidth=49.0% ] from the width of the broadened resonances , the absorption in the mirror coatings can be inferred .
we measure the total width , as defined in @xcite , and express it as a multiple @xmath299 of the unperturbed ( low - power ) linewidth @xmath164 ( i.e. , the total width is @xmath302 ) . in the adiabatic limit , where the temperature distribution in the coating and fiber reaches steady state , @xmath299 is determined by @xcite @xmath303 with @xmath304 the effective thermal expansion coefficient of the coating and the fiber , @xmath305 a geometrical factor ,
@xmath278 the incident , modematched laser power in the fiber , and @xmath306 the thermal conductivity of the fiber .
adiabaticity is met for slow scan velocity @xmath307 , where @xmath308 is the scan velocity in units of the cavity linewidth and @xmath309 the thermal diffusion time with the specific heat per unit volume @xmath310 for sio@xmath0 . for our parameters , @xmath311s and
steady state conditions are expected for @xmath312 . due to the limited mechanical stability of the test mount
, we had to use a slightly faster scan speed @xmath313 , so that our measured @xmath299 values are a lower bound to the steady - state value .
[ fig : opticalbistability](b ) shows the measured @xmath299 as a function of the incident power .
we obtain @xmath314mw@xmath315 . to deduce the mirror absorption from this measurement
, the contribution of the coating to the effective thermal expansion coefficient has to be known . from a finite - element simulation
, we find that the transient temperature profile extends @xmath316 m into the fiber , and that the mirror coating contributes @xmath317 to the expansion . hence @xmath318 is dominated by the value for sio@xmath0 @xmath319k@xmath315 with a @xmath320 contribution of the ta@xmath0o@xmath198 layers with @xmath321k@xmath315 . with the measured values for the finesse and the mirror transmission , equation ( [ eq : lineshift ] )
yields an absorption loss @xmath322ppm , in good agreement with the value obtained in sec .
[ sec : coatingprops ] .
the error for this value stems from the uncertainty of the steady - state value of @xmath299 and of the ta@xmath0o@xmath198 contribution to the thermal expansion .
a more accurate determination could be obtained by measuring the temporal response of the thermal expansion @xcite . to avoid the bistability effect ,
intracavity power should be limited to @xmath323 .
as the results show , ffpcs with laser - machined mirrors combine several desirable properties in a single device .
the most important of these are a small mode waist , high finesse , efficient and robust fiber coupling , and open access to the cavity mode .
while each of these features individually can also be realized with other cavity types , their combination in a single device is unique to our knowledge , and explains the remarkable interest they are meeting since the publication of @xcite .
we expect them to be useful in a correspondingly wide range of applications .
the devices we have realized so far do not exploit yet the full potential of the laser - machined fiber surface : they are limited by the scatter and absorption losses of sub - optimum mirror coatings . an obvious next step is therefore to have the same fibers coated with `` supermirror '' @xcite coatings , which should enable simultaneous significant increase of finesse and overall transmission .
applications that we are currently investigating in collaboration with specialists from various domains include strong - coupling cavity qed with trapped ions , cavity optomechanics , and coupling to solid - state emitters such as quantum dots and color centers in diamond .
we thank richard warburton for fruitful discussions on fiber mirror fabrication , jean hare and fedja orucevic for kindly giving us access to their co@xmath0 laser setup at lkb , cens ( munich ) for access to their cleanroom , insp ( paris ) , espci ( paris ) and didier chatenay s group at the ens physics department for access to their afm and interferometric microscopes , and stephan gnster of lzh and his team for the mirror coating .
we gratefully acknowledge financial support for this work from a euryi award and the scala integrated project of the eu .
in the following we calculate the coupling efficiency between the mode of a sm fiber and the cavity mode .
both modes are assumed to be gaussian .
we take into account the lensing effect due to the concave fiber mirror , and we include the mismatch of wavefront curvatures . what is still neglected in this calculation
is the finite coating thickness and any misalignment ( centering error ) between the mirror and the fiber core . for concreteness ,
let us consider light in a sm fiber incident from the left onto the first mirror of an ffpc .
the fiber mode is characterized by its mode field radius @xmath118 .
the radius of curvature of the mirror is @xmath25 .
we also know the mode radius @xmath98 of the cavity mode on this mirror .
the power transmission coefficient between two gaussian modes is @xcite @xmath324 where @xmath325 is the distance between the mode waist positions . in our case ,
the first mode in this formula is the one leaving the fiber through the concave mirror , which has undergone diffraction as the mirror acts like a plano - concave lens . as the mirror is very thin , the mode after passing through the mirror still has the radius @xmath118 , but has wavefront curvature @xmath326 where @xmath113 is the index of the fiber ( we can neglect the very small difference between the core and cladding index ) . knowing its radius and curvature at the mirror position , the first mode is fully determined .
the second mode in eq .
( [ eq : mmfulllit ] ) is the cavity mode , which is fully determined by @xmath25 and @xmath98 .
deriving @xmath9 , @xmath327 , @xmath328 and @xmath329 and inserting into eq .
( [ eq : mmfulllit ] ) gives the final result of eq .
[ eq : mmfull ] .
neglecting the lensing effect of the mirror and the mismatch of wavefront curvature amounts to assuming @xmath330 which is a good approximation in many cases , as shown in fig .
[ fig : fceff ] . | we have realized a fiber - based fabry - perot cavity with co@xmath0 laser - machined mirrors .
it combines very small size , high finesse @xmath1 , small waist and mode volume , and good mode matching between the fiber and cavity modes .
this combination of features is a major advance for cavity quantum electrodynamics ( cqed ) , as shown in recent cqed experiments with bose - einstein condensates enabled by this cavity [ y. colombe et al . ,
nature * 450 * , 272 ( 2007 ) ] .
it should also be suitable for a wide range of other applications , including coupling to solid - state emitters , gas detection at the single - particle level , fiber - coupled single - photon sources and high - resolution optical filters with large stopband . |
recent studies of the kinematics of the nlrs of seyfert galaxies have taken advantage of the high resolution of _
hst_to map the velocities of these regions .
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) .
although there have been a number of papers on the dynamical aspects of the nlr , most of these studies relied on ground - based data limited to spatial resolutions of @xmath1 50 pc for even the most nearby agn ( see * ? ? ?
* and references therein ) .
these studies relied primarily upon spatially integrated line profiles to understand the dynamics of the nlr @xcite , but the problem arose that the emission - line profiles of the nlr can be explained by many different types of dynamical models , such as infall , rotation and outflows @xcite . with the launch of _ hst_and its high angular resolution ( @xmath0 0@xmath21 ) , the nlrs of seyfert galaxies have received considerable attention . with the limited long - slit capability of the faint object camera ( foc ) , and later the expanded capability of stis , detailed constraints on the kinematics of the nlrs in seyfert and
other galaxies became possible . in turn , these kinematic studies provided good diagnostics upon which dynamical analyses can be based .
the structure of the nlr resembles a bicone as expected from a simple unified model of seyfert galaxies , due to collimation by a thick torus @xcite .
both @xcite and @xcite have modeled the nlr of ngc 4151 and found it to be consistent with a biconical geometry . in an _
hst_study done on a sample of seyfert 1 and 2 galaxies , @xcite compared both seyfert types to study their nlr morphologies , and found triangular structures in most of their seyfert 2s , and circular structures in most of their seyfert 1s , consistent with the unified model and biconical structure for the nlr ( see also * ? ? ?
* ; * ? ? ?
@xcite modeled the inner regions of the seyfert galaxy ngc 2992 , and found that the ionized gas can be fitted with a biconical structure .
we have recently completed a study of the kinematics of the nlrs in two seyfert galaxies ( @xcite 2005 , hereafter paper i , and @xcite 2006 , hereafter paper ii ) . since
our dynamical work in this paper is a direct extension of the works of those two papers , we summarize our results below . in paper
i , kinematic models were developed to match the emission - line velocities from high - resolution stis spectra within @xmath0 400 pc of the central black hole of the seyfert 1 galaxy ngc 4151 .
the nlr gas showed strong evidence of acceleration from @xmath0 0 pc out to @xmath0 100 pc , then showed deceleration back to systemic velocity at @xmath0 400 pc with velocity roughly proportional to distance in each case .
the maximum velocity of the outflowing gas at the turnover point ( 96 pc ) was @xmath0 800 km s@xmath3relative to the black hole .
based on our kinematic model , the nlr could be represented by a bicone , with inner and outer half opening angles of 15 and 33 respectively , and inclination of @xmath0 45 with respect to the plane of the sky , consistent with previous kinematic work done on ngc 4151 with different slit positions @xcite .
some of the fainter nlr clouds showed evidence of backflow at the point where the clouds turnover in their velocities .
the radio jet was found to have little effect on the kinematics of the nlr clouds , however there was some evidence of radial velocity splitting of the fainter nlr clouds near bright knots in the radio jet .
the brighter clouds were not accelerated by the jet . in paper ii , we developed a similar model for the seyfert 2 galaxy ngc 1068 again with high - resolution spectra taken with the stis aboard _
hst_. with seven parallel slit positions covering the entire nlr , we extracted radial velocity profiles and matched them with our newly developed 3-d biconical models .
our kinematical models showed that the nlr gas accelerated out to @xmath0 140 pc ( the turnover point ) , then decelerated back to systemic velocity at a distance of @xmath0 400 pc from the central black hole .
the maximum velocity of outflow of the gas was @xmath0 2000 km s@xmath3 , with respect to the black hole , and the model predicted and inner and outer half opening angles of the bicone of 20 and 40 respectively , and an inclination of 5 out of the plane of the sky .
we used high resolution radio maps of the nlr obtained from @xcite to search for jet / cloud interactions .
evidence showed that , similar to ngc 4151 , the fainter nlr clouds were split near bright knots in the radio jets , whereas the brighter nlr clouds remained unaffected by the jet .
other seyfert galaxies show similar flow patterns to ngc 4151 and ngc 1068 , such as the seyfert 2 galaxy mrk 3 @xcite .
the radial velocities of these galaxies were matched with a common kinematic model with little variation in the parameter space , which begs the question , what are the physical processes involve that would cause such a similar flow pattern in both types of seyfert galaxies ?
it is this question that motivated us to carry out this dynamical study . in the first truly dynamical model of the nlr based on _
hst_kinematics , @xcite attempted to fit the nlr velocities in ngc 4151 , based on measurements done by @xcite .
they tested an isothermal parker wind model which assumes thermally expanding winds @xcite . by assuming a spherical cloud geometry
, they let the parker wind drag along the embedded nlr clouds .
as the parker wind accelerates by thermal expansion and slowly loses heat by adiabatic cooling , the clouds are also accelerated to high speeds .
they then let the parker wind run into a low density ambient medium to slow the velocity of the wind , and hence the clouds , to the systemic velocity .
the model explains the velocity profile of the nlr of ngc 4151 , but suffers from the fact that an isothermal wind can not be sustained out to large distances .
also , the mass profile of the smbh plus galaxy , which determines the temperature profile for their model , is not exactly known for ngc 4151 .
hence while their model was not successful on physical grounds , it is worth noting its relative success in matching the nlr kinematics . in this paper ,
the main question we want address is : how can we constrain the dynamics of the nlr in seyfert galaxies with the detailed knowledge that we have gained from our kinematic studies ? here , we concentrate on the dynamics of the nlr in ngc 1068 , since it has the best constraints
. we will start with a simple construction of the enclosed mass function based on data from previous studies and eventually formulate a radiation pressure - gravity tug - of - war on the nlr clouds .
the questions we will attempt to answer include the following : 1 ) if the nlr gas is in outflow , is radiation pressure really the best driving mechanism ?
2 ) if the nlr gas is turning over its velocity and decelerating back to systemic , is gravity responsible for stopping the gas ?
3 ) can we fit the velocity profile of the data with a simple radiation - gravity law , or do we need to include another force ( such as drag ) ? our analysis applies to ngc 1068 in particular , but has relevance to all seyferts in general that show signs of gas outflow and subsequent deceleration @xcite . to test whether gravity is playing any role in stopping and turning back the gas velocity , we will construct a mass profile within @xmath4 10,000 pc from the nucleus of ngc 1068 . with the mass profile in hand , we will test whether the gas kinematics are dominated by rotation .
such a test might prove fruitful for rotation in ngc 4151 , which shows redshifts northeast and blueshifts southwest of the nucleus , but from a geometrical point of view , rotation of the gas will prove difficult to match the velocity field of ngc 1068 , which shows blueshifts and redshifts on each side of the nucleus .
next we plot the escape velocity with distance to see if the gas should escape or not , given the velocities seen in the data , and whether or not the kinematics of the nlr can be dominated by gravitational infall .
then we concentrate on outflow assuming spherical symmetry and pure radial motion .
first we determine if the deceleration of the gas can be attributed to the enclosed mass , regardless of the outward accelerating force .
we then apply radiative line driving plus gravitational forces and compare the results to the observed velocity law of ngc 1068 derived from the kinematic models .
finally , we introduce a drag force due to an ambient medium on the nlr clouds , in addition to radiation pressure and gravity , to determine if it can improve the fit to either the accelerating or decelerating portions of the observed velocity curve .
in building a model of the enclosed mass as a function of distance from the central smbh , we incorporate various subsystems into our mass profile . these include contributions from the smbh , the nuclear stellar cluster , and the bulge . for all these systems ,
we assume spherical symmetry for simplicity .
we have assumed that the stellar cluster and bulge extend all the way inward to the smbh , which may overestimate the mass close in .
the size of the stellar cluster was estimated to be @xmath0 140 pc based on a study by @xcite and the bulge was assumed to extend up to 10000 pc .
mass contribution from the galactic disk of ngc 1068 was neglected because the galactic potential is dominated by the large bulge to at least 1000 pc , which is well beyond the extent of the nlr .
ngc 1068 is one of only a few agn that shows an edge - on disk of h@xmath50maser emission close to the smbh .
the disk shows the signature of a rotational velocity curve , which can be used to determine the mass of the smbh .
@xcite estimated the mass of ngc 1068 to be 1.5@xmath610@xmath7within 0.65 pc , based on the velocity field of the h@xmath50maser emission observed with the vlba and the vla .
we will therefore use this estimate for the mass of the smbh .
since we are going out to @xmath4 10,000 pc , which encloses the nlr and the enlr clouds , we need an accurate assessment of the total mass within this region . at small distances ( @xmath8 1 pc ) , the smbh is dominant , although in the case of ngc 1068 , a concentrated stellar cluster is also providing substantial gravity close in . for an estimate of the bulge mass in ngc 1068 , we rely on the work of @xcite .
they found a tight correlation between black - hole mass and bulge mass for a sample of 30 galaxies , including ngc 1068 .
they determined the bulge mass by modeling the bulge with the jeans equation in spherical form .
they assumed the bulge to be isotropic and spherically symmetric , which might lead to an overestimation of the bulge mass ; however they also neglect any contribution from dark matter , which would tend to underestimate the bulge mass .
their value for the bulge mass in ngc 1068 is 2.3@xmath610@xmath9within a radius @xmath10 = 3r@xmath11 ( 3 effective radii ) , where r@xmath11 = 3.1 @xmath12 0.8 kpc , a value taken from the surface brightness deconvolution of @xcite .
the effective radius r@xmath11 is defined to be such that half of the total light from the galaxy is predicted to be contained within the isophotal ellipse that has area @xmath13 @xcite .
elliptical galaxies and bulges surface brightnesses can be well described by the empirical formula developed by @xcite @xmath14},\ ] ] where @xmath15 = i(@xmath16 ) . in the 1980s1990s
, a family of stellar density curves emerged that modeled both elliptical galaxies and bulges well .
these curves are of the form @xmath17 where @xmath18 ( in pc ) is a scaling radius and m is the total mass of the bulge @xcite .
the parameter @xmath19 determines different types of models , where the @xmath19 = 2 cases corresponds to previous density models by @xcite and the @xmath19 = 1 cases to models by @xcite .
these models , when integrated over a spherical volume , yield the following enclosed mass profile : @xmath20 for the special case when @xmath19 = 1.5 , the density profile of equation [ dp ] yields a surface density distribution that closely matches the de vaucouleurs surface brightness profile of equation [ sbp ] to within 15% over nearly 4 decades in radius @xcite .
therefore we adopt the following form of the mass function as the bulge profile @xmath21 we find a suitable value for @xmath18 in equation [ mf ] by first defining the ` half - mass radius ' @xmath22 : @xmath23 which yields the following relation for @xmath18 @xmath24 @xcite has found a simple approximation for @xmath25 that depends only slightly on @xmath19 . for a @xmath26 , he found that @xmath27 therefore we can find a value for @xmath18 by using @xmath19 = 1.5 , @xmath16 given above , and equation [ cona ] .
we find that @xmath28 pc .
we also know that m(3r@xmath11 ) = 2.3@xmath610@xmath9 , so that @xmath29^{1.5}\\ & = 3.2\times10^{10}{\mbox{\ m\ensuremath{_{\odot}}}\xspace}.\tag{8}\end{aligned}\ ] ] the bulge mass distribution can finally be written as @xmath30 it is known that ngc 1068 has a compact nuclear stellar cluster @xmath31 140 pc in radius @xcite , which contributes significant mass to the total mass profile of the nlr of ngc 1068 . therefore this mass must be taken into account when deriving the mass profile .
@xcite found a mass of 6.8@xmath610@xmath32within 1 of the smbh of ngc 1068 , assuming a virialized , isotropic , and isothermal distribution of the stars .
they used the stellar velocity distribution ( @xmath33 ) found in @xcite , a value of 143 @xmath12 5 km s@xmath3at @xmath4 1 , to calculate the total dynamical mass within 1 of the nucleus of ngc 1068 .
@xmath34 where @xmath35 pc for ngc 1068 .
this mass includes contributions from the stellar cluster , the nucleus , and the bulge within a radius of 72 pc . therefore to find just the mass from the stellar cluster @xmath36 , we took out the rest of the mass contribution from the total mass .
@xmath37 making the various substitutions we have @xmath38 based on the radial surface brightness profile presented in figure 4 of @xcite , @xcite computed a power law consistent with the form @xmath39 to fit the data .
they claim that if the profile traces the stellar mass distribution , then they would expect that the mass profile would have the form m@xmath40 , given a spherical , isothermal distribution of stars in the cluster . using their distribution
we can estimate the stellar mass function based on the condition @xmath41= @xmath42 , so that @xmath43pc@xmath3 .
the stellar cluster mass function is therefore @xmath44 beyond the observed extent of the stellar cluster , we assumed that @xmath45 is a constant .
finally the total enclosed mass function for the nlr of ngc 1068 is given by @xmath46 or @xmath47
a figure representing the total mass enclosed within @xmath48 is shown in the top panel of figure [ mass_rot_esc_vel_pro ] .
the mass profiles for each contribution , the smbh , bulge , and cluster are also shown in the figure .
close in toward the nucleus , the nuclear stellar cluster dominates up to its entire extent , while the bulge takes over from there .
the black hole mass dominates at @xmath8 2 pc .
the kink in the total mass curve at @xmath4 140 pc is because the stellar cluster was cut off abruptly at this location .
we could have modeled the cluster to assume an exponential drop - off after 140 pc , but this would not have contributed much to the gravitational force exerted by the total mass . with the total mass profile
we calculate the circular rotational velocity and plot it as a function of distance as shown in the middle panel of figure [ mass_rot_esc_vel_pro ] .
the rotation velocity only depends on the enclosed mass at radius @xmath48 , and is given by the formula below @xmath49 for demonstration purposes , we show the rotational velocities as if only each mass component was present , as well as the velocity for the total mass profile .
the rotational velocity profile indicates to us that rotation can not dominate the kinematics of the nlr of ngc 1068 because many observed data points exhibit large velocities ( @xmath1 1000 km s@xmath3 ) , whereas the rotation curve never exceeds @xmath0 220 km s@xmath3at large distances ( 100 pc ) .
again the stellar cluster dominates the velocities up to @xmath4 300 pc .
the escape velocity at a given distance @xmath48 is calculated numerically from the formula @xmath50 based on the enclosed mass function , and is plotted in the bottom panel of figure [ mass_rot_esc_vel_pro ] . according to our kinematic model , the maximum velocity at the turnover radius @xmath51 = 140 pc is 2000 km s@xmath3(paper ii ) ; therefore figure [ mass_rot_esc_vel_pro ] tells us that the nlr clouds should have escaped after 140 pc , where the escape velocity is only @xmath4 500 km s@xmath3 . in the data however , clouds at 140 pc are at much higher velocity than escape velocity ; yet after the turnover point , the clouds start to decrease their velocity and return to systemic .
thus , some force other than gravity is causing the clouds to decelerate at r @xmath1 140 pc .
the infall velocity profile is equivalent to the escape velocity profile , except that the velocity vector is now directed inward .
therefore gravitational infall also can not account for the faster moving clouds at 140 pc , and in general does not match the observed velocity profile ( paper ii ) . to test the importance of the force of gravity alone on slowing down the outflowing nlr clouds , we give the clouds a maximum velocity at the turnover point and let gravity do the rest .
in other words , we assume that there is no outward driving force after the turnover point and let the clouds coast under the force of gravity .
the top panel of figure [ coasting_velocity_profiles ] ( top ) and 400 km s@xmath3(bottom ) shown in solid lines for the clouds in ngc 1068 .
the other mass components effect are shown in different linestyles represented by the key above the curves .
the clouds barely slow down out to 10,000 pc when launch with 1000 km s@xmath3 .
however , gravity was able to slow down the clouds down when they were launched with 300 km s@xmath3.,scaledwidth=45.0% ] shows that with a maximum velocity @xmath1 1000 km s@xmath3at 140 pc , there is little deceleration with radius .
however , with maximum velocities @xmath8 300 km s@xmath3gravity can slow the clouds down to rest , as seen in the bottom panel of figure [ coasting_velocity_profiles ] .
the maximum velocities in the nlrs of some seyfert galaxies are on the order of @xmath4 400 km s@xmath3@xcite , and gravitational deceleration may be important in these cases .
the kinematic model of ngc 1068 , however , shows maximum velocities of up to @xmath4 2000 km s@xmath3 , clearly out of the reach of gravitational deceleration .
gravity alone can not slow down the outflowing clouds in this case and there must be some other force or forces involved . to compound the problem ,
suppose we let radiation or some other force push on the gas while gravity is trying to pull it back
. in this case the gravitational deceleration will be even less .
however , to further explore this issue , we generate a line - driven radiation model , with gravity competing to slow the gas down .
this will eventually lead us in the direction of including a drag force to complete the analysis .
the various mechanisms to push the gas out from close to the nucleus include a radiation pressure driven wind , a thermally driven wind , or a magneto - hydrodynamic ( mhd ) wind @xcite .
the latter two methods are discussed in more detail in @xcite and @xcite , and were summarized in [ flow ] .
however , no dynamical model to date has led to a satisfactory description of the kinematics in ngc 1068 or ngc 4151 .
the radiation driven wind mechanism is most efficient when the momentum imparted to the gas is due to line - driving ( bound - bound transitions ) , although bound - free and free - free electron transitions ( thomson scattering ) also contribute to driving the gas out @xcite .
the radiation force is dependent on the ionization state of the gas , with lower ionization states more efficient due to the greater availability of electrons in the bound states .
if dust is mixed in with the nlr gas , then it will compete with the gas in absorbing ionizing photons and hence radiation pressure on the dust can become an important contributor to the velocities of the outflowing gas in the nlr @xcite .
if the dust grains are electrically charged , they can drag the ionized gas along to similar velocities as the dust . in this section
we ignore the effects of dust , which would only increase the radiative acceleration .
thus we consider a radiation driving mechanism coupled with the effects of gravity to find the velocity profile of the nlr gas .
we start with the acceleration due to radiation on a point mass , @xmath52 where @xmath53 is the acceleration , @xmath54 is the bolometric luminosity of ngc 1068 , @xmath55 is the thomson scattering cross section for the electron , @xmath10 is the distance , @xmath56 is the speed of light , @xmath57 is the mass of the proton , and @xmath58 is the force multiplier . as mentioned above , to really drive the gas out efficiently , we need to incorporate other sources of opacity such as bound - bound and bound - free opacity in addition to those from thomson scattering .
these additional opacities are included via the force multiplier , @xmath58 in equation [ acc ] , which is primarily a function of ionization parameter is defined as @xmath59 = @xmath60 at the ionized face of the clouds , where @xmath61 = 13.6 ev and @xmath62 is the hydrogen number density . ]
@xmath63 for a given spectral energy distribution ( * ? ? ?
* and references therein ) .
the acceleration due to gravity per mass is simply given by @xmath64 where @xmath65 is the total enclosed mass within @xmath10 parsecs , and @xmath66 is the universal gravitational constant .
putting equations [ acc ] and [ acc2 ] together we have @xmath67 now we will have to rewrite the acceleration in terms of velocity as a function of radius , and then solve for @xmath68 .
@xmath69 therefore we can now write equation [ acc3 ] as a simple separable differential equation @xmath70 substituting for the constants in equation [ vel2 ] and converting to appropriate units of km s@xmath3and pc , then integrating and setting the initial velocity to zero yields the following form for @xmath68 @xmath71dt},\ ] ] where @xmath72 is luminosity in units of @xmath73 ergs s@xmath3 and @xmath65 is in units of @xmath74 .
the constraints on the luminosity and the force multiplier are presented in the next section .
since ngc 1068 is a seyfert 2 galaxy , we can not measure its luminosity directly . from @xcite , the total luminosity of ngc 1068
is given by @xmath75 where @xmath76 is the fraction of nuclear flux observed as scattered radiation , @xmath77 is the distance to ngc 1068 , and @xmath78 is the solar luminosity .
the most uncertain term in equation [ lum ] is @xmath76 . @xcite
have summarized a range of values for @xmath76 that have been determined previously by several authors .
the range in @xmath76 spans a few orders of magnitude , from 0.0010.05 .
they claim that the best estimate comes from @xcite , who had determined a value for @xmath76 to be 0.015 , based on observations of [ oiii]and broad @xmath79 luminosity and their ratio .
@xcite concluded that @xmath76 is probably within a factor of a few of 0.01 , hence we have adopted a value for @xmath76 of 0.015 because it is the `` best '' estimate and close to the average value adopted by @xcite .
we already know the distance to ngc 1068 as 14.4 mpc @xcite so the bolometric luminosity of ngc 1068 is given by @xmath80 a value which could be uncertain by a factor of 0.315 , depending on @xmath76 .
the emission lines arising from the nlr gas are best fitted with a two component photoionization model at each position , based on _ hst_/stis long - slit spectra @xcite . according to @xcite
, the ionization state of the gas ranges from @xmath81@xmath82 for the two components but seems to vary little with distance .
using @xmath63 and the sed for ngc 1068 , we found the force multiplier to vary from @xmath83 5006000 for the front face ( ionized face ) of the clouds , based on cloudy models @xcite .
the mass function was derived in previous sections , and @xmath72 = 2.4 , so we can now numerically solve equation [ finvel ] for @xmath68 , assuming @xmath58 is constant with distance .
the results are presented in the next section .
equation [ finvel ] has only two parameters that we can vary to find the velocity @xmath84 as a function of distance @xmath10 : the launch radius @xmath85 , and the force multiplier @xmath58 .
we plotted @xmath68 for various combinations of launch radii and force multipliers of the gas .
the top panel of figure [ fm_high_low_ro_velocity_profiles ] with a fixed launch radius @xmath85 = 1 pc .
none of the curves show a significant decrease in velocity up to 10,000 pc .
+ middle : the velocity profiles for several lower force multipliers with @xmath85 = 1 pc .
curves with force multipliers less than 40 turn over before reaching 10,000 pc , although their maximum velocities do not fit the data of ngc 1068 .
+ bottom : the velocity profiles for several different launch radii ( top to bottom curves r@xmath86 = 1 , 2 , 4 , 8 , 16 , 32 pc ) all modeled with a constant force multiplier of 500 .
the bottom curve was able to turn over and return to systemic at @xmath0 1000 pc .
the maximum velocity and maximum distance reached for these curves do not fit the data.,scaledwidth=45.0% ] shows that with a launch radius of @xmath85 = 1 pc , the velocity of the gas increases with force multiplier , but quickly reaches a terminal velocity and does not slow down significantly , even out to @xmath4 10,000 pc .
this is not a surprise , as previous plots had shown that the mass is not enough to slow down the high - velocity clouds .
in addition , the maximum velocities for @xmath87 500 are too high compared to the observations , indicating that we must increase the launch radius @xmath85 .
the physical constraints on the nlr gas suggested that the force multiplier is larger than 500 .
however , we generated plots for @xmath58 smaller than 500 to see whether or not gravity can slow the clouds down , if the clouds are launched from 1 pc .
in the middle panel of figure [ fm_high_low_ro_velocity_profiles ] , we see that with force multipliers @xmath8 40 , gravity does slow the clouds down , but that the maximum velocities are too low to fit the data of ngc 1068 .
such models may prove useful for seyferts with lower outflow velocities in their nlrs , such as ngc 4151 and mkn 3 @xcite .
we next fix the force multiplier to the lowest value of 500 ( because higher values of @xmath58 will be even more problematic to slow down the clouds ) and vary the launch radius .
these plots are presented in the bottom panel of figure [ fm_high_low_ro_velocity_profiles ] .
the maximum velocity falls below the observed values at @xmath88 2 pc , and the final velocity does not return to zero until we reach large launch radii . as the launch radius increases to @xmath4 32 pc , the maximum attainable velocity of the outflowing clouds decreases , and
the enclosed mass is finally able to turn the velocity around .
however by that time , the maximum velocity is much lower than predicted by the kinematic model .
also , the launch radius and the maximum distance of outflow become too large to fit the data , since the nlr clouds generally launch from close in and drop in radial velocity within 400 pc ( paper ii ) .
clearly these curves can not fit the observed nlr velocities of ngc 1068 .
the radiation - gravity interaction on the nlr clouds of ngc 1068 fails to reproduce its velocity profile . for reasonable parameters ,
the velocity increases quickly close to the nucleus and mostly remains constant over large distances regardless of launch radius .
the maximum outflow velocity is rather sensitive to launch radius and decreases with increasing @xmath85 but the clouds velocities never turn over and decrease .
fine - tuning @xmath85 and @xmath58 outside of the range of reasonable parameters can lead to a deceleration profile , but the resulting velocity profile and amplitude does not match the observed trend . therefore we conclude that there must be additional forces at play in the nlr to account for the velocity profile that we see in the data .
one such force that could explain the trend in the data is drag , whereby the clouds are slowing down in a more diffuse , hotter , and higher ionization medium @xcite .
the drag force exerted on a cloud by an ambient medium is @xmath89 where @xmath90 is the mass density of the ambient medium , @xmath91 and @xmath92 are the velocities of the cloud and medium respectively , and @xmath93 represents the cross sectional area of a cloud @xcite .
following everett & murray , we assume the clouds to be spherical with mass @xmath94 = @xmath95 , where @xmath96 is the radius of a cloud , and @xmath97 is its mass density .
the acceleration on the clouds due to drag is then @xmath98 where @xmath99 is the hydrogen number density and @xmath57 is the mass of the proton . in our case ,
we assume that the velocity of the ambient medium is zero and the radius of the cloud remains constant . therefore with @xmath100 we can write @xmath101 where @xmath102 ( @xmath103 ) is the hydrogen column density of the cloud , and the minus sign represents deceleration .
combined with the radiative and gravitational acceleration of equation [ acc3 ] , the total acceleration on the clouds becomes @xmath104 and in differential form is @xmath105 substituting the various constants and converting to units of pc and km s@xmath3 , we have the following inseparable differential equation to solve @xmath106 where @xmath107 . both ngc 1068 and ngc 4151 show evidence for highly ionized gas extended throughout their nlrs , based on chandra x - ray observatory images @xcite
thus , we assume an ambient medium that is highly ionized , with ionization parameter @xmath108 10 .
the ionization parameter , which is inversely related to the density and radius , can be written as follows @xmath109 therefore if the ambient and the nlr clouds see the same ionizing luminosity ( l@xmath110 ) at a particular distance @xmath10 from the source , we can write @xmath111 @xcite provided good constraints on the parameters on the right side of equation [ ratio ] .
the densities of the nlr clouds are almost constant out to large distances from the nucleus with a typical value of @xmath112 10@xmath113 @xmath114for clouds with @xmath115 10@xmath116
. if we substitute these estimates in equation [ ratio ] , we will have a typical estimate for @xmath117 as follows @xmath118 @xcite found column densities for the nlr clouds in the range @xmath119 = 10@xmath12010@xmath121 @xmath122 , which corresponds to @xmath123 = 0.110 . since we are interested in the ratio @xmath124 , varying either parameter while keeping the other constant will result in the same curves . in this paper
, we choose to keep @xmath123 constant at the average value of 1 and vary @xmath117 , because @xmath117 is the most unknown quantity .
we already know that @xmath58 can take values from 5006000 , so equation [ diff ] can now be solved numerically for @xmath68 with various values of the launch radius ( @xmath85 ) , force multiplier ( @xmath125 ) , and the ratio @xmath124 .
we used mathematica v5.2 , which employs the most efficient choices among various flavors of runge - kutta algorithms , to solve for @xmath68 .
the results are presented in the next section .
the top panel of figure [ fm_500_r_1_fm_6000_r_20 ] shown by the numbers .
+ bottom : the velocity profiles for a force multiplier of 6000 , a launch radius of 20 pc , and varying medium densities similar to the top panel.,scaledwidth=45.0% ] represents several plots with a force multiplier of 500 , a launch radius of 1 pc , a column density of @xmath123 = 1 , and various densities of the ambient medium .
the figure shows that when launched at 1 pc , the gas accelerates to a maximum velocity inside 10 pc , then slows down again .
the velocity slows down faster with increasing drag forces , as measured by increasing densities of the ambient medium .
however the point of maximum velocity is too close in to match the data , which has a turnover velocity at @xmath4 140 pc .
the curve with medium density @xmath117 = 0.33 shows gradual deceleration out to @xmath4 500 pc , but its maximum velocity is slightly too high .
any other curve has either too high velocity at turnover or the velocity drops too quickly .
note that the velocity of the data we are trying to match has @xmath4 2000 km s@xmath3at a turnover of @xmath4 140 pc for ngc 1068 . with force multipliers higher than 500 , the same trend as in figure [ fm_500_r_1_fm_6000_r_20 ]
is seen in the velocity model , except that the velocities are much higher ( @xmath1 6000 km s@xmath3 ) .
again , the curve that seems best to match the deceleration part in the data is with medium density @xmath117 = 0.333 .
the rest of the curves decelerate too quickly or too slowly .
the major problem with all of these curves is that the velocity turnover point is much closer to the nucleus than 140 pc .
another way to decrease the overall velocity is to increase the launch radius , and tweak the force multiplier and ambient density .
we tried a launch radius of 10 pc with force multiplier ranging from 5006000 .
the maximum velocity drops as expected , reaching @xmath0 700 km s@xmath3for a force multiplier of 500 and climbs to @xmath0 2500 km s@xmath3when we increase the force multiplier to 6000 .
we noticed that the turning point increases by a factor of 5 as we increase the launch radius from 1 to 10 pc .
however the turnover point is still too low to fit the data well . in order to have a good fit ,
we first need to maximize the launch radius and then tweak the force multiplier and medium density to get the closest match to the data as possible .
the resolution in our data is @xmath4 10 pc and we should not launch much beyond this distance since we see clouds close to the smbh with near zero velocity .
therefore , in the bottom panel of figure [ fm_500_r_1_fm_6000_r_20 ] , we present a plot with a launch radius of @xmath85 = 20 pc , a force multiplier of 6000 , and a column density of @xmath123 = 1 .
the maximum velocity reached is @xmath4 1700 km s@xmath3with @xmath117 = 10 .
the best curve to represent the data seems to be the one with @xmath117 = 0.33 , whose maximum velocity is @xmath4 1500 km s@xmath3 , although the turnover point at @xmath4 60 pc is still too low according to our kinematic model .
furthermore , we have had to tweak the launch radius and force multiplier to very specific values , such that only a narrow range of the observed values gives a reasonably decent fit . however ,
to directly test this velocity profile , we generate biconical models similar to our previous kinematic models ( paper ii ) , to determine whether a more dynamical velocity law , rather than the simple linear law , can reasonably match the data .
we applied our kinematic models of paper ii with our dynamical velocity law to the data .
previously we had used a simple kinematic velocity law , which is based on the relation @xmath126 .
now instead we substitute the velocity relation based on equation [ diff ] , which represented a more physical situation . a model with the new velocity law of equation [ diff ]
is presented in figure [ r20fm6000nh3model ] , , and the best - fit velocity profile from the bottom panel of figure [ fm_500_r_1_fm_6000_r_20 ] with @xmath117 = 0.33 .
clearly this model is a poor match to the data.,scaledwidth=45.0% ] using the input parameters from table [ tabsevena ] , cccccr 450 & 10 & 40 & 5 & 57.8 & equation [ diff ] + and using the best - fit curve from the bottom panel of figure [ fm_500_r_1_fm_6000_r_20 ] , with @xmath123 = 1 and @xmath117 = 0.33 .
our biconical model was constructed with an inner and outer half opening angle of 10@xmath127 and 40@xmath127 respectively , and inclination of 5@xmath127 , and position angle of 57.8@xmath127 , and a half - size of 450 pc .
points on the bicone were assigned velocities according to equation [ diff ] .
the center slit ( slit 4 , see paper ii ) is used here for comparison .
we extracted the velocities from the model at a position equivalent to slit 4 and those are shown in the shaded regions in figure [ r20fm6000nh3model ] .
the data from slit 4 are shown in small triangles .
the model represents a poor fit to the data from slit 4 ( the center slit ) of ngc 1068 .
this was expected from looking at the previous velocity plots , as the turnover point was too low , the launch radius was a bit large , and the velocity profile did not resemble our kinematically derived linear profiles .
we have varied the input parameters from table [ tabsevena ] , but this makes very little improvement .
the turnover point , starting distance , and maximum velocity of the bicone can not be varied without changing the drag parameters , as these are implicitly defined in equation [ diff ] .
the thickness of the bicone can be increased to accommodate more data points , but the bicone will show lots of unnecessary shaded regions . changing the maximum extent of the bicone @xmath128 will have absolutely no effect on the shaded region , except interrupting the shaded regions before or continuing them beyond @xmath12 6 . that leaves us with only two parameters to vary , the inclination of the bicone axis , and its position angle in the sky , and varying these two alone did not fix the model .
with radiation pressure driving the nlr clouds , their velocities will accelerate very quickly , within a few parsecs of the nucleus , assuming the clouds are indeed launched close in . with the introduction of the drag forces , the overall velocities are lowered , but even so the velocities reach maximum too quickly .
the data suggest that the clouds are gradually increasing their velocities to a maximum at about 140 pc .
this gradient in the velocity can not be simply accounted for by radiative forces driving the clouds .
it seems therefore , that radiation pressure may not be the only driving mechanism for the nlr clouds , or that other forces are involved to steer the clouds to that particular gradient .
the velocity profile shown in the data resembles one with a linear or ` hubble flow ' law .
the inclusion of a drag force only serves to reduce the overall high velocities due to radiation driving , but the maximum velocities are reached very close to the nucleus , too close to match the data effectively .
gravitational forces alone can not stop the fast moving clouds observed in the nlr of ngc 1068 . with velocities as high as 1500
km s@xmath3at 140 pc , the clouds should have escaped the nlr . to compound the problem , when radiation forces are added , the cloud velocities are boosted even more . yet
we see clouds that are slowing down and gradually reaching systemic velocity .
therefore , the data suggest that there is a powerful force dragging on the clouds to slow their velocity .
the drag force that we introduce can have a significant effect on the clouds velocities .
we can conclude therefore that the drag forces are a strong competitor to the radiative forces , strong enough to bring the clouds to a halt even close to the nucleus , depending on the column densities of the outflowing clouds and the densities of the ambient medium .
however the overall velocity profiles generated with radiative , gravitational , and drag forces do not match the data for ngc 1068 .
assuming that the mass profile of ngc 4151 is similar to the one for ngc 1068 , it will prove difficult to match its observed velocity profile , because the same linear trends are seen in the velocity of the outflowing clouds .
the same can be said for mkn 3 @xcite .
we would like to thank douglas r. gies and paul j. wiita for providing insights into some of our modeling work .
we would also like to thank john .
e. everett for helpful discussions concerning the driving mechanisms of the nlr gas .
the data used in this paper were obtained from the multimission archive at the space telescope science institute ( mast ) .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
this research has made use of nasa s astrophysics data system . | we present dynamical models based on a study of high - resolution long - slit spectra of the narrow - line region ( nlr ) in ngc 1068 obtained with the space telescope imaging spectrograph ( stis ) aboard _ the hubble space telescope ( hst)_. the dynamical models consider the radiative force due to the active galactic nucleus ( agn ) , gravitational forces from the supermassive black hole ( smbh ) , nuclear stellar cluster , and galactic bulge , and a drag force due to the nlr clouds interacting with a hot ambient medium .
the derived velocity profile of the nlr gas is compared to that obtained from our previous kinematic models of the nlr using a simple biconical geometry for the outflowing nlr clouds .
the results show that the acceleration profile due to radiative line driving is too steep to fit the data and that gravitational forces along can not slow the clouds down , but with drag forces included , the clouds can slow down to the systemic velocity over the range 100400 pc , as observed
. however , we are not able to match the gradual acceleration of the nlr clouds from @xmath0 0 to @xmath0 100 pc , indicating the need for additional dynamical studies . |
undersampled images are common in astronomy , because instrument designers are frequently forced to choose between properly sampling a small field - of - view , or undersampling a larger field .
nowhere is this problem more acute than on the hubble space telescope , whose corrected optics now provide superb resolution ; however , the detectors on hst are only able to take full advantage of the full resolving power of the telescope over a limited field of view . for instance
, the primary optical imaging camera on the hst , the wide field and planetary camera 2 @xcite , is composed of four separate 800x800 pixel ccd cameras , one of which , the planetary camera ( pc ) has a scale of @xmath0 per pixel , while the other three , arranged in a chevron around the pc , have a scale of @xmath1 per pixel .
these latter three cameras , referred to as the wide field cameras ( wfs ) , are currently the primary workhorses for deep imaging surveys on hst .
however , these cameras greatly undersample the hst image .
the width of a wf pixel equals the full - width at half - maximum ( fwhm ) of the optics in the the near - infrared , and greatly exceeds it in the blue .
in contrast , a well - sampled detector would have @xmath2 pixels across the fwhm .
other hst cameras such as nicmos , stis and the future advanced camera for surveys ( acs ) also suffer from undersampling to varying degrees .
the effect of undersampling on wf images is illustrated by the
great eye chart in the sky " in figure 1 .
further examples showing astronomical targets are given in section 8 .
when the true distribution of light on the sky @xmath3 is observed by a telescope it is convolved by the point - spread function of the optics @xmath4 to produce an observed image , @xmath5 , where @xmath6 represents the convolution operator .
this effect is shown for the hst and wfpc2 optics by the upper - right panel in figure 1 .
pixelated detectors then again convolve this image with the response function of the electronic pixel @xmath7 , thus @xmath8 .
the detected image can be thought of as this continuous convolved image _ sampled _ at the center of each physical pixel .
thus a shift in the position of the detector ( know as a `` dither '' ) can be thought of as producing offset samples from the same convolved image .
although pixels are typically square on the detector , their response may be non - uniform , and indeed , may , because of the scattering of light and charge carriers , effectively extend beyond the physical pixel boundaries .
this is the case in wfpc2 .
by contrast , in the nicmos detectors @xcite , the electronic pixel is effectively smaller than the physical pixel .
fortunately , much of the information lost to undersampling can be restored . in the lower right of figure 1
we display an image made using one of the family of techniques we refer to as `` linear reconstruction . ''
the most commonly used of these techniques are shift - and - add and interlacing . in interlacing ,
the pixels from the independent images are placed in alternating pixels on the output image according to the alignment of the pixel centers in the original images .
the image in the lower right corner of figure 1 has been restored by interlacing dithered images . however , due to the occasional small positioning errors of the telescope and the non - uniform shifts in pixel space caused by the geometric distortion of the optics , true interlacing of images is often infeasible . in the other standard linear reconstruction technique ,
shift - and - add , a pixel is shifted to the appropriate location and then added onto a sub - sampled image
. shift - and - add can easily handle arbitrary dither postions , but it convolves the image yet again with the original pixel , compounding the blurring of the image and the correlation of the noise . in this case , two further convolutions are involved .
the image is convolved with the physical pixel @xmath9 , as this pixel is mathematically shifted over and added to the final image .
in addition , when many images with different pointings are added together on the final output grid , there is also a convolution by the pixel of the final output grid @xmath10 .
this produces a final image @xmath11 the last convolution rarely produces a significant effect , however , as the output grid is usually considerably finer than the detector pixel grid , and convolutions add roughly as a sum of squares ( the summation is exact if the convolving functions are gaussians ) .
the importance of avoiding convolutions by the detector pixel is emphasized by comparing the upper and lower right hand images in figure 1 .
the deterioration in image quality between these images is due entirely to the single convolution of the image by the wf pixel .
the interlaced image in the lower - right panel has had the sampled values from all of the input images directly placed in the appropriate output pixels without further convolution by either @xmath9 or @xmath10 .
here we present a new method , drizzle , which was originally developed for combining the dithered images of the hubble deep field north ( hdf - n ) @xcite and has since been widely used both for the combination of dithered images from hst s cameras and those on other telescopes .
drizzle has the versatility of shift - and - add yet largely maintains the resolution and independent noise statistics of interlacing .
while other methods ( _ c.f . _ lauer 1999 ) have been suggested for the linear combination of dithered images , drizzle has the advantage of being able to handle images with essentially arbitrary shifts , rotations and geometric distortion , and , when given input images with proper associated weight maps , creates an optimal statistically summed image .
drizzle also naturally handles images with `` missing '' data , due , for instance , to corruption by cosmic rays or detector defects .
the reader should note that drizzle does not attempt to improve upon the final image resolution by enhancing the high frequency components of the image which have been suppressed either by the optics or the detector .
while such procedures , which we refer to as `` image restoration '' ( in contrast to `` image reconstruction '' ) , are frequently very valuable ( see hanisch and white 1994 for a review ) , they invariably trade signal - to - noise for enhanced resolution .
drizzle , on the other hand , was developed specifically to provide a flexible and general method of image combination which produces high resolution results without sacrificing the final signal - to - noise .
although the effect of drizzle on the quality of the image can be profound , the algorithm is conceptually straightforward .
pixels in the original input images are mapped into pixels in the subsampled output image , taking into account shifts and rotations between images and the optical distortion of the camera . however , in order to avoid re - convolving the image with the large pixel footprint " of the camera , we allow the user to shrink the pixel before it is averaged into the output image , as shown in figure 2 . the new shrunken pixels , or drops " , rain down upon the subsampled output . in the case of the hdf - n wfpc2 imaging , the drops were given linear dimensions one - half that of the input pixel slightly larger than the dimensions of the output pixels .
the value of an input pixel is averaged into an output pixel with a weight proportional to the area of overlap between the drop " and the output pixel .
note that if the drop size is sufficiently small not all output pixels have data added to them from each input image .
one must therefore choose a drop size that is small enough to avoid degrading the image , but large enough so that after all images are drizzled the coverage is reasonably uniform .
the drop size is controlled by a user - adjustable parameter called pixfrac , which is simply the ratio of the linear size of the drop to the input pixel ( before any adjustment due to the geometric distortion of the camera ) .
thus interlacing is equivalent to drizzle in the limit of @xmath12 , while shift - and - add is equivalent to @xmath13 .
the degree of subsampling of the output is controlled by the user through the scale parameter @xmath14 , which is the ratio of the linear size of an output pixel to an input pixel . when a pixel @xmath15 from an input image @xmath16 with data value @xmath17 and user defined weight @xmath18 is added to an output image
pixel @xmath19 with value @xmath20 , weight @xmath21 , and fractional pixel overlap @xmath22 , the resulting values and weights of that same pixel , @xmath23 and @xmath24 are @xmath25 where a factor of @xmath26 is introduced to conserve surface intensity , and where @xmath16 and @xmath27 are used to distinguish the input and output pixel indices . in practice , drizzle applies this iterative procedure to the input data , pixel - by - pixel , image - by - image .
thus , after each input image is processed , there is a usable output image and weight , @xmath28 and @xmath29 .
the final output images , after all inputs have been processed , can be written as @xmath30 where for these equations , 4 and 5 , we use the einstein convention of summation over repeated indices , and where the input indices @xmath31 and @xmath32 extend over all input images .
it is worth noting that in nearly all cases @xmath33 , since very few input pixels overlap a given output pixel .
this algorithm has a number of advantages over the more standard linear reconstruction methods presently used .
it preserves both absolute surface and point source photometry ( though see section 5 for a more detailed discussion of point source photometry ) .
therefore flux can be measured using an aperture whose size is independent of position on the chip . and
as the method anticipates that a given output pixel may receive no information from a given input pixel , missing data ( due for instance to cosmic rays or detector defects ) does not cause a substantial problem , so long as there are enough dithered images to fill in the gaps caused by these zero - weight input pixels . finally , the linear weighting scheme is statistically optimum when inverse variance maps are used as the input weights .
drizzle replaces the convolution by @xmath9 in equation 1 with a convolution with @xmath34 , the pixfrac . as
this kernel is usually smaller the full pixel , and as noted earlier convolutions add as the sum of squares , the effect of this replacement is often quite significant .
furthermore , when the dithered positions of the input images map directly onto the centers of the output grid , and pixfrac and scale are chosen so that @xmath34 is only slightly greater than @xmath14 , one obtains the full advantages of interlacing : because the power in an output pixel is almost entirely determined by input pixels centered on that output pixel , the convolutions with both @xmath34 and @xmath10 effectively drop away .
nonetheless , the small overlap between adjacent drops fills in missing data .
few hst observing proposals have sufficient time to take a number of exposures at each of several dither positions . therefore , if dithering is to be of wide - spread use , one must be able to remove cosmic rays from data where few , if any , images are taken at the same position on the sky .
we have therefore adapted drizzle to the removal of cosmic rays .
as the techniques involved in cosmic ray removal are also valuable in characterizing the image fidelity of drizzle , we will discuss them first .
here then is short description of the method we use for the removal of cosmic rays : 1 . drizzle each image onto a separate sub - sampled output image using @xmath13 .
2 . take the median of the resulting aligned drizzled images .
this provides a first estimate of an image free of cosmic - rays .
3 . map the median image back to the input plane of each of the individual images , taking into account the image shifts and geometric distortion .
this can done by interpolating the values of the median image using a program we have named `` blot '' .
4 . take the spatial derivative of each of the blotted output images .
this derivative image is used in the next step to estimate the degree to which errors in the computed image shift or the blurring effect of taking the median could have distorted the value of the blotted estimate .
compare each original image with the corresponding blotted image . where the difference is larger than can be explained by noise statistics , the flattening effect of taking the median or an error in the shift ,
the suspect pixel is masked .
repeat the previous step on pixels adjacent to pixels already masked , using a more stringent comparison criterion .
finally , drizzle the input images onto a single output image using the pixel masks created in the previous steps . for this final combination a smaller pixfrac than in step 1 ) will usually be used in order to maximize the resolution of the final image .
figure 3 shows the result of applying this method to data originally taken by cowie and colleagues @xcite .
the reduction was done using a set of iraf @xcite scripts which are now available along with drizzle in the dither package of stsdas .
in addition to demonstrating how effectively cosmic rays can be removed from singly dithered images ( i.e. images which share no common pointing ) , this image also displays the degree to which linear reconstruction can improve the detail of an image . in the drizzled image
the object to the upper right clearly has a double nucleus ( or a single nucleus with a dust lane through it ) , but in the original image the object appears unresolved .
the drizzling algorithm was designed to obtain optimal signal - to - noise on faint objects while preserving image resolution .
these goals are unfortunately not fully compatible . as noted earlier , non - linear image restoration procedures which attempt to remove the blurring due to the psf and the pixel response function through enhancing the high frequencies in the image , such as the richardson - lucy @xcite and maximum entropy methods @xcite , directly exchange signal - to - noise for resolution . in the drizzling algorithm
no compromises on signal - to - noise have been made ; the weight of an input pixel in the final output image is entirely independent of its position on the chip .
therefore , if the dithered images do not uniformly sample the field , the `` center of light '' in an output pixel may be offset from the center of the pixel , and that offset may vary between adjacent pixels . dithering offsets combined with geometric distortion generally produce a sampling pattern that varies across the field .
the output psfs produced by the combination of such irregularly dithered datasets may , on occasion , show variations about the true psf .
fortunately this does not noticeably affect aperture photometry performed with typical aperture sizes .
in practice the variability appears larger in wfpc2 data than we would predict based on our simulations .
examination of more recent dithered stellar fields leads us to suspect that this excess variability results from a problem with the original data , possibly caused by charge transfer errors in the ccd @xcite .
camera optics generally introduce geometric distortion of images . in the case of the wfpc2 ,
pixels at the corner of each ccd subtend less area on the sky than those near the center .
this effect will be even more pronounced in the case of the advanced camera for surveys ( acs ) .
however , after application of the flat field , a source of uniform surface brightness on the sky produces uniform counts across the ccd .
therefore , point sources near the corners of the chip are artificially brightened compared to those in the center . by scaling the weights of the input pixels by their areal overlap with the output pixel , and by moving input points to
their corrected geometric positions , drizzle largely removes this effect . in the case of @xmath35 ,
this correction is exact . in order to study the ability of drizzle to remove the photometric effects of geometric distortion
when pixfrac is not identically equal to one , we created a four times sub - sampled grid of @xmath36 artificial stellar psfs .
this image was was then blotted onto four separate images , each with the original wf sampling , but dithered in a four - point pattern of half - pixel shifts . as a result of the geometric distortion of the wf camera ,
the stellar images in the corners of these images appear up to @xmath37 brighter in the corners of the images than near the center .
these images were then drizzled with a @xmath38 and @xmath39 .
aperture photometry on the @xmath36 grid after drizzling reveals that the effect of geometric distortion on the photometry has been dramatically reduced : the rms photometric variation in the drizzled image is 0.004 mags .
of course this is not the final photometric error of a drizzled image ( which will depend on the quality of the input images ) , but only the additional error which the use of drizzle would add under these rather optimal circumstances . in practice users
may not have four relatively well interlaced images but rather a number of almost random dithers , and each dithered image may suffer from cosmic ray hits . therefore , in a separate simulation , we have used the shifts actually obtained in the wf2 f814w images of the hdf - n as an example of the nearly random sub - pixel phase that large dithers may produce on hst .
in addition , we have associated with each image a mask corresponding to cosmic ray hits from one of the deep hst wf images used in creating figure 3 .
when these simulated images are drizzled together , the root mean - square noise in the final photometry ( which does not include any errors that could occur because of missed or incorrectly identified cosmic rays ) is @xmath40 mags .
figure 4 displays the results of this process .
we have also evaluated the effect of drizzling on astrometry .
the stellar images described in the previous section were again drizzled using the hdf shifts as above , setting @xmath38 and @xmath39 .
both uniform weight files and cosmic ray masks were used .
the positions of the drizzled stellar images were then determined with the imexam " task of iraf , which locates the centroid using the marginal statistics of a box about the star .
a box with side equal to 6 _ output _ pixels , or slightly larger than twice the full - width at half maximum of the stellar images , was used .
a root mean square scatter of the stellar positions of @xmath41 _ input _ pixels about the true position was found for the images created with uniform weight files and the cosmic - ray masks .
however , we find an identical scatter when we down - sample the original four - times oversampled images to the two - times oversampled scale of the test images .
thus it appears that _ no _ additional measurable astrometric error has been introduced by drizzle .
rather we are simply observing the limitations of our ability to centroid on images which contain power that is not fully nyquist sampled even when using pixels half the original size .
drizzle frequently divides the power from a given input pixel between several output pixels . as a result , the noise in adjacent pixels will be correlated . understanding this effect in a quantitative manner is essential for estimating the statistical errors when drizzled images are analysed using object detection and measurement programs such as sextractor @xcite and daophot @xcite .
the correlation of adjacent pixels implies that a measurement of the noise in a drizzled image on the output pixel scale underestimates the noise on larger scales .
in particular , if one block sums a drizzled image by @xmath42 pixels , even using a proper weighted sum of the pixels , the per - pixel noise in the block summed image will generally be more than a factor of n greater than the per - pixel noise of the original image .
the factor by which the ratio of these noise values differs from n in the limit as @xmath43 we refer to as the noise correlation ratio , @xmath44 .
one can easily see how this situation arises by examining figure 5 . in this figure
we show an input pixel ( broken up into two regions , a and b ) being drizzled onto an output pixel plane .
let the noise in this pixel be @xmath45 and let the area of overlap of the drizzled pixel with the `` primary '' output pixel ( shown with a heavier border ) be @xmath46 , and the areas of overlap with the other three pixels be @xmath47 and @xmath48 , where @xmath49 , and @xmath50 .
now the total noise power added to the image variance is , of course , @xmath51 ; however , the noise that one would measure by simply adding up the variance of the output image pixel - by - pixel would be @xmath52 the inequality exists because all cross terms ( @xmath53 ) are missed by summing the squares of the individual pixels . these terms , which represent the correlated noise in a drizzled image , can be significant .
in general , the correlation between pixels , and thus @xmath44 , depends on the choice of drizzle parameters and geometry and orientation of the dither pattern , and often varies across an image .
while it is always possible to estimate @xmath44 for a given set of drizzle parameters and dithers , in the case where all output pixels receive equivalent inputs ( in both dither pattern and noise , though not necessarily from the same input images ) the situation becomes far more analytically tractable . in this case , calculating the noise properties of a single pixel gives one the noise properties of the entire image .
consider then the situation when @xmath54 , @xmath34 , is set to zero .
there is then no correlated noise in the output image since a given input pixel contributes only to the output pixel which lies under its center , and the noise in the individual input pixels is assumed to be independent .
let @xmath55 represent a pixel from any of the input images , and let @xmath56 be the set of all @xmath55 whose centers fall on a given output pixel of interest .
then it is simple to show from equations 4 and 5 that the expected variance of the noise in that output pixel , when @xmath57 , is simply @xmath58 where @xmath59 is the standard deviation of the noise distribution of the input pixel @xmath55 .
we term this @xmath60 , as it is the standard deviation calculated with the pixel values added only to the pixels on which they are _
centered_. now let us consider a drizzled output image where @xmath61 . in this case
, the set of pixels contributing to an output pixel will not only include pixels whose centers fall on the output pixel , but also those for which a portion of the drop lands on the output pixel of interest even though the center does not . we refer to the set of all input pixels whose drops overlap with a given output pixel as @xmath62 and note that @xmath63 .
the variance of the noise in a given output pixel is then @xmath64 where @xmath65 is the fractional area overlap of the drop of input data pixel @xmath55 with the output pixel @xmath27 .
here we choose the symbol @xmath66 to represent the standard deviation calculated from all pixels that contribute to the output pixel when @xmath67 .
the degree to which @xmath68 and @xmath69 differ depends on the dither pattern and the values of @xmath34 and @xmath14 . however , as more input pixels are averaged together to estimate the value of a given output pixel in @xmath62 than in @xmath56 , @xmath70 .
when @xmath71 , @xmath66 is by definition equal to @xmath60 .
now consider the situation where we block average a region of @xmath72 pixels of the final drizzled image , doing a proper weighted sum of the image pixels .
this sum is equivalent to having drizzled onto an output image with a scale size @xmath73 .
but as @xmath74 , this approaches the sum over c , or , in the limit of large @xmath75 , @xmath76 . however , a prediction of the noise in this region , based solely on a measurement of the pixel - to - pixel noise , without taking into account the correlation between pixels would produce @xmath77 .
thus we see that @xmath78 one can therefore obtain @xmath44 for a given set of drizzle parameters and dither pattern by calculating @xmath60 and @xmath66 and performing the division .
however , there is a further simplification that can be made .
because we have assumed that the inputs to each pixel are statistically equivalent , it follows that the weights of the individual output pixels in the final drizzled image are independent of the choice of @xmath34 . to see this , notice that the total weight of a final image ( that is the sum of the weights of all of the pixels in the final image ) is independent of the choice of @xmath34 . ignoring edge pixels , the number of pixels in the final image with non - zero weight is also independent of the choice of @xmath34 . yet
as the fraction of pixels within @xmath34 of the edge scales as @xmath79 , and the weight of an interior pixel can not depend on n , we see that the weight of an interior pixel must also be independent of @xmath34 . as a result @xmath80 . therefore , we find that @xmath81 although @xmath44 must be calculated for any given set of dithers , there is perhaps one case that is particularly illustrative .
when one has many dithers , and these dithers are fairly uniformly placed across the pixel , one can approximate the effect of the dither pattern on the noise by assuming that the dither pattern is entirely uniform and continuously fills the output plane . in this case
the above sums become integrals over the output pixels , and thus it is not hard ( though somewhat tedious ) to derive @xmath44 .
if one defines @xmath82 where @xmath83 and @xmath84 , then in the case of a filled uniform dither pattern one finds , if @xmath85 @xmath86 and if @xmath87 , @xmath88 using the relatively typical values of @xmath89 and @xmath90 , one finds @xmath91 .
this formula can also be used when block summing the output image .
for example , a weighted block - sum of @xmath92 pixels is equivalent to drizzling into a single pixel of size @xmath93 .
therefore , the correlated noise in the blocked image can be estimated by replacing @xmath14 with @xmath93 in the above expressions .
drizzle has now been widely used for many astronomical image combination problems . in this section
we briefly note some of these and provide references where further information may be obtained .
drizzle was developed for use in the original hubble deep field north , a project to image an otherwise unexceptional region of the sky to depths far beyond those of previous astronomical images .
exposures were taken in four filter bands from the near ultraviolet to the near infra - red .
the resulting images are available in the published astronomical literature @xcite as well as from the space telescope science institute via the world wide web at http://www.stsci.edu/ftp/science/hdf/hdf.html .
subsequently drizzle has also been applied to the hubble deep field south @xcite . in this case
it was used for the combination of images from nicmos @xcite and stis @xcite as well as wfpc2 @xcite . in order to obtain dithered nicmos and wfpc2 images in parallel with stis spectroscopy ,
hst was rotated , as well as shifted , between observations during the hdf - s .
all the software developed to handle such challenging observations is now publicly available ( see section 9 ) .
the hdf imaging campaigns are atypical as they had a large numbers of dither positions .
a more usual circumstance , matching that described in section 3 , is the processing of a small number of dithers without multiple exposures at the same pointing .
a good example of such imaging and its subsequent processing is provided in fruchter _
et al . _
( 1999 ) where gamma ray burst host galaxies were observed using the stis and nicmos hst cameras to obtain morphological and photometric information .
similarly bally , odell and mccaughrean ( 2000 ) have used drizzle to combine dithered wfpc2 images with single exposures at each dither position in a program to observe disks , microjets and wind - blown bubbles in the orion nebula .
examination of these published images may help the reader to obtain a feeling for the results of using the drizzle program .
in addition an extensive set of worked examples of combining dithered data using drizzle is available in the dither handbook @xcite distributed by stsci .
drizzle provides a flexible , efficient means of combining dithered data which preserves photometric and astrometric accuracy , obtains optimal signal - to - noise , and approaches the best resolution that can be obtained through linear reconstruction .
an extensively tested and robust implementation is freely available as an iraf task as part of the stsdas package and can be retrieved from the space telescope science institute web page ( http://www.stsci.edu ) .
in addition to drizzle , a number of ancillary tasks for assisting with determining the shifts between images and the combination of wfpc2 data are available as part of the dither " package in stsdas .
we are continuing to improve drizzle , to increase both ease of use and generality .
new versions of drizzle will be incorporated into stsdas software updates .
additional capabilities will soon make the alignment of images simpler , and will provide the user with a choice of drizzling kernels , including ones designed to speed up the image combination with minimal change to the output image or weight an enhancement which may prove particularly useful in the processing acs images .
although these additions may make drizzle somewhat more flexible , the basic algorithm described here will remain largely unchanged , as it provides a powerful , general algorithm for the combination of dithered undersampled images .
drizzle was developed originally to combine the hdf - n datasets .
we wish to thank our colleagues in the hdf - n team , and bob williams in particular , for encouraging us , and for allowing us to be part of this singularly exciting scientific endeavor .
we also thank ivo busko for his work on the original implementation of the stsdas dither package , hans - martin adorf for many entertaining and thought provoking discussions on the theory of image combination , and stefano casertano for inciting us to develop a more general theory of the correlated noise in drizzled images .
finally , we are grateful to anton koekemoer for a careful reading of the text , and to our referee , tod lauer , for numerous suggestions which significantly improved the clarity and presentation of this paper . | we have developed a method for the linear reconstruction of an image from undersampled , dithered data .
the algorithm , known as variable - pixel linear reconstruction , or informally as `` drizzle '' , preserves photometry and resolution , can weight input images according to the statistical significance of each pixel , and removes the effects of geometric distortion both on image shape and photometry .
this paper presents the method and its implementation .
the photometric and astrometric accuracy and image fidelity of the algorithm as well as the noise characteristics of output images are discussed .
in addition , we describe the use of drizzling to combine dithered images in the presence of cosmic rays . |
in de sitter geometry the problem of particle production was considered by many authors .
these studies show that the space expansion generate particle production using various methods @xcite .
however a detailed study of interaction between scalar field and maxwell field minimally coupled with gravitation using perturbative methods and exact calculations received little attention .
we focus our attention on this subject because the last astronomical observations shows that the expansion of the universe is accelerating , thus increasing the interest on de sitter model .
the study of interactions between bosons and scalar fields on a curved background will help us to understand how the scalar electrodynamics from flat space will be modified when a gravitational field is present .
important results such as normalised solutions was obtained for dirac field @xcite,@xcite and scalar field @xcite,@xcite,@xcite on de sitter geometry . in the case of the electromagnetic field
the situation is simple if one recalls the conformal invariance between de sitter and minkowski metrics .
this means that the theory of free electromagnetic field is the same as in minkowski geometry if we use the metric with conformal time @xcite . in the present paper
our we will study on the interaction between charged scalar field and electromagnetic field .
we want to study the effect of electromagnetic interaction upon particle production in de sitter expanding universe .
the intention of this paper is to develop the theory of the scalar qed on de sitter spacetime and to use this formalism for studying the following processes : @xmath0 in an external coulomb field and @xmath1 . here
@xmath2 represents the charged scalar particle , @xmath3 represents the charged scalar antiparticle and @xmath4 the photon . the annihilation of scalar pair into vacuum in the presence of coulomb field and annihilation of the triplet scalar pair and a photon into de sitter vacuum represents the time reversed processes .
our results prove that these probabilities are nonvanishing only in strong gravitational fields .
our exact calculations allows us to study the behaviour of our amplitudes / probabilities for large / small values of the expansion factor , making in this way the connection with the scalar qed from minkowski space .
we must specify that in our calculations we will use the solutions of klein - gordon and maxwell equations which have a defined momentum .
these reactions are forbidden in flat space theory because of the energy - momentum conservation laws . in de sitter case
the energy is no longer conserved and we must expect to obtain a nonvanishing probability for these processes . we start in the second section with a review of the results concerning the free scalar field theory on de sitter space . in the third section
we construct the theory of interaction between charged scalar field and electromagnetic field on de sitter space and then our attention will be focused on developing the reduction formalism for scalar field . in this section
is also studied the perturbation theory which help us to obtain the definition of scattering amplitudes .
the first order processes that generate particle production will be studied in sections four and five .
further we study the probabilities that generate particle production using a graphical method .
our conclusions are summarized in section six , and in the appendix we give the main steps that help us to calculate the transition amplitudes .
the exact solutions of the free klein - gordon equation on de sitter space - time , was obtained in @xcite,@xcite .
the calculations was done using the de sitter line element @xcite : @xmath5 where @xmath6 is the expansion factor . in the chart @xmath7 with cartesian
coordinates the klein - gordon equation reads @xcite : @xmath8 it is known that the klein - gordon equation ( [ eq : kg1 ] ) on de sitter space can be analytically solved in terms of bessel functions @xcite,@xcite . in our study we will use the normalized solutions of positive frequencies obtained in @xcite : @xmath9 where @xmath10 is the hankel function of first kind , @xmath11 and the notations are : @xmath12 with @xmath13 .
the fundamental solutions of negative frequencies are @xmath14 . for @xmath15 ,
the modes defined above behave as a positive / negative frequency modes with respect to the conformal time @xmath16 : @xmath17 this is the behavior in the infinite past for positive / negative frequency modes which defines the bunch - davies vacuum .
the fundamental solutions of klein - gordon equation , satisfy the ortonormalizations relations @xcite : @xmath18 where the integration extends on an arbitrary hypersurface @xmath19 and @xmath20 .
the completeness condition satisfied by these solutions reads @xcite : @xmath21 the quantization can be done considering the plane wave in momentum representation : @xmath22,\label{eq : field1}\ ] ] where the particle @xmath23 and antiparticle ( @xmath24 operators satisfy the standard commutation relations in momentum representation : @xmath25=[b(\vec{p}\ , ) , b^{\dagger}(\vec{p}^{\,\,\prime } ) ] = \delta^3 ( { \vec{p}}-\vec{p}^{\,\,\prime})\,.\label{eq : com1}\ ] ] these operators can be calculated using the inversion formulas : @xmath26 all these operators act on the fock space supposed to have an unique vacuum state @xmath27 accomplishing @xmath28 in this way the field @xmath29 is correctly quantized according to the canonical rule @xcite : @xmath30=ie^{-3\omega t}\delta^3(\vec{x}-\vec{x}^{\,\,\prime})\,.\ ] ] in the quantum theory of fields it is important to study the green functions related to the partial commutator functions ( of positive or negative frequencies ) defined as : @xmath31,\ ] ] and the total one , @xmath32 .
these functions are solutions of the klein - gordon equation in both sets of variables and obey @xmath33^*=d^{(\mp)}(x , x')$ ]
. then the partial commutator functions will have the expressions : @xmath34 these functions help us to introduce the principal green functions .
in general , @xmath35 is a green function of the klein - gordon equation if satisfies @xcite : @xmath36 the properties of the commutator functions allow us to construct the green function just as in the scalar theory on minkowski spacetime .
we assume that the retarded , @xmath37 , and advanced , @xmath38 , green functions read : @xmath39 while the feynman propagator , @xmath40\,|0\rangle\nonumber\\ & = & \theta ( t - t ' ) d^{(+)}(t , t',\vec{x}-\vec{x}^{\,\,\prime})-\theta(t'-t)d^{(-)}(t , t',\vec{x}-\vec{x}^{\,\,\prime})\,,\end{aligned}\ ] ] is defined as a causal green function .
thus we have now all the elements that we need for begin our study related to the reduction formulas for scalar field on de sitter space - time .
the goal of the quantum field theory is to describe the dynamics of interacting particles .
we are interested to obtain the amplitudes for single particles to propagate in a curved space - time and the transition amplitudes for interacting particles between different initial and final states .
our main attention in this paper will be focused on interaction between charged massive scalar field and electromagnetic field on de sitter space - time . in this work we use for our calculation the same formalism as in @xcite-@xcite , and we adopt the minimal coupling corresponding to the classical interaction of a point charge as the prescription for introducing electrodynamic couplings . also we will note the interacting fields with @xmath41 and @xmath42 . these fields will satisfy a system of equations that can be obtained from the following lagrangian density : @xmath43&=&\int d^{4}x[\mathcal{l}_{sc}(\phi)+\mathcal{l}_{m}(a)+\mathcal{l}_{i}(\phi , a)]\nonumber\\ & & = \sqrt{-g}\biggl\{(\partial^{\mu}\phi^{\dagger}\partial_{\mu}\phi - m^2 \phi^{\dagger } \phi)-\frac{1}{4}f_{\mu\nu}f^{\mu\nu}\nonumber\\ & & -ie[(\partial_{\mu}\phi^{\dagger})\phi-(\partial_{\mu}\phi)\phi^{\dagger}]a^{\mu } + e^{2}\phi^{\dagger}\phi a_{\mu}a^{\mu}\biggl\},\label{eq : action}\end{aligned}\ ] ] where the density of lagrangian of interaction reads @xmath44a^{\mu } + \sqrt{-g}e^{2}\phi^{\dagger}\phi a_{\mu}a^{\mu}$ ] .
this relation is obtained after we make the substitution @xmath45 in the lagrangian of free scalar field . from action ( [ eq : action ] ) , we obtain the system of equations for interacting fields on a curved background as follows : @xmath46 + 2e^{2}\phi^{\dagger } \phi a^{\nu},\nonumber\\ & & \frac{1}{\sqrt{-g}}\,\partial_{\mu}\left[\sqrt{-g}\ , g^{\mu\nu}\partial_{\nu}\phi - ie\sqrt{-g}\phi a^{\mu}\right]+m^2\phi = ie(\partial_{\mu}\phi)a^{\mu}+e^{2}\phi a_{\mu}a^{\mu}. \label{eq : eqs}\end{aligned}\ ] ] the problem of finding exact solutions for the system ( [ eq : eqs ] ) , which contain coupled nonlinear equations for interacting fields proves to be as in the flat space case , a complicated task .
it is immediate from here that for finding solutions of the coupled equations , we can try to follow the same approximation method as in flat space case @xcite-@xcite .
the system of coupled equations is replaced with one system of integral equations that contain information about initial conditions .
for that we select the green function @xmath47 corresponding to one initial condition , which help us to write the solution of the second equation from system ( [ eq : eqs ] ) as follows : @xmath48\nonumber\\ & & + i[\partial_{\mu}\phi(y)]a^{\mu}(y)+e\phi(y ) a_{\mu}(y)a^{\mu}(y)\biggl\},\label{eq : int}\end{aligned}\ ] ] where @xmath49 are free fields . one can verify that ( [ eq : int ] ) is the exact solution of the second equation from ( [ eq : eqs ] ) , applying @xmath50 to @xmath41 , with the observation that @xmath51\hat{\phi}(x)=0 $ ] .
we introduce the following convenable notation @xmath52 . using the fact that green functions must satisfy equation ( [ eq : kgg ] ) , it is not hard to verify that the solution ( [ eq : int ] ) is justified .
since our main scope in this paper is to study transition amplitudes , we want to construct the states which give a simple description of the physical system at the time @xmath53 .
at this time the particles have not yet interacted with each other and propagate under the influence of their self - interactions only .
this implies that for constructing a theory of interactions , we must seek for operators that create independent particle states with each particle propagating with its physical mass .
let us start to construct these operators , using ( [ eq : int ] ) which offers us the possibility of constructing free fields , which are asymptotic equal ( at @xmath54 ) with solutions of system ( [ eq : eqs ] ) . as we know the retarded green functions @xmath55 vanishes at @xmath56 while the advanced one @xmath57 vanishes for @xmath58 . thus equation ( [ eq : int ] )
can be expressed with retarded and advanced functions as follows : @xmath59\nonumber\\ & & + i[\partial_{\mu}\phi(y)]a^{\mu}(y)+e\phi(y ) a_{\mu}(y)a^{\mu}(y)\biggl\}\label{eq : phi}.\end{aligned}\ ] ] now we can define the free fields , @xmath60 that satisfy : @xmath61 the free fields @xmath62 and @xmath63 have mass @xmath64 and are equal at @xmath65 with exact solutions of the coupled equations and represent the fields before and after the interaction .
it is known that the mass of the interacting fields may differ from the mass of free fields . for obtaining the mass correction we will ad to the lagrangian of interaction a term of the form @xmath66 , which help us to rewrite the second equation from system ( [ eq : eqs ] ) : @xmath67+m^2\phi = ie(\partial_{\mu}\phi)a^{\mu}+e^{2}\phi a_{\mu}a^{\mu}-\delta m^{2}\phi,\label{eq : dm}\ ] ] where the current @xmath68 is now the source giving rise to the scalar fields .
further we can follow the method from minkowski case @xcite-@xcite , where the free fields @xmath62 and @xmath63 are defined up to a normalization constant noted with @xmath69 .
this allows us to define the @xmath70 fields in order to construct the states of single particles , as follows : @xmath71\nonumber\\ & & + i[\partial_{\mu}\phi(y)]a^{\mu}(y)+e\phi(y ) a_{\mu}(y)a^{\mu}(y)-\delta m^{2}\phi(y)\biggl\}. \label{eq : inout1}\end{aligned}\ ] ] the @xmath70 free fields defined above satisfy klein - gordon equation , are written with the help of creation and annihilation operators and in addition satisfy the conditions : @xmath72 it is useful to express these fields with the help of equation ( [ eq : dm ] ) , in the form : @xmath73\phi(y).\label{eq : inout2}\ ] ] from the above normalization we can define the creation and annihilations operators as in section ii , @xmath74 the creation and annihilation operators defined above satisfy the commutation relations ( [ eq : com1 ] ) and from that it follows that all the properties of free fields will be preserved . before starting our calculations
we make a few remarks about the scattering operator . denoting the vacuum state @xmath27 ( which is suppose to be unique ) ,
the one particle / antiparticle states can be written : @xmath75 if we consider two states @xmath76 and @xmath77 , then the probability of transition from state @xmath78 to state @xmath79 is defined as the scalar product of the two states : @xmath80 .
these are just the elements of matrix for scattering operator @xmath81 , which are of physical interest .
this operator assures the stability of the vacuum state and one particle state , and in addition transform any @xmath82 field in the equivalent @xmath83 field .
our remaining task is to construct and study , general matrix elements which describe the dynamical behavior of interacting particles . as in minkowski
theory @xcite-@xcite , we can construct @xmath84particle @xmath83 and @xmath82 states as in the free scalar field theory by repeated application to the vacuum of @xmath85 and @xmath86 .
the reduction formulas for scalar particles can be obtained as in minkowski theory @xcite-@xcite .
let us consider the amplitude for the process in which particles from @xmath83 states denoted with @xmath78 , together with a scalar particle @xmath87 pass in @xmath82 state , in which we have a scalar particle @xmath88 and particles denoted by @xmath79 .
the reduction from out state can be done using the following relations : @xmath89 @xmath90\phi(y),\label{eq : dif}\ ] ] where we noted @xmath91 .
the final result for reduction of scalar particle from @xmath82 state is : @xmath92\nonumber\\ & & \times\langle out \beta|\phi(y)|in \alpha,1(\vec{p}\,)\rangle d^{4}y.\label{eq : fout}\end{aligned}\ ] ] this method can be also applied for the reduction of particles from @xmath83 state and the final result read : @xmath93f_{\vec{p}}\,(y)d^{4}y.\end{aligned}\ ] ] let us resume the above results related to reduction method .
we showed that all particles , will be replaced by formulas that were obtained in the reduction of one particle .
when more particles are reduced , in the matrix element appears the time ordered products of corresponding field operators . every particle will be replaced after we apply the above reduction method with expressions which depend on the field operator @xmath41 .
as in the flat space case , we can associate a green function to any process of interaction that must be calculated using perturbation theory . expanding the transition amplitudes and matrix elements in a power series in the interaction strength ,
we can obtain the rules for computing the terms in the expansion .
the green functions of the interacting fields can not be calculated exact and for that reason we must use perturbation methods .
the form of the amplitudes obtained from reduction formalism allows us to use perturbation calculus as in minkowski theory @xcite-@xcite . for calculating the green functions
it will be helpful to write them in terms of free fields as follows : @xmath94|0\rangle= \frac{\langle0|t[\hat{\phi}(y_{1})\hat{\phi}(y_{2}) ... \hat{a}(y_{n}),\mathbf{s}]|0\rangle}{\langle0|\mathbf{s}|0\rangle}. \label{eq : g}\ ] ] then the entire perturbation calculus is based on expansion of the operator @xmath95 : @xmath96d^{4}x_{1} ... d^{4}x_{n } , \label{eq : s}\end{aligned}\ ] ] where the density lagrangian of interaction is ( we neglect the second term in the density of lagrangian of interaction ) : @xmath97a^{\mu}:$ ] . each term
from ( [ eq : s ] ) corresponds to an order from perturbation theory . replacing ( [ eq : s ] ) in the expression of green function ( [ eq : g ] ) we obtain perturbation series which allows us to calculate the transition amplitude in any order . with this method
, the definition of transition amplitudes in any order of perturbation theory can be obtained .
the green functions from equation ( [ eq : g ] ) can be expressed in terms of feynmann propagators considering all the t contractions from wick theorem .
for example in the first order of perturbation theory the transition amplitude reads : @xmath98a^{\mu}(x)d^{4}x , \label{eq : ampl2}\ ] ] where the bilateral derivative acts as : @xmath99 .
the above expression is just the scattering amplitude in the first order of the perturbation theory on de sitter expanding universe . with the above method
were established the general rules of calculation for deducing the expression for the scattering amplitudes in any order of the perturbation theory .
the results obtained in the last two sections will be applied to calculate the processes which generate particle production .
in this section we want to study the process @xmath100 , in the presence of coulomb field , using the scalar qed formalism developed in the previous sections .
more precisely , we want to show that there are nonvanishing probabilities for scalar pair production in the case of expanding de sitter universe .
it is well known that the coulomb field could not generate particle production in scalar qed from minkowski space as a perturbative phenomenon .
contrary to this , in the de sitter case we must expect that this process to be allowed by the fact that energy is no longer conserved .
the studies done before with fermions indeed show that there are nonvanishing probabilities for pair generation in external coulomb field @xcite .
the ingredients of our calculations will be the solutions of klein - gordon equation in the momentum basis and the form of coulomb field in de sitter geometry . using the conformal invariance we can establish the expression for coulomb field in this geometry in the natural frame @xcite : @xmath101
the amplitude corresponding to scalar pair production in coulomb field on de sitter spacetime , reads : @xmath102a^{0}(x)d^{4}x , \label{eq : ampl2}\ ] ] where the notations @xmath103 stands for a scalar particle of momentum @xmath104 and negative charge and @xmath105 denotes the antiparticle with positive charge and momentum @xmath106 . in order to obtain the expression of the probability of pair production ,
we first need to calculate the amplitude of transition ( [ eq : ampl2 ] ) .
we use in our calculations the solutions of klein - gordon equation in momentum basis ( [ eq : sol ] ) .
these solutions are written in terms of hankel functions with imaginary index , provided that @xmath107 .
the spatial integral is the same as in the flat case and the result is incorporated in the amplitude formula .
the temporal integrals can be expressed with the solutions of the free scalar field ( [ eq : sol ] ) .
after we pass to the new variable of integration @xmath108 and calculate the bilateral derivative , the temporal integrals become : @xmath109.\ ] ] using the relation for derivative of hankel functions ( [ hn ] ) , and the relation between hankel functions and bessel @xmath110 functions ( [ hk ] ) from appendix , we finally obtain the expression of transition amplitude in the form : @xmath111 \nonumber\\ & & + p\int_{0}^{\infty}ds s^2 \left[k_{-ik}(ip\,'s)k_{-ik-1}(ips)+k_{-ik}(ip\,'s)k_{-ik+1}(ips)\right]\biggl\}.\end{aligned}\ ] ] the results of these integrals are discussed in appendix ( [ a0 ] ) . the final result for the amplitude
is expressed in terms of hypergeometric gauss functions , gamma euler functions and unit step functions as follows : @xmath112,\label{an}\end{aligned}\ ] ] where the newly introduced functions @xmath113 have the following expressions ( with the specification that @xmath114 is obtained when @xmath115 ) : @xmath116\nonumber\\ & & = \left(\frac{m}{\omega}\right)^{-1}\left[\left(\frac{p}{p\,'}\right)^{-ik}\gamma(3)\gamma(1-ik)\gamma(ik)\,_{2}f_{1}\left(2,1-ik;1-ik;\left(\frac{p}{p\,'}\right)^{2}\right)\right.\nonumber\\ & & \left.+\left(\frac{p}{p\,'}\right)^{ik}\gamma(3)\gamma(2+ik)\gamma(-ik)_{2}f_{1}\left(1,2+ik;1+ik;\left(\frac{p}{p\,'}\right)^{2}\right)\right.\nonumber\\ & & \left . + \left(\frac{p}{p\,'}\right)^{-ik}\gamma(3)\gamma(2-ik)\gamma(ik)_{2}f_{1}\left(1,2-ik;1-ik;\left(\frac{p}{p\,'}\right)^{2}\right ) \right.\nonumber\\ & & \left .
+ \left(\frac{p}{p\,'}\right)^{ik}\gamma(3)\gamma(1+ik)\gamma(-ik)\,_{2}f_{1}\left(2,1+ik;1+ik;\left(\frac{p}{p\,'}\right)^{2}\right)\right]\end{aligned}\ ] ] @xmath117\nonumber\\ & & = \left(\frac{m}{\omega}\right)^{-1}\left[\left(\frac{p}{p\,'}\right)^{-ik}\gamma(3)\gamma(1-ik)\gamma(1+ik)\,_{2}f_{1}\left(1,1-ik ;- ik;\left(\frac{p}{p\,'}\right)^{2}\right)\right.\nonumber\\ & & \left.+\left(\frac{p}{p\,'}\right)^{2+ik } \gamma(3)\gamma(2+ik)\gamma(-1-ik)\,_{2}f_{1}\left(2,2+ik;2+ik;\left(\frac{p}{p\,'}\right)^{2}\right)\right.\nonumber\\ & & \left.+\left(\frac{p}{p\,'}\right)^{-ik+2}\gamma(3)\gamma(2-ik)\gamma(-1+ik)\,_{2}f_{1}\left(2,2-ik;2-ik;\left(\frac{p}{p\,'}\right)^{2}\right)\right.\nonumber\\ & & \left.+\left(\frac{p}{p\,'}\right)^{ik}\gamma(3)\gamma(1-ik)\gamma(1+ik)\,_{2}f_{1}\left(1,1+ik;ik;\left(\frac{p}{p\,'}\right)^{2}\right)\right]\end{aligned}\ ] ] the result of the integrals in the case @xmath118 , give terms proportional with delta dirac function @xmath119 .
however when we sum all the contributions of each integral with hankel functions these terms cancel between them . from equations ( [ eq : gn ] ) and ( [ eq : hn ] ) , we can observe that the momentum is not conserved in the process of scalar pair production in external coulomb field .
the breaking of momentum conservation law is due to the external coulomb field , because the geometry ( [ metr ] ) has spatial translation invariance as an exact symmetry .
if one had just the de sitter background , without the coulomb field , there would still be certain types of particle production as we will show in the next section , but no violation of momentum conservation law @xcite .
it is clear that the contribution specific to the de sitter geometry in our amplitude is contained in the functions @xmath120 .
these quantities depend of the parameter @xmath121 , and contain the effect of space expansion on the pair production process .
these functions will define the probability of pair production , which is the outcome of our perturbative approach .
the probability for pair production in coulomb field is obtained by squaring the amplitude : @xmath122\\ \nonumber & & + \frac{1}{p\,'^4}\theta(p\,'-p)\left[\left|h_{k}\left(\frac{p}{p\,'}\right)\right|^{2}+\left|g_{k}\left(\frac{p}{p\,'}\right)\right|^{2 } -g_{k}\left(\frac{p}{p\,'}\right)h_{k}^{*}\left(\frac{p}{p\,'}\right)-h_{k}\left(\frac{p}{p\,'}\right ) g_{k}^{*}\left(\frac{p}{p^{'}}\right)\right]\biggl\}.\label{prob}\end{aligned}\ ] ] the total probability is obtained by integrating after the final momenta @xmath123 the probability given above .
this is a complicated task and we restrict to study the properties of the probability given in ( [ prob ] ) .
since our calculations are exact we can study the results for probability in terms of the ratio mass of the particle / expansion factor . in what follows we plot the probability as function of the parameter @xmath121 for different values of the momenta ratio @xmath124 .
these results are presented in the figs.([f1]-[f3 ] ) . as a function of @xmath125 for @xmath126 . ] as a function of @xmath125 for @xmath127 . ] as a function of @xmath125 for @xmath128 . ]
our results from the above graphs shows the dependence of the probability in terms of parameter @xmath121 , for different values of the ratio @xmath129 .
this result was obtained with the exact solution of klein - gordon equation expressed with hankel function of imaginary index @xmath130 , where @xmath131 , solutions that are valid for @xmath107 . in the above graphs the parameter @xmath125 was taken in the interval @xmath132 .
only for these values our analytical and graphical results are valid .
we observe that the probabilities have a variation in terms of the ratio @xmath129 . for the momenta ratio closer to one
the probability is very small as shown in fig.([f1 ] ) . as the momenta ratio become very small the probability increase
this result show that the transitions from vacuum in which the momenta ratio are very small will be favoured in external coulomb field ( see fig.([f3 ] ) ) . this is equivalent to say that there are bigger probabilities to produce pairs in which the momenta of the particle @xmath133 , is very small comparatively with the momenta of the antiparticle @xmath134 @xmath135 .
another observation that emerges from graphs figs.([f1]-[f3 ] ) is that the probability decreases as the ratio @xmath136 increase .
in other words , as the expansion parameter become smaller the probabilities for pair production reduces to zero .
when the parameter @xmath137 , which corresponds to the flat space limit ( zero expansion factor ) , the probabilities are zero as we can observe from figs.([f1]-[f3 ] ) .
this result is also obtained when we replace the ratio @xmath138 in the expressions for amplitudes and probabilities .
we can draw the conclusion that in the minkowski limit there are no production of scalar particles in external coulomb field .
further we can address the problem of separation between particles and antiparticles .
the probability is proportional with the factor @xmath139 .
the modulus from nominator can be expanded as : latexmath:[\[\label{sm } @xmath79 is the angle between the two momenta vectors .
from here we can see that for @xmath141 the momenta of particle and antiparticle are parallel and have the same orientation . in this case
the probability is minim and the pair will annihilate in vacuum .
the opposite case is when @xmath142 , which corresponds to the situation when the momenta of the particle and antiparticle are parallel and have opposite orientation .
this is the case when the probability is maxim . only in this case
the particles and antiparticles could separate .
the process : @xmath143 , in which scalar pair and a photon is generated from de sitter vacuum is the topic of this section .
this process is forbidden in scalar qed from flat space by the momentum and energy conservation laws . in de sitter case
the energy is no longer conserved and the probability for production of the triplet is no longer zero . for computing the transition amplitude we will use the solution of the klein - gordon equation and the solution of maxwell equation on de sitter background , written in the momentum basis .
the theory of maxwell field in de sitter geometry is simple is we recall the conformal invariance @xmath144 ( where @xmath145 is the vector potential in flat space ) .
then the fundamental solution of maxwell equation in de sitter geometry in momentum basis is @xcite : @xmath146 the transition amplitude for this process reads : @xmath147w^i(x)^*d^{4}x.\ ] ] our calculations will be valid only in the case @xmath148 and we will use the solutions of the klein - gordon equation expressed in terms of hankel functions of imaginary index . replacing all the quantities of interest in our amplitude we obtain that the spatial integrals give the delta dirac function @xmath149 , which assures the momentum conservation in this process . in the case of the temporal integrals we pass from hankel functions to bessel @xmath150 functions , arriving in this way at integrals ( [ in1]-[in2 ] ) , which are discussed in appendix .
the transition amplitude can be written passing to the new variable of integration @xmath108 : @xmath151 . \nonumber\\\end{aligned}\ ] ] the final result is expressed in terms of gauss hypergeometric functions @xmath152 and appell hypergeometric functions of double argument @xmath153 : @xmath154.\end{aligned}\ ] ] the functions which define our amplitude are defined as follows : @xmath155 @xmath156\label{h}\end{aligned}\ ] ] in equation ( [ fv],[hv ] ) , the newly introduced notation @xmath157 is defined as : @xmath158 we observe that the momentum is conserved in the process @xmath143 . in the previous section we obtained that the process of pair production in external field broke the momentum conservation law . from here
we can draw the conclusion that the particle production processes in the presence of gravity , preserve the momentum conservation law as long as there are no external fields ( electric fields or magnetic fields ) .
the probability is obtained squaring the amplitude and summing after the photon polarizations @xmath159 .
since our amplitude is proportional with a delta dirac term @xmath160 , by squaring this term we obtain @xmath161 . in this way
we will define the probability per unit of volume : @xmath162^{2}\,\delta^{3}(\vec{p}+\vec{p\,'}+\vec{q}\,)}{64 \,(2\pi)^{3 } } \left[|h_{k}(p , p\,',q)|^{2}+ |h_{-k}(p , p\,',q)|^{2 } \right.\nonumber\\ & & \left.+|f_{k}(p , p\,',q)|^{2}+ |f_{-k}(p , p\,',q)|^{2 } + h_{k}(p , p\,',q)h_{-k}^{*}(p , p\,',q)+h_{k}(p , p\,',q)f_{k}^{*}(p , p\,',q)\right.\nonumber\\ & & \left.+h_{k}(p , p\,',q)f_{-k}^{*}(p , p\,',q ) + h_{-k}(p , p\,',q)h_{k}^{*}(p , p\,',q ) + h_{-k}(p , p\,',q)f_{k}^{*}(p , p\,',q)\right.\nonumber\\ & & \left.+h_{-k}(p , p\,',q)f_{-k}^{*}(p , p\,',q ) + f_{k}(p , p\,',q)h_{k}^{*}(p , p\,',q)+f_{k}(p , p\,',q)h_{-k}^{*}(p , p\,',q)\right.\nonumber\\ & & \left.+f_{k}(p , p\,',q)f_{-k}^{*}(p , p\,',q ) + f_{-k}(p , p\,',q)h_{k}^{*}(p , p\,',q)+f_{-k}(p , p\,',q)h_{-k}^{*}(p , p\,',q)\right.\nonumber\\ & & \left.+f_{-k}(p , p\,',q)f_{k}^{*}(p , p\,',q)\right].\end{aligned}\ ] ] the probability depend on the factor : @xmath163 . using that the polarization vectors are orthogonal on photon momentum @xmath164 and for any polarization @xmath165
, they satisfy : @xmath166 we can express the sum from probability taking into account that the momentum conservation as given by the delta dirac term gives @xmath167 : @xmath168 where @xmath169 is the angle between the momenta vectors @xmath106 and @xmath170 .
this result shows that the probability will increase when the angle between the two vectors approach the value @xmath171 and vanishes for @xmath172 .
the conclusion is that the scalar particle with momentum @xmath173 and the photon will be emitted on different directions .
the same conclusion emerges if we use the angle between @xmath104 and @xmath170 .
the minkowski limit can be obtained when the expansion factor vanishes or equivalently when @xmath174 . in this limit the probability for generation of the triplet vanishes , thus obtaining the well known result from flat space where this process is forbidden by the energy conservation law . in this subsection the probability given in eq.([prv ] ) will be analysed in terms of the parameter @xmath121 , using numerical values for the momenta @xmath175 . as a function of @xmath125 for @xmath176 . ] as a function of @xmath125 for @xmath177 . ] for plotting the probability we take the appel hypergeometric functions @xmath153 to be equal with one , since these functions are less studied . using the definition of functions @xmath153
, one can observe that for large @xmath125 these functions are very convergent , and from here our approximation is well justified .
also one can observe from ( [ fv ] ) , that the shape of our graphs will be mainly determined by the factor @xmath178 , which is very convergent at large @xmath125 .
we observe from our graphs that the probability drops rapidly to zero when the expansion factor @xmath6 decreases . from graphs
figs.([f4]-[f5 ] ) , it is clear that in the minkowski limit the probability vanishes confirming in this way our analytical results for @xmath137 .
we studied in this paper the interaction between maxwell field and charged scalar field in de sitter geometry using the perturbative methods as in flat space case .
we established the equations of the interacting fields in the hypothesis of minimally coupling .
using then the solutions of the equations for interacting fields we construct the reduction formalism for the scalar field in de sitter geometry .
the definition of transition amplitudes was also obtained using the same perturbative methods as in flat space case .
we compare two kinds of particle production in de sitter geometry .
our study show that the particle production processes in external fields , do not preserve the momentum conservation law .
contrary to this we obtain that the mechanism of production of particles from de sitter vacuum always preserve the momentum conservation law .
our analytical and graphical results prove that the probability of particle production is significant only in strong gravitational fields and vanishes in the flat space limit @xmath179 .
indeed a direct perturbative calculation in minkowski scalar qed prove that the amplitudes corresponding to our processes are zero .
the main steps leading to amplitude of pair production in coulomb field ( [ al1 ] ) , will be detailed in this section .
first , we will use the relation for the derivative of hankel functions @xcite : @xmath180 and then the connection between hankel functions and bessel @xmath110 functions @xcite : @xmath181 arriving in this way at integrals of the type @xcite : @xmath182 in our case @xmath183 and the second condition for convergence is satisfied .
we also observe that in our case @xmath184 are complex and for solving our integrals we add to @xmath185 a small real part @xmath186 , with @xmath187 and in the end we take the limit @xmath188 .
this assure the convergence of our integral and will correctly define the unit step functions and @xmath189 functions .
for the second process of particle production from de sitter vacuum , we use the well known formula for hankel functions in terms of bessel @xmath150 functions @xcite : @xmath192 together with : @xmath193 and we arrive at the following integrals @xcite : @xmath194 which depends on variable @xmath195 .
the second type of integrals used are @xcite : @xmath196 these integrals are convergent for @xmath197 .
for this reason we must replace @xmath198 , when we evaluate these integrals . in the end we take the limit @xmath199 . for establishing the final results we use the following relations for legendre functions of second kind @xcite : @xmath200,\end{aligned}\ ] ] when @xmath201 and their relations with hypergeometric functions @xmath152
@xcite @xmath202.\ ] ] the legendre functions of first kind can be written in terms of hypergeometric functions @xmath152 @xcite : @xmath203 mihaela - andreea bloi was supported by the strategic grant posdru/159/1.5/s/137750 , project `` doctoral and postdoctoral programs support for increased competitiveness in exact sciences research '' cofinanced by european social fund within the sectoral operational programme human resources development 2007 - 2013 .
cosmin crucean was supported by a grant of the romanian national authority for scientific research , programme for research - space technology and advanced research - star , project nr .
72/29.11.2013 between romanian space agency and west university of timisoara .
99 c. w. misner , k.s.thorne and j.a.wheleer , _ gravitation _ ( w.h.freeman and company new york , 1973 ) .
r. m. wald , _ general relativity _ ,
( university of chicago press , chicago , 1984 ) .
s. drell and j. d. bjorken , _ relativistic quantum fields _ ( me graw - hill book co. , new york 1965 ) .
l. landau and e. m. lifsit , _ theorie quantique relativiste _ ( mir moscou 1972 ) . m. l. goldberger , k. m. watson , _ collision theory _
( wiley , new york , 1964 ) .
n. d. birrel and p. c. w. davies , _ quantum fields in curved space _ ( cambridge university press , cambridge 1982 ) .
s. weinberg , _ the quantum theory of fields _ , ( cambridge university press , cambridge , 1995 ) . c. itzykson , j. b. zuber , _ quantum field theory _ ( me graw - hill inc .
a. o. barut and i. h. duru , _ phys.rev .
d _ * 36 * , 3705 ( 1987 ) .
v. s. otchik , _ class .
quantum grav .
_ * 2 * , 539 ( 1985 )
. g. v. shishkin , _ class .
quantum grav . _
* 8 * , 175 ( 1991 ) .
i. i. cotescu , _ phys.rev .
d _ * 65 * , 084008 ( 2002 ) .
i. i. cotescu , c. crucean , a. pop , _ int .
* 23 * ( 2008 ) 2563 .
i. i. cotescu and c. crucean , _ progr .
* 124 * ( 2010 ) 1051 .
c. crucean , _
mod.phys.lett.a _ * 22 * , 2573 ( 2007 ) . c. crucean , _ phys .
d _ * 85 * , ( 2012 ) 084036 - 1 .
i. i. cotescu , c. crucean , _ phys .
d _ * 87 * , ( 2013 ) 044016 - 1 .
m. a. bloi , _ mod .
a _ * 29 * , 1450138 ( 2014 ) .
r. blaga , _ arxiv : hep - th 1411.1902 _ , ( 2014 ) . l. parker , _ phys .
* 21 * ( 1963 ) 562 .
l. parker , _ phys .
rev . _ * 183 * ( 1969 ) 1057 .
l. parker , _ phys .
d _ * 3 * ( 1971 ) 346 .
e. schr " odinger , _ physic_a * 6 * , 899 ( 1939 ) . k. h. lotze ,
_ nucl.phys.b _ * 312 * , 673 ( 1989 ) .
k. h. lotze , _ class.quant.grav . _
* 2 * , 351 ( 1988 ) .
k. h. lotze , _ class.quant.grav . _
* 5 * , 595 ( 1985 ) . v. m. villalba , _
phys.rev.d _ * 52 * , 3742 ( 1995 ) . j. garriga , _
phys.rev.d _ * 49 * , 6343 ( 1994 ) .
buchbinder , e.s .
fradkin and d.m .
gitman , _ fortschr .
_ * 29 * , 187 ( 1981 ) .
i. l. buchbinder and l. i. tsaregorodtsev , _ int .
a _ * 7 * , 2055 ( 1992 ) .
h.nomura , m.sasaki , k.yamamoto , _ jcap _ * 11 * , 13 ( 2006 ) .
l. i. tsaregorodtsev , _ russian phys . journal _ * 41 * , ( 1989 ) 1028 .
m. abramowitz and i. a. stegun , _ handbook of mathematical functions _
( dover , new york , 1964 ) .
g. n. watson , _ theory of bessel functions _
( cambridge university press , 1922 ) .
i. s. gradshtein and i. m. ryzhik , _ table of integrals , series and products _
( academic press inc .
, san diego 1980 ) . | we study the theory of interaction between charged scalar field and maxwell field in de sitter background . solving the equation of interacting fields we define the in - out fields as asymptotic free fields and construct the reduction formalism for scalar field .
then we derive the perturbation expansion of the scattering operator .
the first order transition amplitudes corresponding to particle production from de sitter vacuum and pair production in an external field are analysed .
we show that all these effects are important only in strong gravitational fields and vanish in the flat limit . |
with the recent results from the degree angular scale interferometer ( dasi ; @xcite ) , the cosmic anisotropy mapper ( capmap ; @xcite ) and the cosmic background imager ( cbi ; @xcite ) , and the imminent two - year results from the _ wilkinson microwave anisotropy probe _ ( wmap ; @xcite ) , the characterization of the polarization of the cosmic microwave background ( cmb ) polarization is gathering pace .
the ultimate goal of such experiments is to constrain parameters describing the cosmological model and its fluctuations , and hence test models for the generation of these fluctuations , such as inflation . in gaussian theories ,
all of the cosmological information encoded in the cmb anisotropies and their polarization is contained in their power spectra .
for this reason , estimating cmb power spectra is a useful and efficient form of data compression in analysis pipelines @xcite .
however , estimating the spectra is not enough : it is also essential to have an accurate covariance matrix , and preferably an accurate model for the shape of the likelihood function ( see e.g. @xcite ) .
modelling the covariance matrix for a particularly efficient class of polarization power spectra estimators heuristically - weighted quadratic estimators , or pseudo-@xmath0s is the subject of this paper .
the polarization field can be decomposed uniquely on the full sky into a gradient part , with electric ( @xmath1 ) parity , and a curl part with magnetic ( @xmath2 ) parity @xcite .
the cosmological importance of this decomposition is that linear density perturbations can not produce magnetic polarization , and so , in standard scenarios and at linear order , large - angle @xmath2-mode polarization is produced solely by gravitational waves .
second - order effects , most notably weak gravitational lensing by large - scale structure , can generate @xmath2 modes from primary @xmath1 modes @xcite .
the lens - induced @xmath2 modes have a blue spectrum , and should dominate any primary signal from gravitational waves on sub - degree scales . on larger scales , they may still dominate if the gravitational wave amplitude is sufficiently low .
either way , the variance of the @xmath2 modes is expected to be much smaller than that of the @xmath1 modes . given the different parity properties of the @xmath1 and @xmath2 fields , cmb polarization in a statistically - isotropic and parity - respecting universe is characterized by two power spectra , @xmath4 and @xmath5 , and also the cross - power @xmath6 between @xmath1 and the temperature anisotropies .
a number of different methods have been developed to estimate these cmb polarization power spectra .
the optimal method , brute - force maximum likelihood ( e.g. @xcite ) , and its quadratic approximation @xcite , should produce the most accurate ( minimum - variance ) estimates .
however , these methods require the inversion of @xmath7 matrices , where @xmath8 is the number of pixels in the map , which is prohibitively slow for the analysis of mega - pixel maps . given these limitations , a number of fast alternative methods have been suggested @xcite building on earlier work on the analysis of the temperature anisotropies @xcite .
these fast methods are closely related , and all involve compressing heuristically - weighted maps of the stokes parameters into a set of pseudo-@xmath0s making use of spherical transforms a step that requires only @xmath9 operations with fast fourier transform methods on iso - latitude pixelisations such as the widely - used healpix @xcite .
the methods differ in the details of how they transform the pseudo-@xmath0s to power spectrum estimates . in the absence of instrument noise ,
the mean of the pseudo-@xmath0s is linear in the true power spectra , so , for observations over a large enough part of the sky , unbiased power spectrum estimates can be obtained by applying the inverse operation to the pseudo-@xmath0s @xcite . if the observed part of the sky is sufficiently small , direct inversion is not possible and some form of regularisation is required .
@xcite advocate using likelihood methods , while @xcite make use of correlation functions as an intermediate data product
. a notable feature of the latter approach is that the inversion is constructed so that electric and magnetic polarization power is not mixed in the mean .
pseudo-@xmath0 methods have now been applied to a number of real data sets , both for temperature @xcite and polarization analyses @xcite . despite the attention that pseudo-@xmath0 techniques have received , relatively little
work has been done on trying to _ understand _ their error properties .
this paper aims to fill that gap by providing the basis for a reliable , semi - analytic error analysis technique for cmb polarization power spectrum estimates obtained with pseudo-@xmath0 methods .
this extends earlier analytic work that discusses temperature anisotropies only @xcite , and the recent thesis work by one of us @xcite where both temperature and polarization are considered .
a significant complication that arises for polarized data is the mixing of @xmath1 and @xmath2 modes for observations covering only part of the sky @xcite . we develop approximations to the pseudo-@xmath0 sample covariances , taking this mixing into account at leading order for maps with smoothly - apodized edges .
our approximations are accurate in the regime where the mixing is perturbative , i.e. for @xmath10 , where @xmath11 is the effective band - limit of the function used to weight the stokes parameters .
it is difficult to construct good approximations for the covariance on large scales , @xmath12 .
this is partly because of the increased importance of @xmath1-@xmath2 mixing there , but also because the polarization power spectra do not go smoothly to zero on large scales in the presence of reionization .
( the behaviour of the temperature - anisotropy power spectrum on large scales produces a similar effect for the temperature pseudo-@xmath0 covariance . )
following @xcite , the low-@xmath13 sector of the covariance matrices can be constructed directly with exact methods , and used to overwrite the inaccurate approximation .
being able to evaluate accurate covariances efficiently is important in any application where multiple evaluations are required .
examples include parameter estimation , where the covariance generally depends on position in parameter space , and optimising the design and analysis of surveys , where one might wish to assess a large number of weighting schemes . throughout , we ignore the effects of instrument noise so we deal only with the sample covariance of the @xmath14 .
this is to highlight the geometric effects of heuristic weighting that are peculiar to polarization analysis , particularly @xmath1-@xmath2 mixing .
an obvious extension of this work that will be required for application to real data is to include the contribution from instrument noise .
for uncorrelated noise this should be straightforward , but in a more realistic set - up , one will probably have to resort to monte - carlo simulations for the noise contribution . as noted above ,
one of the future goals of cmb polarization observations is to detect ( or constrain ) the stochastic gravitational wave background via its signature in large - angle @xmath2-mode polarization @xcite .
magnetic polarization allows much smaller amplitude gravitational wave backgrounds to be detected than with temperature or electric polarization alone , since , in the latter case , the gravitational - wave signal is swamped by the cosmic variance of the signal from the dominant density perturbations . methods to separate out exactly pure magnetic modes from observations with incomplete sky coverage exist at the map level @xcite .
such methods have the desirable property that quadratic power spectrum estimates formed from the pure-@xmath2 modes have no cosmic variance if the @xmath2-mode power is zero . were it not for gravitational lensing , separating @xmath2 modes at the level of the map would thus allow the detection of arbitrarily low amplitude gravitational wave backgrounds in the absence of instrument noise . in practice ,
imperfections in removing the lensing signal limit the amplitude to tensor - to - scalar ratios @xmath15 ( @xcite ; see @xcite for refinements from more optimal lensing - reconstruction methods ) .
heuristically - weighted pseudo-@xmath0 methods do not share the property that the cosmic variance of the estimated @xmath5 is immune to @xmath1-mode power , even if we take care to construct unbiased estimates so that e.g.the mean of our @xmath5 estimator decouples from @xmath4 @xcite .
we investigate this issue further here , exploring the limits of detectability for @xmath16 that arise solely from cosmic variance of the @xmath1-mode polarization leaking into the variance of quadratic estimates of @xmath5 .
this paper is organised as follows .
section [ sec : properties ] reviews the properties of the polarization pseudo-@xmath0s and establishes our notation . following a brief discussion of the exact pseudo-@xmath0 covariance matrices in section [ sec : exact - cov ] ,
we develop analytic approximations to them that can be rapidly computed in section [ sec : approx - cov ] .
we begin with a toy model , based on a gaussian weight function applied to a small survey area in the flat - sky limit .
( some of the more involved manipulations are relegated to appendix [ app : flat ] . )
this instructive example illustrates the key issues that are specific to the analysis of polarization , i.e. the geometric mixing of @xmath1 and @xmath2 modes .
we then move on to develop approximations for arbitrary ( smooth ) weight functions , and test these against the exact covariances in cases where we can easily compute the latter ( in particular , for azimuthally - symmetric weight functions ) .
simple rules - of - thumb for band - power variances are also derived in section [ subsec : rot ] as limiting cases of the more general expressions ; some of the manipulations of the @xmath17 symbols required for these derivations are relegated to appendix [ app : useful_sums ] .
finally , in section [ sec : implications ] we consider the implications of our results for the detection of gravitational waves with @xmath2-mode polarization estimated with pseudo-@xmath0 methods . throughout
, we illustrate our results with the best - fitting power - law concordance @xmath18cdm model to the _ wilkinson microwave anisotropy probe _
( _ wmap _ ) and 2df galaxy redshift survey data ( @xcite ; table 7 ) , but with the tensor - to - scalar ratio @xmath19 corresponding to @xmath20 inflation .
we include the effects of gravitational lensing in the @xmath2-mode power spectrum , but ignore effects due to its non - gaussianity .
theoretical power spectra , pseudo-@xmath0s and estimated power spectra in all figures include the smoothing effect of a gaussian beam of ten - arcmin fwhm .
we define stokes parameters along the line of sight @xmath21 using the basis vectors of a spherical polar coordinate system .
we take the local @xmath22 axis to be @xmath23 and the @xmath24 axis @xmath25 ; these form a right - handed triad with the radiation propagation direction @xmath26 which defines the local @xmath27 axis .
the complex polarization @xmath28 is then spin -2 @xcite and can properly be expanded in terms of spin-2 spherical harmonics ( see e.g. @xcite for a review ) : @xmath29 reality of the stokes parameters ensures that @xmath30 and similarly for @xmath31 , while under parity transformations @xmath32 ( electric parity ) but @xmath33 ( magnetic parity ) .
the polarization power spectra are defined in statistically - isotropic and parity - respecting models by @xmath34 the @xmath1-@xmath2 power vanishes ( and @xmath2 is uncorrelated with the temperature anisotropies ) by parity . in pseudo-@xmath0 methods
, one adopts a heuristic weighting of the complex polarization with a scalar function @xmath35 .
the weight is necessarily zero in those regions that are not observed , but typically it is desirable to excise a larger region to remove further regions of high foreground emission .
one can attempt to choose @xmath35 to improve the accuracy of subsequent power spectrum estimates .
for example , on small scales , where the instrument noise dominates the signal , the optimal weighting is essentially inverse - noise - variance weighting @xcite , while on large scales uniform weighting is preferable . in this paper
we shall assume that the weight is chosen to be real and to go smoothly to zero as excised regions are approached .
the former restriction ensures that the weighting does not alter the polarization direction , while the latter makes @xmath35 approximately band - limited to some @xmath11 .
such edge apodization allows us to deal with the effects of the resulting @xmath1-@xmath2 mixing in the pseudo-@xmath0 covariance matrices in a perturbative manner ; see section [ sec : approx - cov ] .
the problem of @xmath1-@xmath2 mixing without edge apodization , i.e. taking @xmath35 everywhere one or zero , is discussed extensively by @xcite and @xcite in the context of isolating @xmath2 modes at the map level . following @xcite
, we define pseudo - multipoles @xmath36 and @xmath37 to be the @xmath1 and @xmath2-mode multipoles of the weighted polarization map , so that @xmath38 where @xmath39 are the hermitian coupling matrices @xmath40 using @xmath41 , and the reality of @xmath35 , we have @xmath42 which ensures that @xmath43 and a similar result for @xmath37 . if we expand the weight function in terms of its spherical multipoles , @xmath44 , the coupling matrices can be evaluated in terms of wigner-@xmath17 symbols : @xmath45 from equation ( [ eq:5 - 25 ] ) we can separate the @xmath1 and @xmath2 pseudo - multipoles as @xmath46 where we define the hermitian @xmath47 to be @xmath48 given the @xmath36 and @xmath37 , we form their pseudo-@xmath0s according to @xmath49 we could also form @xmath50 , but , as we show shortly , this vanishes in the mean in the absence of parity violations .
however , it may provide a useful test on parity - violating physics or unaccounted - for systematics , and so future data should , of course , be validated with @xmath51 . in the absence of instrument noise , the mean pseudo-@xmath0s ( i.e.averaged over cmb realisations )
are related to the true power spectra by convolutions .
we find @xcite @xmath52 where @xmath53 \left ( \begin{array}{ccc } l & l ' & l \\ -2 & 2 & 0 \end{array } \right)^2 , \label{eq:11 } \\ m_{ll ' } & \equiv & \frac{1}{2l+1 } \sum_{mm ' } |{}_- i_{(lm)(lm)'}|^2 = \frac{2l'+1}{8\pi } \sum_l ( 2l+1 ) w_l [ 1-(-1)^k ] \left ( \begin{array}{ccc } l & l ' & l \\ -2 & 2 & 0 \end{array } \right)^2 .
\label{eq:12}\end{aligned}\ ] ] here @xmath54 and we have introduced the power spectrum of the weight function @xmath55 .
a non - zero matrix @xmath56 biases the pseudo-@xmath0s by mixing e.g. @xmath1-mode power into @xmath57 . for observations with uniform weight over the full sky , @xmath58 and @xmath59 since then only @xmath60 is non - zero and the @xmath17 symbol forces @xmath61 and hence @xmath62 to be even .
the sum of @xmath63 and @xmath56 is approximately equal to the matrix that relates the temperature @xmath0s to the mean pseudo-@xmath0s @xcite at high @xmath13 .
their sum differs from the temperature case only by the presence of non - zero azimuthal quantum numbers @xmath64 in the @xmath17 symbol .
the mean of @xmath51 evaluates to @xmath65 and so vanishes if parity is preserved in the mean ( @xmath66 ) .
an example of the mean pseudo-@xmath0s for observations over a circular region of radius @xmath67 is shown in fig .
[ fig : power_and_window ] .
the weight function is uniform inside a circle of radius @xmath68 , but the remaining @xmath69 is apodized as @xmath70 $ ] so that @xmath35 falls smoothly to zero at the edge of the observed region . for this weight function
the matrix @xmath63 is roughly two orders of magnitude larger than @xmath56 , and both are narrow compared to features in @xmath4 and @xmath5 except for the reionization signal on the largest scales . on small scales
the mean pseudo-@xmath0s can be approximated by removing @xmath71 and @xmath72 from the convolutions in equation ( [ eq:10 ] ) to give @xmath73 for @xmath74 the mixing term described by @xmath56 is essentially negligible , and the mean of @xmath74 is approximately a scaled version of the underlying power spectrum . the relevant scaling is discussed below . for @xmath57 mixing is important , despite @xmath63 dominating over @xmath56 , since @xmath75 .
this additional contribution to @xmath76 from geometric mixing of @xmath1 and @xmath2 modes can be clearly seen in fig .
[ fig : power_and_window ] , and , as expected , looks like a scaled version of the @xmath1-mode power spectrum .
the normalisations @xmath77 and @xmath78 evaluate to @xmath79 where @xmath80 is defined by @xmath81 , and @xmath82 is the fraction of sky with non - zero weighting .
equation ( [ eq:5 - 35a ] ) follows from the expansions of @xmath63 and @xmath56 in equations ( [ eq:11 ] ) and ( [ eq:12 ] ) , and the orthogonality of the @xmath17 symbols . it is the same normalisation as for the temperature anisotropies . equation ( [ eq:5 - 35 ] ) can be established following the method used to derive equation ( 66 ) in @xcite ; see also appendix [ app : useful_sums ] .
note that for @xmath13 much greater than the assumed band - limit of @xmath35 , the normalisation @xmath83 is suppressed by a factor @xmath84 compared to @xmath85 .
this is consistent with the expectation that @xmath1-@xmath2 mixing is suppressed at high @xmath13 @xcite .
we can gain further insight into this result by noting that [ for smooth @xmath35 ] @xmath87 where @xmath88 is @xmath35 smoothed with a top - hat ( @xmath13-space ) filter .
for band - limited weight functions , at @xmath10 the smoothing has little effect and so to leading order @xmath89 at a given scale @xmath13 , the relative importance of mixing of @xmath1- and @xmath2-mode power in the pseudo-@xmath0s is thus controlled by @xmath90 times the ratio of the average squared gradient of the weight function to the average square weight function .
for the purpose of parameter estimation , it is not strictly necessary to attempt to invert the pseudo-@xmath0s to ( unbiased ) power spectrum estimates .
forming the pseudo-@xmath0s can be regarded as a mildly - lossy form of data compression of the maps of stokes parameters , and inferences about cosmological models can be made directly from them if their covariance or , better , their full sampling distribution can be calculated @xcite . the calculation of the signal contribution to these covariances is the main topic of this paper .
however , for presentation purposes it is useful to be able to plot quantities that are simply related to the underlying power spectra . at
the least this requires some approximate deconvolution of the geometric effects of the weight function from the pseudo-@xmath0s . for observations covering enough of the sky .
] , the matrices @xmath63 and @xmath56 will be invertible . in this case
we easily obtain unbiased estimates of @xmath4 and @xmath5 from @xmath92 alternatively , we can form unbiased estimates of the correlation functions of the stokes parameters from the polarization pseudo-@xmath0s ( or directly from the maps ) , and then invert these with the appropriate legendre transforms @xcite . for observations over smaller parts of the sky ,
@xmath63 and @xmath56 will not be invertible . in this case we do not have the spectral resolution to obtain unbiased estimates of the @xmath0s at every @xmath13 .
various approaches have been suggested to regularise the inversion including the use of band - powers @xcite , or apodized legendre transforms of the ( incomplete ) correlation functions @xcite .
the latter method provides estimates of the @xmath0s convolved with a window function that depends only on the choice of apodizing function that is applied to the correlation function .
@xcite , building on earlier work by @xcite , showed how to generalise the correlation function method to ensure that the estimated polarization power spectra did not mix @xmath1 and @xmath2 power in the mean .
this is particularly useful for visualisation of the @xmath2-mode power spectrum which can easily be swamped by the dominant @xmath1 modes .
we do not reproduce the method of @xcite here , but note only that it provides a convenient way of constructing pseudo - inverses @xmath93 and @xmath94 to @xmath63 and @xmath56 respectively with the properties that @xmath95 these relations ensure that our estimates of the power spectra , @xmath96 satisfy @xmath97 and the equivalent relation for @xmath5 .
the normalised window function @xmath98 , where @xmath99 is given by @xcite @xmath100 here , @xmath101 are the reduced wigner rotation matrices and @xmath102 is the apodizing function that is applied to the correlation functions .
an example of the pseudo - inverses obtained via this route is shown in fig . [
fig : power_and_window ] , along with the mean of the recovered @xmath0s .
the correlation functions have been apodized with a gaussian with hwhm of @xmath103 .
the apodizing function was deliberately chosen to be narrow to ensure that the resulting pseudo - inverses @xmath93 and @xmath94 are well localised .
given that the pseudo - inverses are significantly broader than @xmath63 and @xmath56 , the window function @xmath99 ( not shown in the figure ) inherits the width of the pseudo - inverses .
for this reason the spectral resolution of the recovered @xmath0s is rather poor compared to what should ultimately be achievable for a @xmath67-radius survey .
the effect of this is that the recovered @xmath0s accurately follow the true @xmath0s only for @xmath104 .
the covariance of any quadratic power spectrum estimates that derive from the pseudo-@xmath0s by linear transformations follow simply from the covariance of the latter .
for this reason , in this paper we concentrate on the pseudo-@xmath0 covariance .
furthermore , we focus on the sample covariance ( i.e. we neglect instrument noise ) to highlight those aspects specific to cmb polarization that arise from mode - mixing on the incomplete sky . including instrument noise is obviously important for application to near - future surveys since it will dominate sample variance .
extending the analytic results derived in this paper to include simple white noise should be straightforward ( see @xcite for the case of the temperature anisotropies ) .
however , for more realistic correlated noise , monte - carlo simulations will probably be required to account properly for the noise contribution to the covariances .
we shall assume throughout that the cmb signal is gaussian .
if the initial fluctuations were very - nearly gaussian , as expected in simple single - field inflation models ( e.g. @xcite ) , non - linear effects will introduce only a small level of non - gaussianity into @xmath1 modes on the scales of interest for primary cmb science ( e.g. @xcite ) . however , if the primordial tensor - to - scalar ratio @xmath105 , then the dominant power in @xmath2 modes is expected to come from gravitational - lensing conversion of gaussian @xmath1 modes into non - gaussian @xmath2 modes . in this limit , our assumption of gaussianity will be violated and the expressions derived in this paper will require some modification to account for the non - zero connected four - point function . while a full treatment of the pseudo-@xmath0 covariance in the presence of lens - induced @xmath2 modes is beyond the scope of the current paper , we note that the lensing contribution to @xmath106 has been calculated recently for uniform weighting in the flat - sky limit @xcite . in this section
we calculate the exact sample covariance of @xmath74 and @xmath57 ( see also @xcite for a similar calculation ) and present the results of numerical computations for two types of survey : a full - sky survey with a simple galactic cut , and a @xmath67-radius circular survey of the type that is optimal for gravitational wave searches in the presence of instrument noise at the level achievable with current technology @xcite . for gaussian cmb fields , the pseudo - multipoles are also gaussian ,
so the covariance of their power spectra are given by e.g.@xmath107 if we expand the pseudo - multipoles in terms of the true multipoles using equation ( [ eq:5 - 25 ] ) , we find @xmath108 evaluating these expressions numerically is straightforward in principle , but is computationally expensive , with an operation count of @xmath109 if multipoles up to @xmath13 are retained .
this is prohibitive for @xmath13 above a few hundred , and so , in general , there is a need for faster approximate calculations .
one such approximation is to compute the covariances from monte - carlo simulations ; this has the benefit of allowing one easily to include a number of real - world effects that are more difficult to include in the analytic calculation .
an alternative , which is developed here in section [ sec : approx - cov ] , is to use analytic approximations ( that can be calibrated with simulations if required ) . as we show later , it is difficult to find good analytic approximations at low @xmath13 because of the expected sharp rise in the polarization @xmath0s there , and the increased importance of @xmath1-@xmath2 mixing .
a similar problem arises for the temperature anisotropies because the @xmath0s do not go smoothly to zero as @xmath110 .
@xcite suggests replacing the block of the approximate covariance matrices with small @xmath13 and @xmath111 with the exact expression , since the latter can be calculated quickly at small @xmath13 and @xmath111 . for diagonally - dominant matrices
, this procedure returns an accurate covariance matrix ; a similar procedure can be implemented for polarization using equations ( [ eq:5 - 37])([eq:5 - 39 ] ) and the approximations developed in section [ sec : approx - cov ] . as our first application of equations ( [ eq:5 - 37])([eq:5 - 39 ] ) , we consider a full - sky survey from which we remove a @xmath112 band around the galactic plane .
we adopt a north - south symmetric weighting which is uniform ( which we expect to be nearly optimal in the noise - free limit ) except for the edge of the retained survey which we apodize with the square of a cosine . in the northern hemisphere @xmath113 & \mbox{for $ 65\degr \le \theta \le 70\degr$. } \end{array } \right .
\label{eq : exact_cov1}\ ] ] since @xmath35 is azimuthally - symmetric , the matrices @xmath114 are block diagonal and the covariances can be evaluated exactly in @xmath115 operations .
figures [ fig : exactee_galcosine][fig : exacteb_galcosine ] show the exact covariance matrices @xmath116 , @xmath117 and @xmath118 respectively .
the prefactors have been chosen to make the elements near the diagonal more nearly uniform . given that the weight function is symmetric under parity , @xmath119 and @xmath120 , so that @xmath116 and @xmath106 vanish if @xmath121 is odd , while @xmath122 vanishes if @xmath121 is even ( and hence on the diagonal ) . in each of the three covariance matrices ,
the intense feature for @xmath13 and @xmath123 ( which saturates the colour scale ) arises predominantly from the additional power in @xmath4 on large scales due to reionization .
( it vanishes if we substitute for power spectra with no reionization . ) in the case of @xmath116 , the first term inside the summation over @xmath124 and @xmath125 in equation ( [ eq:5 - 37 ] ) is always dominant .
the covariance is thus very similar to that for the temperature anisotropies with the same weight function .
the matrix is diagonally dominant , except on large scales , with width inversely related to the angular dimension of the retained region . on large scales
the covariance matrix is dominated by the additional polarized power generated by reionization .
note that this power propagates to somewhat smaller scales in the pseudo-@xmath0 covariance due to residual fourier ringing in the matrix @xmath126 .
( for the weight function adopted here , the width of the main peak of @xmath47 is given by the inverse of the survey size , but the effective band - limit is larger being set by the inverse size of the apodized region . ) to emphasise the importance of @xmath1-@xmath2 mixing in the structure of the covariance matrix for @xmath57 , in fig
. [ fig : exactbb_galcosine ] we show both the exact @xmath127 and the matrix we obtain by setting @xmath128 .
the latter is very inaccurate on the diagonal at @xmath129 , where @xmath1-@xmath2 mixing is most important , and particularly so at the largest scales where it further fails to pick up the additional power in @xmath4 from reionization .
setting @xmath4 to zero also artificially suppresses the correlations : there is significantly more structure off the diagonal in the left - hand plot in fig .
[ fig : exactbb_galcosine ] than on the right .
this reflects the fact that @xmath130 is intrinsically broader than @xmath131 ( see fig . [
fig : power_and_window ] ) .
the covariance matrix @xmath122 , shown in fig .
[ fig : exacteb_galcosine ] , is only non - zero because of @xmath1-@xmath2 mixing .
it is correspondingly broader than @xmath116 .
the dominant contribution is from @xmath1-mode power , so the covariance matrix peaks near the diagonal at the positions of the acoustic peaks in @xmath4 . as our second example we consider the @xmath67-radius survey with cosine apodization over the last @xmath69 that we used in fig .
[ fig : power_and_window ] .
we should expect the relative importance of @xmath1-@xmath2 mixing to extend to higher @xmath13 for this smaller survey , and this is indeed seen in fig .
[ fig : exactbb_10_15_cosine_and_exactbb_10_15_cosine_bonly ] where we compare the exact @xmath132 with the covariance obtained by setting @xmath128 .
the @xmath1-mode power dominates the covariance for all scales plotted ( maximum @xmath133 ) , and the magnitude of the diagonal follows the acoustic peaks in @xmath4 .
surveys of this size are being considered by a number of upcoming experiments targeting @xmath2-mode polarization from gravitational waves .
it is clear that , even for the relatively optimistic tensor - to - scalar ratio adopted here ( @xmath19 ) , the application of pseudo-@xmath0 techniques to such surveys will demand careful modelling of the impact of @xmath1-@xmath2 mixing in the covariance matrix for @xmath57 on any scales where sample variance is expected to dominate .
given that exact evaluation of the pseudo-@xmath0 covariance matrices is , in general , prohibitively expensive , in this section we aim to derive accurate analytic approximations that allow fast computation of the covariances at high @xmath13 .
this problem has been well - studied for temperature anisotropies @xcite , but , as yet , there has been no comparable treatment for polarization . as with the temperature case ,
the main simplification arises from assuming that we are working at sufficiently high @xmath13 , and with a weight function that is sufficiently band - limited , that we can approximate the @xmath0s as being constant over the width of all coupling matrices .
this removes the @xmath0s from all convolutions . for @xmath134 , and its products ,
this approximation should be good , as in the temperature case , for smoothly - apodized observations covering a connected patch of linear extent more than several degrees .
the approximation is poorer for @xmath135 , and breaks down completely if there is no apodization , i.e. the weight function @xmath35 is set to unity or zero .
( see @xcite for examples of elements of @xmath136 in this case . ) here we restrict ourselves to situations where these assumptions hold .
we first consider a specific case of gaussian weighting over small patches of the sky , before considering general ( but smooth ) weighting on the sphere in section [ subsec : general ] . as a warm - up to the approximate calculation of the pseudo-@xmath0 covariances for general weighting functions
@xmath35 , we first consider gaussian weighting over small patches of the sky .
for such observations we can approximate the spherical - harmonic analysis by fourier analysis , and we are able to make analytic progress more simply than for the general case .
the results of this subsection are also of considerable interest for observations of cmb polarization with interferometers [ such as dasi @xcite , cbi @xcite and the array for microwave background anisotropy ( amiba ) ] , since they directly sample the fourier transform of the sky multiplied by the primary beam .
the latter thus acts as a weight function @xmath35 for an interferometer .
furthermore , the primary beam can often be approximated as a gaussian . in the flat - sky approximation , for lines of sight close to the @xmath27 axis ,
we adopt a global polarization basis that is aligned with the @xmath22 and @xmath137 axes .
the radiation propagation direction at the centre of the field is along @xmath138 .
stokes parameters defined on this basis , @xmath139 and @xmath140 , at position @xmath141 on the sky can be decomposed into fourier modes @xmath142 and @xmath143 : @xmath144 from these we can define the fourier transforms of the electric and magnetic parts of the polarization , @xmath145 where @xmath146 is the angle between @xmath147 and the @xmath22 axis .
note that this is a rotation of the stokes parameters in fourier space onto a basis adapted to the wavevector @xmath147 .
the fourier transforms @xmath148 and @xmath149 can be shown to be related to the spherical multipoles @xmath150 and @xmath31 for @xmath151 by the same correspondence as for the temperature anisotropies , i.e.@xmath152 a similar relation holds for @xmath2 . in terms of the power spectra ,
we have e.g. @xmath153 .
if the polarization is weighted by @xmath154 , then the fourier transform of the stokes parameters for the weighted field ( or the visibilities in the case of an interferometer ) are @xmath155 with a similar expression for @xmath156 . here
@xmath157 is the fourier transform of @xmath154 . for @xmath154 real , @xmath158 and similarly for @xmath156 .
decomposing @xmath159 and @xmath156 as in equation ( [ eq : flat2 ] ) , we find @xmath160 or equivalently @xmath161 , \label{eq : flat6 } \\
\tilde{b}({{\bmath{l } } } ) & = & \int \frac{d^2 { { \bmath{l}}}'}{2\pi } [ { } _ + i({{\bmath{l}}},{{\bmath{l } } } ' ) b({{\bmath{l } } } ' ) - i { } _ -i({{\bmath{l}}},{{\bmath{l } } } ' ) e({{\bmath{l } } } ' ) ] , \label{eq : flat7}\end{aligned}\ ] ] where the hermitian [ for @xmath154 real ] kernels are @xmath162 these are the flat - sky limits of the @xmath47 in equations ( [ eq:5 - 27 ] ) and ( [ eq:5 - 28 ] ) .
the geometric mixing of e.g. @xmath148 into @xmath163 is suppressed by the trigonometric factor @xmath164 .
if the extent of the support of @xmath157 is @xmath165 [ i.e.the band - limit of @xmath35 ] then the mixing of @xmath1 into the observable @xmath166 will be @xmath167 for @xmath10 .
clearly this is significant if @xmath168 . in the flat - sky limit
the pseudo-@xmath0s are defined by @xmath169 the signal correlations between the @xmath170 and @xmath171 are discussed in appendix [ app : flat ] . for gaussian weighting , @xmath172 , and smooth spectra ( over a range of multipoles @xmath173 ) , we can use equations ( [ eq : flat18])([eq : flat20 ] ) to show that @xmath174 where @xmath175 , \label{eq : flat27 }
\\ \beta(l ) & = & \frac{1}{2 l^2}\left [ 1-\frac{3}{2l^2\sigma^2}+ \frac{1}{2}e^{-l^2\sigma^2 } \left(1+\frac{3}{l^2\sigma^2}\right)\right ] .
\label{eq : flat28}\end{aligned}\ ] ] these results are consistent with equations ( [ eq:5 - 35a ] ) and ( [ eq:5 - 35 ] ) , if we use @xmath176 and @xmath177 for @xmath151 , since the ( spherical ) power spectrum of the gaussian weight function @xmath178 .
the sample covariances of the pseudo-@xmath0s are given by the angular average of the squares of the correlators of @xmath170 and @xmath171 : @xmath179 for an azimuthally - symmetric weight function , the correlators in these integrals are invariant under rigid rotations of @xmath147 and @xmath180 , so the integration over one of @xmath181 and @xmath182 is trivial . to evaluate equations ( [ eq : flat38])([eq : flat40 ] ) for the case of gaussian weighting we make use of the results for the correlators given in appendix [ app : flat ] .
the analysis is rather involved so we simplify things a little by considering only the limits @xmath183 and @xmath128 .
if we consider the covariance of @xmath74 with itself , the case @xmath183 describes the contribution to the covariance if there were no @xmath1-@xmath2 mixing , while the case @xmath128 describes the contribution that arises solely from mixing .
consider @xmath183 first .
then , for azimuthally - symmetric @xmath154 , we have e.g.@xmath184 where @xmath185 , and similar expressions hold for the other covariances . we have not been able to perform the integrals over @xmath186 exactly , but it is possible to make progress in the limit of large @xmath13 and @xmath111 ( @xmath187 ) . in this limit , we can ignore the terms with factor @xmath188 $ ] in equations ( [ eq : flat20 ] ) and ( [ eq : flat23 ] ) in appendix [ app : flat ] .
this leaves a factor of @xmath189 multiplying slowly - varying functions of @xmath186 .
the exponential peaks sharply around @xmath190 ( reflecting the fact that @xmath170 and @xmath191 are only strongly correlated for @xmath147 within a radius @xmath173 of @xmath180 ) , so we can accurately handle the inverse powers of @xmath192 and trigonometric terms in equations ( [ eq : flat20 ] ) and ( [ eq : flat23 ] ) with an expansion in @xmath193 .
to make this expansion it is useful to note that @xmath194 \sin\psi .
\label{eq : flat43 } \end{aligned}\ ] ] integrating the square of the expansions over @xmath186 then yields a series of modified bessel functions @xmath195 .
these series are cumbersome , so we only give the leading - order results in the asymptotic regime ( @xmath13 , @xmath196 ) , where we can replace the bessel functions with their asymptotic expansions : @xmath197 to obtain these results it is necessary to expand the integrand in equation ( [ eq : flat41 ] ) to @xmath198 .
it is straightforward to verify that the covariances of the pseudo-@xmath0s for the other limit , @xmath128 , can be obtained by interchanging @xmath1 and @xmath2 in these expressions . to gain an indication of how the covariance of the pseudo-@xmath0s propagates to estimates of the power spectra , we form simple estimators for @xmath4 and @xmath5 that are local in @xmath74 and @xmath57 by solving equation ( [ eq : flat37 ] ) : @xmath199 , \quad \hat{c}_l^b = \frac{1}{n(l)}[\alpha(l ) \tilde{c}_l^b - \beta(l ) \tilde{c}_l^e ] , \label{eq : flat30}\ ] ] where the normalisation @xmath200 .
the covariance of these estimators then follows from the covariance of the pseudo-@xmath0s which we derived above . to leading order
, we find @xmath201 \quad ( c_l^b=0 ) , \label{eq : flat47}\ ] ] and @xmath202 note that these expressions are symmetric in @xmath13 and @xmath111 as required .
equation ( [ eq : flat48 ] ) , which gives the covariance of the estimated @xmath2-mode power if there were no @xmath1-@xmath2 mixing , is the direct analogue of the asymptotic expression for the temperature anisotropies , where the exact result in the flat - sky limit ( for smooth @xmath203 ) is @xmath204 if we average the estimators into bands of width @xmath205 we obtain quasi - uncorrelated estimates .
the variances of these band - powers follows by integrating equations ( [ eq : flat47 ] ) and ( [ eq : flat48 ] ) over @xmath13 and @xmath111 within the band . for wide bands , we find the approximate results @xmath206 for the @xmath128 case , we recover the approximation given for temperature by @xcite , @xmath207 noting that , for the case of gaussian weighting considered here , @xmath208 and @xmath209 . in section [ subsec : rot ] we develop more general rules of thumb for arbitrary ( but smooth ) weight functions ; these properly reduce to the expressions derived here for gaussian weighting in the flat - sky limit .
our leading - order results show that , for gaussian weighting , the ratio of the cosmic variance contributions to @xmath210 from @xmath1 and @xmath2 is @xmath211 for @xmath212 . in section [ sec : implications ] we consider the implications of this residual variance from @xmath1-mode power in unbiased quadratic estimates of @xmath5 for detecting gravitational waves via @xmath2-mode polarization .
we now consider the case of general ( but smooth ) weight functions and work on the spherical sky .
the starting point is to approximate the correlators of the pseudo - multipoles that appear squared in the pseudo-@xmath0 covariances , equations ( [ eq:5 - 37])([eq:5 - 39 ] ) , by removing the @xmath0s from the convolutions : @xmath213_{(lm)(lm ) ' } - \left ( \sqrt{c_l^e c_{l'}^e } - \sqrt{c_l^b c_{l'}^b } \right ) [ { } _ { -}i^2(w)]_{(lm)(lm ) ' } , \label{eq:5 - 40}\\ \langle \tilde{b}_{lm } \tilde{b}_{(lm)'}^ * \rangle & = & \sum_{lm } { } _ { + } i_{(lm)(lm)}{}_{+}i_{(lm ) ' ( lm)}^ * c_l^b + { } _ { -}i_{(lm)(lm)}{}_{- } i_{(lm ) ' ( lm)}^ * c_l^e \nonumber \\ & \approx & \sqrt{c_l^b c_{l'}^b } [ { } _ { + } i(w^2)]_{(lm)(lm ) ' } + \left ( \sqrt{c_l^e c_{l'}^e } - \sqrt{c_l^b c_{l'}^b } \right ) [ { } _ { -}i^2(w)]_{(lm)(lm ) ' } , \label{eq:5 - 41}\\ \langle \tilde{e}_{lm } \tilde{b}_{(lm)'}^ * \rangle & = & i \sum_{lm } { } _ { + } i_{(lm)(lm)}{}_{-}i_{(lm ) ' ( lm)}^ * c_l^e + { } _ { -}i_{(lm)(lm)}{}_{+ } i_{(lm ) ' ( lm)}^ * c_l^b \nonumber \\ & \approx & \frac{i}{2 } \left(\sqrt{c_l^e c_{l'}^e}+ \sqrt{c_l^b c_{l'}^b}\right ) [ { } _ { -}i(w^2)]_{(lm)(lm ) ' } \nonumber \\ & & \mbox { } + \frac{i}{2 } \left(\sqrt{c_l^e c_{l'}^e}-\sqrt{c_l^b c_{l'}^b}\right ) \left[{}_{+}i(w){}_{-}i(w)- { } _ { -}i(w){}_{+}i(w)\right]_{(lm)(lm ) ' } , \label{eq:5 - 42}\end{aligned}\ ] ] where @xmath214_{(lm)(lm)'}$ ] are the matrices @xmath215 but evaluated with weight function @xmath216 rather than @xmath35 , and @xmath217_{(lm)(lm)'}$ ] is the matrix product @xmath218 . ,
@xmath217_{(lm)(lm)'}$ ] is also equal to @xmath219 . ]
we now consider the derivation of equations ( [ eq:5 - 40])([eq:5 - 42 ] ) in detail .
we start from the completeness relation @xmath220 which ensures that @xmath221_{(lm)(lm ) ' } & = & \sum_{lm } { } _ { \pm 2}i_{(lm)(lm ) } { } _ { \pm 2 } i^*_{(lm ) ' ( lm ) } \nonumber\\ & = & \int { { \rmn{d}}}{\hat{\bmath{n}}}\ ; w^2 ( { \hat{\bmath{n } } } ) { } _ { \pm 2 } y_{lm}^ * ( { \hat{\bmath{n } } } ) { } _ { \pm 2 } y_{(lm)'}({\hat{\bmath{n } } } ) \nonumber \\ & \equiv & [ { } _ { \pm 2 } i(w^2)]_{(lm)(lm)'}. \label{eq:5 - 43}\end{aligned}\ ] ] for @xmath222 or 0 , so that @xmath223 , these relations imply that @xmath224 are projection operators @xcite .
if we express @xmath225 in terms of @xmath226 , using the obvious matrix notation , we find @xmath227 adding and subtracting gives @xmath228 which can be used to verify the second equalities in equations ( [ eq:5 - 40])([eq:5 - 42 ] ) . note that if @xmath4 were equal to @xmath5 , the correlators would only involve @xmath229 and the construction of the covariances would then be no more difficult than for the temperature anisotropies .
the major simplification that arises in this case is that the correlators of the pseudo - multipoles reduce to linear functionals of @xmath216 .
for statistically - isotropic cmb signals , the pseudo-@xmath0 covariance must be invariant under rotations of @xmath35 , so that if the correlators are linear in @xmath216 then the covariance matrix can only depend on rotationally - invariant quadratic combinations of the multipoles of @xmath230 , i.e. the power spectrum of @xmath230 . in the general case , where the pseudo - multipole correlators are more - general quadratic functionals of @xmath35 ,
the covariance matrix can depend on less - specific configurations of the trispectrum of @xmath35 .
( the power spectrum of @xmath216 picks up only rather specific and easy to calculate
configurations of the full trispectrum . ) forming the covariances from equations ( [ eq:5 - 37])([eq:5 - 39 ] ) , we now find @xmath231_{(lm)(lm ) ' } - \left ( \sqrt{c_l^e c_{l'}^e } - \sqrt{c_l^b c_{l'}^b } \right ) [ { } _ { -}i^2(w)]_{(lm)(lm)'}\right|^2 , \label{eq:5 - 46}\\ { \rmn{cov } } \left(\tilde{c}_l^b,\tilde{c}_{l'}^b\right ) & \approx & \frac{2}{(2l+1)(2l'+1 ) } \sum_{mm ' } \left|\sqrt{c_l^b
c_{l'}^b } [ { } _ { + } i(w^2)]_{(lm)(lm ) ' } - \left ( \sqrt{c_l^b c_{l'}^b } - \sqrt{c_l^e c_{l'}^e } \right ) [ { } _ { -}i^2(w)]_{(lm)(lm)'}\right|^2 , \label{eq:5 - 47}\\ { \rmn{cov } } \left(\tilde{c}_l^e,\tilde{c}_{l'}^b\right ) & \approx & \frac{2}{(2l+1)(2l'+1 ) } \sum_{mm ' } \left| \frac{i}{2 } \left(\sqrt{c_l^e c_{l'}^e}+\sqrt{c_l^b c_{l'}^b}\right ) [ { } _ { -}i(w^2)]_{(lm)(lm ) ' } \right .
\nonumber \\ & & \hspace{0.2\textwidth } \mbox { } \left .
+ \frac{i}{2 } \left(\sqrt{c_l^e c_{l'}^e}-\sqrt{c_l^b c_{l'}^b}\right ) \left[{}_{+}i(w){}_{-}i(w)- { } _ { -}i(w){}_{+}i(w)\right]_{(lm)(lm ) ' } \right|^2 .
\label{eq:5 - 48}\end{aligned}\ ] ] it is straightforward to compute @xmath229 in terms of wigner-3@xmath232 symbols using equation ( [ eq:5 - 26 ] ) : @xmath233_{(lm)(lm ) ' } = \frac{1}{2 } \sum_{lm } ( -1)^m ( w^2)_{lm } \sqrt{\frac{(2l+1)(2l'+1)(2l+1)}{4\pi } } [ 1\pm ( -1)^{k } ] \left ( \begin{array}{ccc } l & l ' & l \\ -2 & 2 & 0 \end{array } \right ) \left ( \begin{array}{ccc } l & l ' & l \\ m & -m ' & -m \end{array } \right ) , \label{eq:5 - 49}\ ] ] where @xmath234 are the ( scalar ) multipoles of @xmath216 , and , as before , @xmath235 . for @xmath236 , we go back one step , taking as our starting point the integral expression for @xmath136 : @xmath237 .
\label{eq:5 - 50}\end{aligned}\ ] ] our aim is to find an approximation to @xmath238 that is a functional of the squares of derivatives of @xmath239 , so that its contribution to the covariance matrices will only depend on easily - calculable configurations of the trispectrum of @xmath239 .
writing the spin-2 harmonics in equation ( [ eq:5 - 50 ] ) in terms of spin - weight derivatives of the scalar harmonics ( see e.g. appendix b of @xcite ) , @xmath240 which follow from repeated use of @xmath241 integrating by parts and using @xmath242 , we find that @xmath243 y_{(lm)'}({\hat{\bmath{n } } } ) .
\label{eq:5 - 53}\ ] ] this result shows explicitly that @xmath136 vanishes if @xmath35 is a constant over the full sky .
more generally , for @xmath35 band - limited to @xmath11 , it suggests that @xmath244 for @xmath245 with this leading - order contribution coming from the last two terms in equation ( [ eq:5 - 53 ] ) .
again , this is consistent with our expectation that @xmath1-@xmath2 mixing is suppressed for @xmath10 .
forming the square of @xmath136 , we find @xmath246_{(lm)(lm ) ' } & = & \frac{1}{4 } \int{{\rmn{d}}}{\hat{\bmath{n}}}_1{{\rmn{d}}}{\hat{\bmath{n}}}_2\ , \biggl\ { \left [ \eth^2 w({\hat{\bmath{n}}}_1 ) { } _ { 2}y_{lm}^*({\hat{\bmath{n}}}_1 ) - { \bar{\eth}}^2 w({\hat{\bmath{n}}}_1 ) { } _ { -2}y_{lm}^*({\hat{\bmath{n}}}_1 ) + 2 \eth w({\hat{\bmath{n}}}_1 ) \eth { } _ { 2}y_{lm}^*({\hat{\bmath{n}}}_1 ) - 2 { \bar{\eth}}w({\hat{\bmath{n}}}_1 ) { \bar{\eth}}{}_{-2}y_{lm}^*({\hat{\bmath{n}}}_1 ) \right ] \nonumber \\ & & \times \left.\left [ \eth^2 w({\hat{\bmath{n}}}_2 ) { } _ { 2}y_{(lm)'}^*({\hat{\bmath{n}}}_2 )
-{\bar{\eth}}^2 w({\hat{\bmath{n}}}_2 ) { } _ { -2}y_{(lm)'}^*({\hat{\bmath{n}}}_2 ) + 2 \eth w({\hat{\bmath{n}}}_2 ) \eth { } _ { 2}y_{(lm)'}^*({\hat{\bmath{n}}}_2 ) - 2 { \bar{\eth}}w({\hat{\bmath{n}}}_2 ) { \bar{\eth}}{}_{-2 } y_{(lm)'}^*({\hat{\bmath{n}}}_2 ) \right]^ * \right .
\nonumber \\ & & \times \left .
\sum_{l\ge 2,m } y_{lm}({\hat{\bmath{n}}}_1 ) y_{lm}^*({\hat{\bmath{n}}}_2 ) \frac{(l-2)!}{(l+2 ) ! } \right\}. \label{eq:5 - 54}\end{aligned}\ ] ] although we shall not make use of it here , we note for completeness that the sum over @xmath124 in this equation can be performed exactly as follows : ( i ) apply the operator @xmath247 to remove the factor @xmath248 ; ( ii ) perform the resulting summation with the addition theorem ; ( iii ) solve the differential equation away from @xmath249 ; and ( iv ) enforce the correct behaviour at @xmath249 .
the final result is @xmath250 where @xmath251 .
the first two terms can be shown not to contribute to @xmath252 . for our purposes
, it is more useful to approximate the sum in equation ( [ eq:5 - 54 ] ) , for @xmath253 , by @xmath254 note that this expression only makes sense when considered _ inside the integrals _ in equation ( [ eq:5 - 54 ] ) .
the justification for the approximation is as follows .
equation ( [ eq:5 - 54 ] ) may be regarded as a convolution of the slowly - varying @xmath248 with the product of two functions of @xmath124 : one resulting from the integral over @xmath255 , which is localised to within @xmath11 of @xmath13 , and the other from the integral over @xmath256 , which is within @xmath11 of @xmath111 .
we can thus replace @xmath248 with @xmath257 in the convolution for @xmath13 and @xmath258 .
furthermore , we can then extend the summation to include the @xmath259 and @xmath260 terms since they sum to a constant plus @xmath261 which , like the first two terms on the right of equation ( [ eq : gen1 ] ) , do not contribute to the integral .
( these terms would be negligible anyway , for large @xmath13 and @xmath111 . )
the presence of the delta function in equation ( [ eq:5 - 55 ] ) ensures that we achieve our aim of reducing @xmath262 to a linear functional of squares of derivatives of @xmath35 .
however , before proceeding we make one further approximation which can easily be relaxed if more accuracy is required .
since @xmath35 is assumed to be slowly varying compared to the spherical harmonics , we ignore terms like @xmath263 compared to e.g. @xmath264 , so that we retain only the leading - order contribution to @xmath262 .
finally , with these approximations we have @xmath246_{(lm)(lm ) ' } & \approx & \sqrt{\frac{(l-2)!}{(l+2)!}}\sqrt{\frac{(l'-2)!}{(l'+2 ) ! } } \int { { \rmn{d}}}{\hat{\bmath{n}}}\left\ { \left [ \eth w({\hat{\bmath{n } } } ) \eth { } _ { 2}y_{lm}^*({\hat{\bmath{n } } } ) - { \bar{\eth}}w({\hat{\bmath{n } } } ) { \bar{\eth}}\ ! { } _ { -2}y_{lm}^*({\hat{\bmath{n } } } ) \right ] \right . \nonumber \\ & & \mbox { } \left . \times \left [ \eth w({\hat{\bmath{n } } } ) \eth { } _ { 2}y_{(lm)'}^*({\hat{\bmath{n } } } ) - { \bar{\eth}}w({\hat{\bmath{n } } } ) { \bar{\eth}}{}_{-2}y_{(lm)'}^*({\hat{\bmath{n } } } ) \right]^ * \right\ } \nonumber \\ & = & \frac{(-1)^{m+1}}{\sqrt{l(l+1)l'(l'+1)}}\sum_{lm } \sqrt{\frac{(2l+1)(2l'+1)(2l+1)}{4\pi } } \left ( \begin{array}{ccc }
l & l ' & l \\ m & -m ' & -m \end{array } \right ) \nonumber \\ & & \times \left\ { \left| \eth w \right|^2_{lm } [ 1+(-1)^{k } ] \left ( \begin{array}{ccc } l & l ' & l \\ -1 & 1 & 0 \end{array } \right ) + \left [ { } _ { 2}(\eth w)^2_{lm } + ( -1)^{k } { } _ { -2}({\bar{\eth}}w)^2_{lm } \right ] \left ( \begin{array}{ccc } l & l ' & l \\ -1 & -1 & 2 \end{array } \right ) \right\ } , \label{eq:5 - 56}\end{aligned}\ ] ] where we have used equation ( [ eq:5 - 52 ] ) . note that this approximation preserves hermiticity of @xmath262 . in deriving equation ( [ eq:5 - 56 ] )
we have expanded products of the derivatives of @xmath35 in the appropriate spin - weight harmonics : @xmath265^ * & = & { \bar{\eth}}w({\hat{\bmath{n } } } ) [ { \bar{\eth}}w({\hat{\bmath{n}}})]^ * = |\eth w({\hat{\bmath{n}}})|^2 = \sum_{l\ge 0,\,m } |\eth w|^2_{lm } y_{lm}({\hat{\bmath{n } } } ) , \\
\eth w({\hat{\bmath{n } } } ) [ { \bar{\eth}}w({\hat{\bmath{n}}})]^ * & = & [ \eth w({\hat{\bmath{n}}})]^2 = \sum_{l\ge 2,\,m } { } _ { 2}(\eth w)^2_{lm } { } _ { 2}y_{lm}({\hat{\bmath{n } } } ) , \\ { \bar{\eth}}w({\hat{\bmath{n } } } ) [ \eth w({\hat{\bmath{n}}})]^ * & = & [ { \bar{\eth}}w({\hat{\bmath{n}}})]^2 = \sum_{l\ge 2,\,m } { } _ { -2}({\bar{\eth}}w)^2_{lm } { } _ { -2}y_{lm}({\hat{\bmath{n } } } ) .
\label{eq:5 - 57}\end{aligned}\ ] ] note that @xmath266 is the complex conjugate of @xmath267 since @xmath239 is real , and that the spin-2 multipoles satisfy @xmath268 . substituting equations ( [ eq:5 - 49 ] ) and ( [ eq:5 - 56 ] ) into equations ( [ eq:5 - 46 ] ) and ( [ eq:5 - 47 ] ) , we have @xmath269 \left| \sqrt{c_l^e c_{l'}^e } ( w^2)_{lm } \left ( \begin{array}{ccc } l & l ' & l \\ -2 & 2 & 0 \end{array } \right ) \nonumber \right .
\\ & & \mbox { } \left .
+ \frac{2\left(\sqrt{c_l^e c_{l'}^e}-\sqrt{c_l^b c_{l'}^b}\right ) } { \sqrt{l(l+1)l'(l'+1 ) } } \left [ |\eth w|^2_{lm } \left ( \begin{array}{ccc } l & l ' & l \\ -1 & 1 & 0 \end{array } \right ) + { { \mathcal{e}}}_{lm } \left ( \begin{array}{ccc } l & l ' & l \\ -1 & -1 & 2 \end{array } \right ) \right ] \right|^2 \biggr\ } \nonumber \\ & & \mbox { } + \frac{1}{\pi } \left [ \frac{\sqrt{c_l^e c_{l'}^e}-\sqrt{c_l^b c_{l'}^b}}{\sqrt{l(l+1)l'(l'+1)}}\right]^2 \sum_{lm } [ 1-(-1)^{k } ] |{{\mathcal{b}}}_{lm}|^2 \left ( \begin{array}{ccc } l & l ' & l \\ -1 & -1 & 2 \end{array } \right)^2 , \label{eq:5 - 58}\end{aligned}\ ] ] and @xmath270 \left| \sqrt{c_l^b c_{l'}^b } ( w^2)_{lm } \left ( \begin{array}{ccc } l & l ' & l \\ -2 & 2 & 0 \end{array } \right ) \right .
\nonumber \\ & & \mbox{}\left .
+ \frac{2\left(\sqrt{c_l^b c_{l'}^b}-\sqrt{c_l^e c_{l'}^e}\right ) } { \sqrt{l(l+1)l'(l'+1 ) } } \left [ |\eth w|^2_{lm } \left ( \begin{array}{ccc } l & l ' & l \\ -1 & 1 & 0 \end{array } \right ) + { { \mathcal{e}}}_{lm } \left ( \begin{array}{ccc } l & l ' & l \\ -1 & -1 & 2 \end{array } \right ) \right ] \right|^2 \biggr\ } \nonumber \\ & & \mbox { } + \frac{1}{\pi } \left [ \frac{\sqrt{c_l^b c_{l'}^b}-\sqrt{c_l^e c_{l'}^e}}{\sqrt{l(l+1)l'(l'+1 ) } } \right]^2 \sum_{lm } [ 1-(-1)^{k } ] |{{\mathcal{b}}}_{lm}|^2 \left ( \begin{array}{ccc } l & l ' & l \\ -1 & -1 & 2 \end{array } \right)^2 , \label{eq:5 - 59}\end{aligned}\ ] ] where we have introduced parity eigenstates @xmath271 and @xmath272 for the multipoles of the spin-@xmath273 objects @xmath274 ^ 2 $ ] and @xmath275 ^ 2 $ ] , which are defined by , @xmath276 and @xmath277 . ] @xmath278 note that the @xmath272 vanish for an azimuthally - symmetric weight function since then @xmath279 .
equations ( [ eq:5 - 58 ] ) and ( [ eq:5 - 59 ] ) achieve our goal of finding approximations to the pseudo-@xmath0 covariance matrices that are efficient to compute , but do take account of @xmath1-@xmath2 mixing effects at leading order for @xmath13 , @xmath258 .
note that , whereas the pseudo-@xmath0 covariance for temperature anisotropies involves only the power spectrum of @xmath230 ( e.g. @xcite ) , the polarization covariances also include the auto and cross - power ( with @xmath230 ) of the square of the gradient of @xmath239 .
we can make a useful check on our approximation for @xmath217_{(lm)(lm)'}$ ] , equation ( [ eq:5 - 56 ] ) , by setting @xmath61 and @xmath280 and summing on @xmath281 .
inspection of equation ( [ eq:5 - 35 ] ) shows that @xmath282_{(lm)(lm ) } = \sum_{l ' } m_{ll'},\ ] ] whereas evaluating the left - hand side by summing the approximation for @xmath262 directly gives @xmath282_{(lm)(lm ) } \approx \frac{1}{\sqrt{\pi } l(l+1)}(|\eth w|^2)_{00 } = \frac{1}{2\pi } \int \frac{|\nabla w|^2}{l(l+1 ) } \ , { { \rmn{d}}}{\hat{\bmath{n}}}.\ ] ] reassuringly , this agrees with the leading - order result for @xmath83 given in equation ( [ eq:15 ] ) .
finally we consider the covariance of @xmath74 and @xmath283 , equation ( [ eq:5 - 48 ] ) .
this involves evaluating the matrix commutator @xmath284 $ ] . following the procedure that led to equation ( [ eq:5 - 53 ] ) for @xmath136 , we can express @xmath134 as @xmath285 + \left[\eth w({\hat{\bmath{n } } } ) \eth { } _ { 2}y_{lm}^*({\hat{\bmath{n } } } ) + { \bar{\eth}}w({\hat{\bmath{n } } } ) { \bar{\eth}}{}_{-2}y_{lm}^*({\hat{\bmath{n}}})\right ] \right .
\nonumber \\ & & \mbox{}\left .
+ \sqrt{\frac{(l+2)!}{(l-2)!}}w({\hat{\bmath{n}}})y_{lm}^*({\hat{\bmath{n } } } ) \right\ } y_{(lm)'}({\hat{\bmath{n } } } ) .
\label{eq:5 - 61}\end{aligned}\ ] ] this is similar to equation ( [ eq:5 - 53 ] ) except for the presence of the last term in the curly brackets ( and the sign differences ) .
the last term will dominate the integral for @xmath13 , @xmath258 , and in this limit @xmath134 will be close to the equivalent coupling matrix for the temperature anisotropies ( see e.g. @xcite ) . forming the matrix product @xmath286 from equations ( [ eq:5 - 53 ] ) and ( [ eq:5 - 61 ] ) , and using the approximation in equation ( [ eq:5 - 55 ] ) , we find at leading order @xmath287 \left ( \begin{array}{ccc } l & l ' & l \\ 0 & 1 & -1 \end{array } \right ) \right ] , \label{eq:5 - 62}\end{aligned}\ ] ] which is @xmath288 as expected . in deriving equation ( [ eq:5 - 62 ] )
we have used @xmath289 .
noting that @xmath290 is the hermitian conjugate of @xmath291 , we have @xmath292 \left ( \begin{array}{ccc } l & l ' & l \\ 1 & 0 & -1 \end{array } \right ) \right ] , \label{eq:5 - 63}\end{aligned}\ ] ] where we have used the symmetries of the @xmath17 symbols and the reality of @xmath230 . differencing the previous two equations gives @xmath293 \left [ \frac{1}{\sqrt{l'(l'+1 ) } } \left ( \begin{array}{ccc } l & l ' & l \\ 1 & 0 & -1 \end{array } \right ) + \frac{1}{\sqrt{l(l+1 ) } } \left ( \begin{array}{ccc } l & l ' & l \\ 0 & 1 & -1 \end{array } \right ) \right ] .
\label{eq:5 - 64}\end{aligned}\ ] ] this can be simplified further by noting that @xmath294\left [ \frac{1}{\sqrt{l'(l'+1 ) } } \left ( \begin{array}{ccc } l & l ' & l \\ 1 & 0 & -1 \end{array } \right ) + \frac{1}{\sqrt{l(l+1 ) } } \left ( \begin{array}{ccc } l & l ' & l \\ 0 & 1 & -1 \end{array } \right ) \right ] = \frac{[1-(-1)^k]}{\sqrt{l'(l'+1 ) } } \left ( \begin{array}{ccc } l & l ' & l \\ 0 & 1 & -1 \end{array } \right ) \left ( 1-\frac{l'(l'+1)}{l(l+1 ) } \right),\ ] ] which follows from a recursion relation for the @xmath17 symbols [ equation ( 4 ) of section 8.6.2 of @xcite ] . since @xmath295 , it follows that the commutator @xmath296_{(lm)(lm)'}$ ] is @xmath297 , and so we can safely neglect its contribution compared to that from @xmath298 in equation ( [ eq:5 - 48 ] ) at high @xmath13 . ( the power spectrum prefactors are approximately equal for both terms . ) in this limit , @xmath122 is symmetric .
therefore our final approximation for the covariance of @xmath1 and @xmath2 is , after summing over @xmath281 and @xmath299 , @xmath300 \left ( \sqrt{c_l^e c_{l'}^e } + \sqrt{c_l^b c_{l'}^b}\right)^2 |(w^2)_{lm}|^2 \left ( \begin{array}{ccc } l & l ' & l \\ -2 & 2 & 0 \end{array } \right)^2 .
\label{eq:5 - 65}\ ] ] this , and our approximations for @xmath301 and @xmath127 , can be computed efficiently . for a general weight function @xmath35
, we can compute the multipoles @xmath302 , @xmath271 and @xmath272 as follows .
first compute the spin-1 field , @xmath303 ; this can be done , for example , by synthesising @xmath304 with fast spherical transforms ( on an iso - latitude pixelisation ) .
next , form the spin - zero field @xmath305 and the spin-2 field @xmath306 in real space by taking the squared modulus and the square of @xmath303 respectively . the scalar multipoles @xmath302 can then be extracted from the spin - zero field with a ( scalar ) spherical transform , while @xmath271 and @xmath272 are just the electric and magnetic multipoles of the spin-2 field which can be extracted with spin-2 spherical transforms . at each @xmath13 and @xmath111 ,
the @xmath17 symbols for all the @xmath124 values required can be computed efficiently with stable recursion relations @xcite . in this subsection
we test the approximate covariance matrices developed above against the exact matrices . since computing the latter on intermediate and small scales is prohibitively slow for general weight functions , we restrict ourselves to the two azimuthally - symmetric examples considered in section [ subsec : ex1 ] .
of course , for any practical application , our approximate covariance matrices should be carefully verified against those obtained from high - volume monte - carlo simulations with the chosen survey geometry .
figure [ fig : exact_and_approxee_10_15_cosine ] shows the covariance matrix @xmath116 for the @xmath67-radius survey computed with the approximation ( [ eq:5 - 58 ] ) and with the exact expression , equation ( [ eq:5 - 37 ] ) .
again , we take the weighting to be uniform except for cosine - squared apodization of the outer @xmath69 annulus .
the approximation is very accurate for the covariance of @xmath74 , except on the largest scales , as expected given that our largest potential source of error the perturbative treatment of @xmath1-@xmath2 mixing is irrelevant in this case since @xmath75 . for @xmath307 ,
the error on the diagonal is better than 2% except around the acoustic trough in @xmath4 , at @xmath308 , where it is around @xmath309% .
quite generally , the error on the diagonal is largest at the acoustic peaks and troughs ; the approximation over - estimates at the peaks , where the curvature of @xmath4 is negative , under - estimates at the troughs , and is best in between where the curvature vanishes .
this is consistent with the errors that we would expect from removing @xmath4 from the convolution in equation ( [ eq:5 - 37 ] ) .
the magnitude of the correlations for @xmath74 fall to lower than 1% for @xmath310 . for @xmath13 and @xmath311 , the maximum error in those off - diagonal elements with correlations greater than 1% is @xmath312 , but the error is localised to a few elements at the extremes of the diagonal band around the acoustic trough at @xmath308 .
this is , however , a very stringent error criterion : to get a 1%-correlated element accurate to 30% with monte - carlo methods would require @xmath313 simulations .
a more relevant measure for the off - diagonal elements is the absolute error in the correlations , since large fractional errors in elements with small correlation are relatively harmless . with this measure
, the maximum error in the correlations for @xmath13 and @xmath311 is @xmath314 .
we have made similar comparisons for the case of the @xmath112 galactic cut described in section [ subsec : ex1 ] . in this case
, the error on the diagonal is better than 1% for @xmath315 .
the only off - diagonal elements with correlations @xmath316 are for @xmath317 and @xmath318 , and for all of these the fractional error is less than 2% .
the same comparisons are made for @xmath127 for the @xmath67-radius survey in fig .
[ fig : exact_and_approxbb_10_15_cosine ] .
the dominant contribution to the covariance matrix over the full range plotted is now from @xmath1-mode power that is mixed into @xmath37 ; see fig .
[ fig : exactbb_10_15_cosine_and_exactbb_10_15_cosine_bonly ] .
as expected , the approximate covariance matrix for @xmath57 is less accurate than for @xmath74 since any errors in our modelling of the mixing are important for the former except at very large @xmath13 .
the implied improvement in accuracy with increasing @xmath13 is apparent in the figure .
more quantitatively , for @xmath319 the error on the diagonal is better than 10% .
interestingly , the error follows the same trend as for @xmath74 , i.e. it is largest in magnitude at the acoustic peaks and troughs of @xmath4 , and vanishes at the inflection points .
this suggests that the main source of error here is from removing @xmath4 from the convolution in equation ( [ eq:5 - 38 ] ) rather than errors in approximating @xmath262 by equation ( [ eq:5 - 56 ] ) .
( since @xmath320 is broader than @xmath126 , the fractional error made in approximating the convolution is worse for the covariance of @xmath57 than for @xmath74 . ) regarding the off - diagonal elements , for @xmath13 and @xmath321 the error _ in the correlation _ is @xmath322 for all elements .
the correlation matrix is thus accurately determined except on the largest scales , and so , in a practical application , could be used as a template for the shape of the covariance matrix with the variances calibrated off simulations . this procedure is similar to that adopted by _
wmap _ for the analysis of the one - year temperature data @xcite .
the @xmath323 over which correlations are significant is scale - dependent for the covariance of @xmath57 unlike that for @xmath74 ; see fig .
[ fig : exact_corr_10_15_cosine ]
. the correlations are broadest at the peaks of @xmath4 , where the contribution of @xmath1 modes to the @xmath57 covariance is greatest .
this behaviour arises from the different widths of the @xmath47 matrices , and would be lost in any approximation that did not model the @xmath1-@xmath2 mixing accurately . repeating the comparison for the case of the galactic cut
, we find that the diagonal elements are accurate to better than 10% for @xmath324 , and better than 1% for @xmath319 . for the off - diagonal elements , the correlations are accurate to better than @xmath314 for @xmath13 and @xmath311 .
finally , we compare the exact and approximate @xmath325 for the @xmath67-radius survey in fig .
[ fig : exact_and_approxeb_10_15_cosine ] . the error on the diagonal is better than 10% for all @xmath326 , and , as above , arises mainly from removing @xmath4 ( which dominates ) from the convolution in equation ( [ eq:5 - 39 ] ) , rather than our ignoring the matrix commutator @xmath327 $ ] in equation ( [ eq:5 - 48 ] ) . for parameter estimation ,
the error in the correlation matrix , @xmath328 is the more relevant quantity . the exact correlation matrix is shown in fig . [ fig : exact_corr_10_15_cosine ] ; there are strong correlations on large scales due to the power in @xmath4 from reionization , but for @xmath13 and @xmath311 , the maximum correlation is only @xmath329 ( and occurs on the diagonal ) . forming the approximate covariance matrix with equations ( [ eq:5 - 58 ] ) , ( [ eq:5 - 59 ] ) and ( [ eq:5 - 65 ] ) , we find that the error is better than @xmath314 for all @xmath13 and @xmath311 . for the case of the galactic cut ,
the correlations are @xmath330 for @xmath13 and @xmath311 , and the maximum error in our approximate correlation matrix is @xmath331 over the same range .
to summarise , the dominant error in our approximate covariance matrices , for the weight functions tested here , is from the treatment of the power spectra in the convolutions in equations ( [ eq:5 - 37])([eq:5 - 39 ] ) .
the biggest errors on large and intermediate scales arise from the terms with the dominant @xmath4 convolved with the broad @xmath320 , and so are most critical for the variance of @xmath57 . in forming the correlation matrices
, the convolution error largely cancels and so our approximate correlation matrices are more accurate than the covariances . for small surveys
it may be necessary to re - calibrate our approximate variances off simulations .
accurate covariance matrices should then be achievable by combining our approximate correlations matrices with the calibrated variances .
in this subsection we use our approximate expressions for the covariances of the pseudo-@xmath0s , equations ( [ eq:5 - 58 ] ) , ( [ eq:5 - 59 ] ) and ( [ eq:5 - 65 ] ) to derive useful ` rules of thumb ' for the variances of quasi - uncorrelated band - power estimates of the power spectra .
our results generalise equations ( [ eq : flat50 ] ) and ( [ eq : flat51 ] ) , derived earlier for gaussian weighting in the flat - sky limit , and extend the temperature result of @xcite , given here as equation ( [ eq : hivon ] ) , to polarization .
following our treatment of gaussian weighting in section [ subsec : gaussian ] , we approximate the inversion from pseudo-@xmath0s to estimated @xmath0s by the local transformation @xmath332 where we have defined @xmath333 and @xmath334 . recall that e.g. @xmath335 is the contribution of @xmath1 modes to the mean of @xmath74 , and @xmath336 is the contribution from @xmath2 modes . the normalisation @xmath337 .
the covariance of the @xmath338s is thus a scaled version of that for the pseudo-@xmath0s . as in the flat - sky limit , we consider only the two extreme cases , @xmath128 and @xmath183 , to keep the algebra manageable .
such limits are sufficient to get the dominant contribution to the variance of @xmath339 , and to assess the importance of @xmath1-@xmath2 mixing for the variance of @xmath210 .
we only need perform the calculation for @xmath340 since the other case then follows by symmetry .
we obtain quasi - uncorrelated estimates by averaging the recovered @xmath0s in bands of width @xmath341 , so we begin by averaging the pseudo-@xmath0 covariances in such bands .
consider first the variance of the binned @xmath74 for @xmath342 .
the leading - order contribution is from the term involving @xmath234 in equation ( [ eq:5 - 58 ] ) . to perform the summation over @xmath111 in a band centred on @xmath343
, we first insert a factor @xmath344 , which is effectively unity for @xmath111 in the band provided that @xmath345 .
so long as we further choose bands that are broad compared to @xmath11 , we can reduce the lower limit of the summation over @xmath111 to @xmath346 and increase the upper limit to @xmath347 .
( recall that @xmath11 sets the range over which the pseudo-@xmath0s are significantly correlated . ) finally , if we approximate @xmath348 as constant over a band we can replace it by @xmath349 , and the summation over the square of the @xmath17 symbol can be done in the same way as in the derivation of equations ( [ eq:5 - 35a ] ) and ( [ eq:5 - 35 ] ) .
we find that , at leading - order , @xmath350 to compute the variance of the binned @xmath74 , we further average over @xmath13 in the same band , treating @xmath4 as constant , to find @xmath351 this result has the same form as that for the temperature anisotropies . for the variance of the binned @xmath352 we must retain all terms in equation ( [ eq:5 - 59 ] ) that survive setting @xmath183 .
we perform the summation over @xmath111 in a similar manner to that described above , making use of some additional results for summing products of @xmath17 symbols that are given in appendix [ app : useful_sums ] .
further averaging @xmath13 within the band , we find , at leading order , @xmath353 this can be simplified further by noting that @xmath354 and , from equation ( [ eq:5 - 60 ] ) and the completeness of the spin - weight harmonics , @xmath355 = \int |\eth w|^4 \ , { { \rmn{d}}}{\hat{\bmath{n}}}. \label{eq : rot5}\ ] ] putting these results together , we find @xmath356 note that in the limit @xmath183 , the variance of @xmath57 arises solely from @xmath1-@xmath2 mixing which explains why only derivatives of @xmath239 appear in this equation .
the final term we require is @xmath357 . setting @xmath183 in equation ( [ eq:5 - 65 ] ) and using similar approximations to those adopted above
, the leading - order result is @xmath358 as expected , the relative sizes of @xmath359 , @xmath360 and @xmath361 for @xmath183 are roughly @xmath362 , @xmath363 and @xmath364 respectively .
we now have all the results required to compute the leading - order variances of the band - power estimates for @xmath4 and @xmath5 . combining the ( co)variances calculated above with equation ( [ eq : rot0 ] ) , we find @xmath365 here , we have introduced the convenient notation @xmath366 , where @xmath367 , @xmath368 or @xmath369 in the above , and approximated @xmath370 and @xmath371 $ ] [ see equations ( [ eq:5 - 35a ] ) and ( [ eq:15 ] ) ] . to obtain the equivalent results for @xmath128
, one should swap @xmath1 and @xmath2 everywhere in equations ( [ eq : rot8 ] ) and ( [ eq : rot9 ] ) .
it is straightforward to verify that for a gaussian weight we recover the variances given in section [ subsec : gaussian ] in the flat - sky limit .
we now see , quite generally , that the ratio of the excess sample variance from @xmath1 modes in band - power estimates of @xmath5 to the variance from @xmath2 modes is @xmath372 .
we consider the implications of this for @xmath2-mode detection with pseudo-@xmath0 techniques in the next section .
the realisation that , in linear theory , scalar perturbations do not generate @xmath2-mode polarization @xcite has paved the way for a future generation of instruments designed for high - sensitivity searches for gravitational waves via @xmath2-mode polarization .
instrument noise can be reduced to the @xmath373-arcmin level by surveying e.g.a few - hundred square degrees for around a year with arrays of hundreds of detectors . with instrument noise at this level , the sample variance from the @xmath2 modes produced by gravitational lensing of the scalar @xmath1 modes dominates the thermal noise in the random error budget for @xmath5 , assuming perfect isolation of @xmath2 modes .
( we are assuming the tensor - to - scalar ratio @xmath374 here , so that sample variance of the linear @xmath2 modes is sub - dominant . )
if we make no attempt to reconstruct and subtract the lensing signal , its variance sets a fundamental limit to the one - sigma error of @xmath375 for a full - sky survey in the null hypothesis , @xmath376 .
this should be compared with the limit of only @xmath377 obtainable by analysing temperature and electric polarization only .
the limit on @xmath16 is much poorer in this case because of the large sample variance of the dominant scalar perturbations .
if pseudo-@xmath0 techniques are to be applied to future , high - sensitivity searches for gravitational waves , it is clearly desirable that the imperfect isolation of @xmath2 modes inherent in such methods should not compromise sensitivity to @xmath16 . in fig .
[ fig : errors_on_recovered_b ] we plot the errors on flat band - power estimates of @xmath5 ( with @xmath378 ) obtained by using the unbiased estimator of @xcite ; see also the discussion in section [ sec : inverse ] .
the weight function applied to the @xmath67-radius survey , and apodization of the correlation function , are the same as in fig .
[ fig : power_and_window ] .
the errors are obtained from the exact covariance matrices for the pseudo-@xmath0s by transforming ( two - sidedly ) with the appropriate pseudo - inverses @xmath93 and @xmath94 .
these pseudo - inverse matrices are shown in fig .
[ fig : power_and_window ] .
the errors are calculated for @xmath376 in two ways .
first , we set @xmath128 so that we obtain the error in the recovered @xmath5 due to sample variance of the lensing @xmath2 modes .
these errors are representative of what would be achieved if we were able to separate @xmath1 and @xmath2 modes perfectly , and without loss , on all scales in the maps ; they agree very well with the rule of thumb in equation ( [ eq : rot8 ] ) if we swap the roles of @xmath1 and @xmath2 there .
we also calculate the band - power errors obtained by setting @xmath183 , i.e. removing the @xmath2-mode power from lensing ; they agree reasonably with the rule of thumb in equation ( [ eq : rot9 ] ) .
these errors arise because , in any realisation , our chosen estimator @xmath210 receives a contribution from both @xmath1 and @xmath2 modes , despite only coupling to @xmath2-mode power in the mean .
such errors can only be removed by isolating @xmath2 modes at the level of the map , or by adopting more optimal , non - local weighting of the data , as in maximum - likelihood power spectrum estimation .
critically , we see from fig . [
fig : errors_on_recovered_b ] that the errors due to the sample variance of the @xmath1 modes that leak into the estimator dominate the sample variance from lensing @xmath2 modes on large scales , where the gravitational - wave signal resides .
we conclude from the figure that pseudo-@xmath0 estimators will be highly non - optimal for future experiments with thermal sensitivity better than @xmath379-arcmin , targeting @xmath2-mode polarization over a few - hundred square degrees .-
mode polarization with a power spectrum of known shape but ( low ) unknown amplitude is a circle of radius around 10 degrees @xcite .
this assumes that the sample variance of the @xmath2 modes from gravitational waves does not dominate the error budget .
if the lens - induced @xmath2 modes are treated as an additional ( gaussian ) noise term , the optimal survey area is increased since sample variance becomes an issue again .
the best compromise is roughly to balance the sample variance of the lens - induced @xmath2 modes with the thermal - noise variance . as a specific example , for 200 _ pairs _ of polarization - sensitive bolometers , with each bolometer having sensitivity @xmath380 , and one year of integration , a survey covering two per cent of the sky balances the thermal and sample variance . ]
we can make a quantitative estimate of the impact of the additional sample variance due to leaked @xmath1 modes on the detectability of gravitational waves as follows . to preserve all information in the pseudo-@xmath0s , we bypass the step of forming unbiased estimates of @xmath5 [ which involves the singular transformation in equation ( [ eq : estimates ] ) ] , and instead estimate @xmath16 directly from the pseudo-@xmath0s .
we concatenate @xmath74 and @xmath57 into a single data vector @xmath14 , whose covariance matrix @xmath381 then consists of blocks formed from @xmath116 , @xmath127 and @xmath122 .
we assume that all parameters except @xmath16 are fixed , and calculate its one - sigma fisher error from @xmath382 in practice , we invert the covariance matrix with a singular - value decomposition which projects out any linear dependencies in the @xmath14 due to the finite sky coverage . for the @xmath67-radius survey , the one - sigma error on @xmath16 in the null hypothesis ( @xmath376 )
is @xmath383 .
our sensitivity to @xmath16 improves dramatically to @xmath384 if we artificially set @xmath128 ( and @xmath385 ) in the analysis .
this latter case is representative of what could be achieved with lossless separation of the @xmath2 modes , followed by a pseudo-@xmath0 analysis of the stokes maps of these @xmath2 modes . in practice , @xmath1-@xmath2 separation is lossy , particularly on the largest scales where there is significant @xmath2-mode power from reionization @xcite .
if we assume , crudely , that large - scale modes with @xmath386 are lost to the separation , we find a @xmath387 for the @xmath67-radius survey .
the effect of not isolating the @xmath2 modes properly in the pseudo-@xmath0 analysis thus degrades the limit on @xmath16 by a factor @xmath388 .
accurate calculation of the covariance matrix is an important part of power spectrum estimation during the analysis of cmb data .
a ` correct ' estimator ( i.e. unbiased ) , but with poorly quantified errors , is useless for subsequent parameter estimation . in this paper
we have constructed the contribution from sample variance to the covariance matrices of heuristically - weighted quadratic estimates of the cmb polarization power spectra .
exact calculation of these matrices for general weighting is numerically feasible at large scales ( @xmath13 less than a hundred or so ) , but the @xmath109 scaling is prohibitive on small scales . to resolve this problem we developed analytic approximations to the polarization covariance matrices , including the effects of mode coupling and , in a perturbative manner , the mixing of @xmath1 and @xmath2 modes that inevitably arises with heuristic weighting .
our approximations , which generalise earlier work on the temperature anisotropies @xcite , can be computed efficiently to high @xmath13 .
accurate covariance matrices can be obtained on all scales by replacing the low-@xmath13 sectors of our approximations with a numerical calculation .
we tested the approximations against numerical results for two toy - model surveys for which azimuthal symmetry allows efficient evaluation of the exact covariances .
for the more challenging survey , a @xmath67-radius survey with cosine - squared apodization of the outer @xmath69 , the most significant errors were on the variance of @xmath57 .
these errors arise from our local treatment of the @xmath1-mode power spectrum when it appears in a convolution with the broad kernel that describes the mixing of @xmath1 and @xmath2 modes in equation ( [ eq:5 - 38 ] ) . for the variance of @xmath74 ( and the temperature anisotropies )
the approximation is more accurate because the dominant convolution involves a kernel that is intrinsically more narrow .
it will probably be very difficult to improve the error on the variance of @xmath57 in an analytic treatment .
fortunately , convolution errors tend to cancel in the approximation of the _ correlation _ matrices .
we observed maximum errors @xmath389 in the correlations , except on the largest scales , suggesting that , in a practical application , it should be possible to construct accurate covariance matrices by combining the approximate correlations with variances calibrated off simulations . by binning our approximate covariances
, we also derived useful rules of thumb for the variances of band - power estimates of @xmath4 and @xmath5 .
these are particularly simple to evaluate , involving only integrals of products of the weight function and its derivatives .
a further aim of this paper was to quantify the conditions under which pseudo-@xmath0 methods can be applied to gravitational - wave searches via @xmath2-mode polarization .
measurements of the @xmath2-mode power spectrum from gravitational waves with future surveys with sub-@xmath390-arcmin thermal noise are potentially limited by the sample variance of lens - induced @xmath2 modes .
we showed that for surveys covering one or two per cent of the sky , the additional variance due to leaked @xmath1 modes dominates the lensing noise .
adopting pseudo-@xmath0 methods would therefore by highly sub - optimal : the error on the tensor - to - scalar ratio would be @xmath388 times higher than if @xmath2 modes had been isolated before forming quadratic estimates .
ultimately , this additional variance would limit the tensor - to - scalar ratio that is detectable with pseudo-@xmath0 methods to @xmath3 .
map - level @xmath1-@xmath2 separation , or more optimal weighting such as in maximum - likelihood power spectrum estimation , will therefore be essential to ensure the maximum returns from future @xmath2-mode surveys .
adc acknowledges a royal society university research fellowship .
gc acknowledges travel support from queens college , cambridge and the cavendish laboratory .
all theoretical power spectra in this paper were computed with camb @xcite .
some of the results in this paper have been derived using the lapack package for matrix decompositions .
barkats d. , et al .
, 2005 , apj , 619 , l127 bartolo n. , komatsu e. , matarrese s. , rioto a. , 2004 , phys . rep .
, 402 , 103 bond j. r. , jaffe a. h. , knox l. , 1998 , phys .
d , 57 , 2117 bond j. r. , jaffe a. h. , knox l. , 2000 , apj , 533 , 19 bunn e. f. , zaldarriaga m. , tegmark m. , de oliveira - costa a. , 2003 , phys . rev .
d , 67 , 3501 chon g. , 2003 , phd thesis , university of cambridge chon g. , challinor a. , prunet s. , hivon e. , szapudi i. , 2004 , mnras , 350 , 914 crittenden r. g. , natarajan p. , pen u .-
l . , theuns t. , 2002 , apj , 568 , 20 efstathiou g. , 2004 , mnras , 349 , 603 grski k. m. , 1994 , apj , 430 , l85 grski k. m. , hivon e. , banday a. j. , wandelt b. d. , hansen f. k. , reinecke m. , bartelman m. , 2004 , preprint ( astro - ph/0409513 ) hansen f. k. , grski k. m. , 2003 , mnras , 343 , 559 hansen f. k. , grski k. m. , hivon e. , 2002 , mnras , 336 , 1304 hinshaw g. , et al . , 2003 ,
apjs , 148 , 135 hivon e. , grski k. m. , netterfield c. b. , crill b. p. , prunet s. , hansen f. , 2002 , apj , 567 , 2 jaffe a. h. , kamionkowski m. , wang l. , 2000 , phys .
d , 61 , 083501 kamionkowski m. , kosowsky a. , stebbins a. , 1997 , phys .
, 78 , 2058 kesden m. , cooray a. , kamionkowski m. , 2002 , phys .
lett . , 89 , 011304 knox l. , song y .- s . , 2002 , phys . rev
, 89 , 1303 kogut a. et al . , 2003 , apjs , 148 , 161 kovac j. , leitch e. m. , pryke c. , carlstrom j. e. , halverson n. w. , holzapfel w. l. , 2002 , nature , 420 , 772 leitch e. m. , kovac j. m. , halverson n. w. , carlstrom j. e. , pryke c. , smith m. w. e. , 2004 , preprint ( astro - ph/0409357 ) lewis a. , 2003 , phys . rev .
d , 68 , 3509 lewis a. , challinor a. , lasenby a. , 2000 , apj , 538 , 473 lewis a. , challinor a. , turok n. , 2002 , phys . rev .
d , 65 , 023505 martin j. , schwarz d. , 2000 , phys .
d , 62 , 103520 mollerach s. , harari d. , matarrese s. , 2004 , phys .
d , 69 , 063002 netterfield c. b. , et al .
, 2002 , apj , 571 , 604 newman e. , penrose r. , 1966 , j. math .
phys . , 863 ponthieu n. , et al . , 2005 ,
preprint ( astro - ph/0501427 ) readhead a. c. s. , et al .
, 2004 , science , 306 , 836 schulten k. , gordon r. g. , 1976 , comput . phys .
commun . , 11 , 269 seljak u. , zaldarriaga m. , 1997 , phys . rev
. lett . , 78 , 2054 seljak u. , hirata c. m. , 2004 , phys . rev .
d , 69 , 3005 smith k. m. , hu w. , kaplinghat , m. , 2004 , phys .
d , 70 , 043002 spergel d. , et al .
, 2003 , apjs , 148 , 175 szapudi i. , prunet s. , colombi s. , 2001 , apj , 561 , l11 tegmark m. , de oliveira - costa a. , 2002 , phys .
d , 64 , 063001 varshalovich d. a. , moskalev a. n. , khersonskii v. k. , 1988 , quantum theory of angular momentum .
world scientific , singapore wandelt b. d. , hivon e. , grski k. m. , 2001 , phys .
d , 64 , 083003 zaldarriaga m. , seljak u. , 1998 , phys .
d , 58 , 023003
in this appendix we derive approximate analytic expressions for the correlations between the flat - sky fourier transforms @xmath170 and @xmath171 for a gaussian weight function @xmath154 .
our expressions are valid for @xmath13 and @xmath111 large compared to @xmath391 . in this limit
, the power spectra can be approximated as being constant over the support of @xmath392 ; see section [ sec : approx - cov ]
. using equations ( [ eq : flat6 ] ) and ( [ eq : flat7 ] ) we can approximate the correlators by @xmath393 for a gaussian weight function we can evaluate the integrals in these equations analytically ; we find @xmath394 .
\label{eq : flat20}\end{aligned}\ ] ] note that @xmath395 = \frac{1}{2 } \sigma^2 \cos2(\phi_{{\bmath{l}}}-\phi_{{{\bmath{l}}}'})e^{-|{{\bmath{l}}}-{{\bmath{l}}}'|^2\sigma^2/4 } , \label{eq : flat21}\ ] ] and the right - hand side is @xmath396 constructed from @xmath397 .
this result follows quite generally from equation ( [ eq : flat8 ] ) . to calculate the covariance of the errors of the power spectrum estimates we also need the following correlator @xmath398 the integrals in this expression evaluate to @xmath399 .
\label{eq : flat23}\end{aligned}\ ] ] we now have @xmath400 = - \frac{i}{2 } \sigma^2 \sin2(\phi_{{\bmath{l}}}-\phi_{{{\bmath{l}}}'})e^{-|{{\bmath{l}}}-{{\bmath{l}}}'|^2\sigma^2/4 } , \label{eq : flat24}\ ] ] and the right - hand side is @xmath401 constructed from @xmath397 .
note that @xmath402 if @xmath147 and @xmath180 are parallel or perpendicular .
in sections [ sec : properties ] and [ subsec : rot ] we make use of a number of results for summing products of @xmath17 symbols . some of these follow trivially from the orthogonality properties of the @xmath17s , but others require more effort . here , we sketch our derivation of these latter results .
we make extensive use of well - known properties of the @xmath17 symbols , such as their symmetries and integral representations ( e.g. @xcite ) . in section [ sec : properties ] we require @xmath403 it is the presence of the factor of @xmath404 , where @xmath235 , that complicates the evaluation of this summation .
to proceed , we express the product of @xmath17 symbols as an integral of reduced wigner matrices using @xmath405 so that @xmath406 the factor @xmath404 has been absorbed in reducing the product of @xmath17 symbols to a square .
the summation over @xmath111 can be performed by first expressing @xmath407 in terms of @xmath408 using the recursion relations for the reduced wigner matrices @xcite , and then using the completeness relation .
full details can be found in @xcite , but the final result is @xmath409 inserting this into equation ( [ eq : appb2 ] ) gives @xmath410 using the differential representation of the reduced wigner matrices ( see section 4.3.2 of @xcite ) , it can be shown that @xmath411 is a polynomial in @xmath261 of degree @xmath412 .
it can thus be expanded as the series @xmath413 , i.e. an expansion in legendre polynomials . using the orthogonality of the @xmath414
, we see that the integral in equation ( [ eq : appb4 ] ) vanishes if @xmath415 .
it can be performed for @xmath416 by replacing @xmath417 by its differential representation and repeatedly integrating by parts .
only boundary terms then survive , and evaluating these establishes the final result @xcite : @xmath418 in section [ subsec : rot ] , we require the evaluation of @xmath419 this follows along similar lines to @xmath420 above .
a useful intermediate result is @xmath421 from which it follows that @xmath422 the summation @xmath423 thus reduces to @xmath424 which , again , vanishes if @xmath415 . inserting the differential representation of @xmath425 and integrating by parts , we find @xmath426 a further result we require in section [ subsec : rot ] is @xmath427 using equations ( [ eq : appb1b ] ) and ( [ eq : appb8 ] ) , we find @xmath428 in this case , @xmath429 is a polynomial in @xmath261 of degree @xmath430 and has a root at @xmath431 .
it can thus be expanded as a series in @xmath432 with non - zero coefficients for @xmath433 , and so the integral in equation ( [ eq : appb12 ] ) vanishes for @xmath434 .
the summation @xmath435 arises when computing band - power variances from equations ( [ eq:5 - 58 ] ) and ( [ eq:5 - 59 ] ) ; it occurs with prefactors @xmath436 and @xmath437 .
since we are assuming the weight function is band - limited , @xmath438 in @xmath435 , so for @xmath10 the contribution from @xmath435 vanishes .
the final summation we require in section [ subsec : rot ] is @xmath439 which reduces to @xmath440 this time , @xmath441 is a polynomial in @xmath261 of degree @xmath430 that has a root at @xmath442 .
the summation @xmath443 can thus be expanded in the @xmath444 with @xmath445 , so that , like @xmath435 , it vanishes for @xmath446 .
the summation arises from the cross term between @xmath447 and @xmath448 in equations ( [ eq:5 - 58 ] ) and ( [ eq:5 - 59 ] ) , and so this contribution to the band - power variance can also be assumed to vanish for @xmath10 . | quadratic methods with heuristic weighting ( e.g. pseudo-@xmath0 or correlation function methods ) represent an efficient way to estimate power spectra of the cosmic microwave background ( cmb ) anisotropies and their polarization .
we construct the sample covariance properties of such estimators for cmb polarization , and develop semi - analytic techniques to approximate the pseudo-@xmath0 sample covariance matrices at high legendre multipoles , taking account of the geometric effects of mode coupling and the mixing between the electric ( @xmath1 ) and magnetic ( @xmath2 ) polarization that arise for observations covering only part of the sky .
the @xmath1-@xmath2 mixing ultimately limits the applicability of heuristically - weighted quadratic methods to searches for the gravitational - wave signal in the large - angle @xmath2-mode polarization , even for methods that can recover ( exactly ) unbiased estimates of the @xmath2-mode power .
we show that for surveys covering one or two per cent of the sky , the contribution of @xmath1-mode power to the covariance of the recovered @xmath2-mode power spectrum typically limits the tensor - to - scalar ratio that can be probed with such methods to @xmath3 . cosmic microwave background
methods : analytical : methods : numerical . |
recent years have seen an enormous enhancement in the understanding of neutrino - oscillation parameters in accelerator - based experiments .
these experiments are however confronted with a number of problems .
these are related to the large systematic uncertainties associated with the neutrino - nucleus signal in the detector .
major issues arise from the fact that the neutrino energy - flux in experiments is distributed over a wide range of energies from very low to a few gev .
hence a number of nuclear effects over a broad kinematical range ( from low - energy nuclear excitations to multinucleon emission ) simultaneously come into play .
the simulation codes used in the analysis of the experimental results are predominantly based on relativistic fermi gas ( rfg ) models .
rfg can describe the quasielastic ( qe ) cross section sufficiently accurate for medium momentum ( @xmath2 500 mev / c ) transfer reactions , but its description becomes poor for low momentum ( @xmath3 300 mev / c ) transfer processes , where nuclear effects are prominent . for the broad neutrino energy - flux used in the experiments , more realistic models are required . in this work
, we present a self - consistent continuum random phase approximation ( crpa ) approach to calculate qe electron and neutrino - scattering cross - sections off the nucleus .
this formalism was used to describe exclusive photo - induced and electron - induced qe scattering @xcite , inclusive neutrino scattering at supernova energies @xcite and charged - current quasielastic ( ccqe ) antineutrino scattering at intermediate energies @xcite .
we will briefly describe the essence of our model , for an updated version of the formalism we refer the reader to ref .
@xcite . the main update in ref .
@xcite from ref .
@xcite , are the inclusion of relativistic corrections and a suppression of the rpa quenching at high @xmath4 .
we start with a mean - field ( mf ) description of the nucleus where we solve the hartree - fock ( hf ) equations with a skyrme ( ske2 ) two - body interaction @xcite to obtain the mf potential .
we obtain the continuum wave functions by integrating the positive energy schrdinger equation with appropriate boundary conditions , hence taking into account final - state interactions in this manner .
long - range correlations are implemented by means of a crpa approach based on a green s function formalism .
the polarization propagator is approximated by iteration of its first - order contribution . in this way
, the formalism takes into account one - particle one - hole excitations out of the correlated nuclear ground state . within the rpa
an excited nuclear state is represented as the coherent superposition of the particle - hole ( @xmath5 ) and hole - particle ( @xmath6 ) excitations out of a correlated ground state @xmath7~,\ ] ] where @xmath8 denotes all quantum numbers identifying an accessible channel .
the rpa polarization propagator can be written as @xmath9 where @xmath10 is the excitation energy of the nucleus and @xmath11 is a short - hand notation for the combination of spatial , spin and isospin coordinates .
the @xmath12 corresponds to the mf contribution and @xmath13 is the antisymmetrized nucleon - nucleon interaction .
we used the modified effective momentum approximation ( mema ) @xcite , in order to take into account the influence of the nuclear coulomb field on the ejected lepton . in order to prevent the ske2 force from becoming unrealistically strong at high virtuality @xmath14
, we introduce a dipole hadronic form factor at the nucleon - nucleon interaction vertices @xcite .
further , we have implemented relativistic kinematic corrections @xcite in an effective manner .
we first test the reliability of the formalism by confronting ( @xmath0 ) scattering cross sections with the data of refs .
thereby , we present updated results of flux - folded charged - current quasielastic ( ccqe ) antineutrino scattering off @xmath1c and compare them with the miniboone measurements @xcite .
further , we discuss the contribution of neutrino - induced low - energy nuclear excitations in the signal of the accelerator - based neutrino oscillation experiments .
c(@xmath0 ) plotted as a function of excitation energy @xmath15 .
the incident electron energy @xmath16 and lepton scattering angle @xmath17 are listed on top of each panel .
solid lines are crpa cross sections and dashed - lines are hf cross sections .
experimental data is taken from refs .
@xcite.,title="fig : " ] c(@xmath0 ) plotted as a function of excitation energy @xmath15 . the incident electron energy @xmath16 and lepton scattering angle @xmath17
are listed on top of each panel .
solid lines are crpa cross sections and dashed - lines are hf cross sections .
experimental data is taken from refs .
@xcite.,title="fig : " ]
we start this section by showing some examples of electron - scattering results . in fig . [ fig_e_scattering ]
, we show our prediction of qe @xmath1c(@xmath0 ) scattering cross - sections and compare them with the measurements of the refs .
our predictions successfully describe the data over the broad kinematical range considered here .
the formalism successfully describes low - energy excitations ( panel ( a ) and ( b ) ) below the qe peak .
the forward scattering cross sections , even for higher incoming electron energies , are dominated by the qe contribution .
however , the data include cross section contributions beyond the qe channel , like @xmath18-excitations and other inelastic channels .
our calculations are intended to predict only the qe behavior .
a detailed comparison of ( @xmath0 ) cross section on @xmath1c , @xmath19o and @xmath20ca is performed in ref .
@xcite . an overall successful description of qe ( @xmath0 ) cross section data and especially low - energy excitations , validates the reliability of our formalism .
c@xmath21 plotted as a function of the direction of the scattering lepton @xmath22 for different values of its kinetic energy @xmath23 values ( top ) , as a function @xmath23 for different ranges of @xmath22 ( bottom ) .
solid lines are crpa and dashed lines are hf calculations .
miniboone data @xcite are filled squares , error bars represent the shape uncertainties and error boxes represent the 17.2% normalization uncertainty . ]
we show the double - differential cross section for @xmath1c@xmath21 , folded with the miniboone antineutrino flux @xcite , in fig .
[ fig_miniboone_anu_ddxs ] .
the top panels show the cross section in @xmath23 bins and the bottom panels show the cross section in @xmath22 bins .
the cross section is integrated over the corresponding bin width .
we adopt an axial mass value of @xmath24 = 1.03 gev , in the dipole axial form factor .
hf and crpa cross sections are compared with the miniboone measurements of ref .
miniboone data is presented with both shape and normalization uncertainties .
overall , crpa and hf calculations successfully reproduce the gross features of the measured cross section .
the predictions tend to underestimate the data .
it has been suggested in refs .
@xcite that the inclusion of multinucleon contributions , which are not included in our calculations , are essential for a more complete reproduction of the data .
c@xmath25 as a function of @xmath22 .
the miniboone data @xcite are integrated over @xmath23 . ]
c@xmath25 as a function of @xmath26 , compared with miniboone data @xcite . ] the flux - folded differential cross section as a function of @xmath22 , is shown in fig .
[ fig_miniboone_anu_dxs ] . for comparison with data ,
we integrate miniboone data over @xmath23 .
it is interesting to note that in the very forward direction the crpa results are larger than the hf ones .
this is due to the collective giant resonance contributions which are absent in the hf approximation but appear in crpa results that include long - range correlations . in fig .
[ fig_miniboone_anu_xs ] , we present total cross sections per target proton as a function of neutrino energy and compare them with the experimental data .
unlike double - differential cross sections , this quantity is model dependent .
the theoretical calculations are function of a true antineutrino energy while the experimental data are function of reconstructed antineutrino energy .
up to @xmath27 = 0.4 gev , the hf results essentially coincide with the crpa ones .
this is due to a compensation between a reduction in the qe region and an enhancement in the giant resonance part of the crpa results . for @xmath28 0.4 gev ,
the crpa results are slightly smaller than the hf ones .
c(@xmath29 ) plotted as a function @xmath23 , for different @xmath22 values . ] in order to illustrate the impact of the low - energy nuclear excitations , in fig .
[ fig_lowenergy ] , we show the double - differential cross - section for fixed neutrino energies and fixed scattering angles . as it appears in panel
( a ) , 150 mev energy neutrinos induce low - lying nuclear excitations at all scattering angles . for neutrino energies of 800 mev , which is near the mean energy of the miniboone @xcite and t2k @xcite fluxes , in panel ( b ) ,
the forward scatterings still show sizable low - energy excitations .
this feature can have a non - negligible contribution to the neutrino signals in these experiments , but can not be accounted for within the rfg - based simulation codes .
as already mentioned , one can observe in fig .
[ fig_miniboone_anu_dxs ] , that at very forward scatterings , @xmath30 1 , the crpa cross section generates more strength ( emerging from the low - lying excitation ) than the hf .
we have presented a continuum random phase approximation approach for quasielastic electron- and neutrino - nucleus scattering .
we validated the reliability of our formalism , in the quasielastic region , by comparing ( @xmath0 ) cross section with the available data .
an interesting feature of our crpa formalism is the successful prediction of low - energy nuclear excitations .
we calculated flux - folded @xmath1c@xmath21 cross sections and compared them with the miniboone antineutrino cross - section measurements .
crpa predictions are successful in describing the gross features of the cross section but seem to underestimate slightly the measured cross section .
we illustrated how low - energy nuclear excitations can possibly account for non - negligible contributions to the neutrino signal in accelerator - based neutrino - oscillation experiments .
this research was funded by the interuniversity attraction poles programme initiated by the belgian science policy office , the research foundation flanders ( fwo - flanders ) and by the erasmus mundus external cooperations window s eurindia project .
99 j. ryckebusch , m. waroquier , k. heyde , j. moreau , and d. ryckbosch , nucl .
a*476 * , 237 ( 1988 ) .
j. ryckebusch , k. heyde , d. van neck , and m. waroquier , nucl .
a*503 * , 694 ( 1989 ) .
n. jachowicz , k. heyde , j. ryckebusch , and s. rombouts , phys .
c*59 * , 3246 ( 1999 ) .
n. jachowicz , k. heyde , j. ryckebusch , and s. rombouts , phys . rev .
c*65 * , 025501 ( 2002 ) .
n. jachowicz , k. vantournhout , j. ryckebusch , and k. heyde , phys .
93 * , 082501 ( 2004 ) .
n. jachowicz , g. c. mclaughlin , phys .
lett . * 96 * , 172301 ( 2006 ) .
n. jachowicz , c. praet , and j. ryckebusch , acta physica polonica b , 40(9 ) , 2559 - 2564 ( 2009 ) .
n. jachowicz , and v. pandey , proceedings nuint12 , to be published .
v. pandey , n. jachowicz , j. ryckebusch , t. van cuyck , and w. cosyn , phys .
c*89 * , 024601 ( 2014 ) .
v. pandey , n. jachowicz , t. van cuyck , j. ryckebusch , and m. martini , arxiv:1412.4624 [ nucl - th ] .
m. waroquier , j. ryckebusch , j. moreau , k. heyde , n. blasi , s.y .
van de werf , and g. wenes , phys . rep . *
148 * , 249 ( 1987 ) .
j. engel , phys .
c*57 * , 2004 ( 1998 ) .
s. jeschonnek , t. w. donnelly , phys .
c*57 * , 2438 ( 1998 ) .
p. barreau _ et al_. , nucl .
a*402 * , 515 ( 1983 ) .
r. m. sealock _ et al_. , phys .
lett.*62 * , 1350 ( 1989 ) .
d. s. bagdasaryan _ et al_. , yerphi-1077 - 40 - 88 ( 1988 ) .
d. b. day _ et al_. , phys .
c * 48 * , 1849 ( 1993 ) .
a. a. aguilar - arevalo et al . , ( miniboone collaboration ) , phys . rev .
d*88 * , 032001 ( 2013 ) .
m. martini , m. ericson , g. chanfray , and j. marteau , phys .
c * 80 * , 065501 ( 2009 ) . m. martini , m. ericson , g. chanfray , and j. marteau , phys .
c * 81 * , 045502 ( 2010 ) .
j. nieves , i. ruiz simo , and m.j .
vicente vacas , phys .
c * 83 * , 045501 ( 2011 ) .
m. martini , and m. ericson , phys .
c * 87 * , 065501 ( 2013 ) .
j. nieves , i. ruiz simo , and m. vicente vacas , phys .
b * 721 * , 90 ( 2013 ) .
a. a. aguilar - arevalo et al . , ( miniboone collaboration ) , phys .
d*81 * , 092005 ( 2010 ) . k. abe et al . , ( t2k collaboration ) , phys .
d*87 * , 092003 ( 2013 ) . | we present a continuum random phase approximation approach to study electron- and neutrino - nucleus scattering cross sections , in the kinematic region where quasielastic scattering is the dominant process .
we show the validity of the formalism by confronting inclusive ( @xmath0 ) cross sections with the available data .
we calculate flux - folded cross sections for charged - current quasielastic antineutrino scattering off @xmath1c and compare them with the miniboone cross - section measurements .
we pay special emphasis to the contribution of low - energy nuclear excitations in the signal of accelerator - based neutrino - oscillation experiments . |
in spite of the prevalence and importance of omnivory food webs @xcite in natural communities their population dynamics to date remain poorly understood , even for only three species in the community . even in simple systems a plethora of nonlinear effects such as flexible consumer behaviour @xcite , intraspecific interactions between competing consumers and resources @xcite , inhomogeneity of the environment @xcite and adaptive foraging @xcite precludes easy theoretical treatment and interpretation .
one example of a non - trivial omnivory food web is a system with intraguild predation @xcite .
intraguild predation assumes that the same organism is both competitor and predator to another member of the food web .
the igp models encompass a rich dynamical behaviour including coexistence @xcite and alternative stable states @xcite .
simple mathematical models @xcite have been evoked in attempt to explain the persistence of igp interactions in natural habitats .
however predictions from the mathematical theory of 3-species igp systems state that a high resource carrying capacity promotes the exclusion of intermediate trophic levels and thus destabilizes interactions @xcite .
what is puzzling that various empirical studies of omnivory document however coexistence , but not exclusion , over the entire range of natural resource productivities @xcite . on the basis of experimental observations a theoretical 3-species omnivory model @xcite predicts the coexistence only at superior competitive abilities of the ig prey for the communal resource @xcite . yet
empirical data suggest a robust persistence of igp systems in both terrestrial @xcite and aquatic communities @xcite .
theoretical models that are focused on the aspects of stability and coexistence of species in 3-level systems with the igp @xcite , as a rule , largely reduce the complexity of interactions observed in realistic systems @xcite .
such oversimplifications can influence the population dynamics as well as critically impact species persistence .
even though the simplest model of the igp encompasses only three species @xcite a number of empirical studies deal with larger food webs that involve more than three species potentially engaged in igp interactions @xcite .
spatiotemporal heterogeneity of the environment often is invoked as one of the explanatory mechanisms for the coexistence between multiple species competing for the same resources @xcite .
it has been observed that such a spatiotemporal heterogeneity can affect the diversity in prey populations @xcite .
indeed an inhomogeneity in prey items that share common resource and predators is critical in determining the responses of ecological community . for systems with multiple prey composition various coexistence patterns can be found depending on the levels of resource productivity @xcite .
it is not clear yet how the diversity in a prey community will affect the behaviour in the igp systems .
the effect of a habitat structure on the igp is discussed in various recent models @xcite .
for example a stable coexistence of the intraguild prey due to inhomogeneity of a habitat can be supported by creating temporal refuges for prey and reducing the encounter rates among preys and predators @xcite .
in addition the stability of the igp can be enhanced by an inclusion of additional factors such as behaviourally mediated effects @xcite . to include the effect of an increasing diversity of resource and ig predators on population dynamics recently the 3-species igp model @xcite
was modified by holt and huxel ( 2007 ) .
the authors extended the basic 3-species omnivory model to the so called `` partial igp '' model in which `` partial '' overlap among competitors for a single resource exists and both predators have exclusive resources to exploit .
it was shown @xcite that an alternative resource enhances the tolerance of the ig prey against attacks from ig predators .
independently of a competitive status of the ig prey in exploitation for a shared resource it can persists by utilizing an alternative resource . an extended formulation of the igp model with trophic supplementation has been proposed by daugherty et al .
the authors investigated three forms of a supplementary feeding outside of the basic igp module and postulated a higher potential for persistence of the ig prey due to its efficient exploitation of external resources .
there is growing evidence that in many systems the ig prey has a mutualistic or at least facilitative relationship with the ig predator @xcite . including such facilitation in ecological theory will fundamentally change many basic predictions and will enable a better understanding of functioning of many natural communities @xcite . especially in the igp systems
an emphasis should be given to the elucidation of the effects of facilitation on community composition and stability @xcite .
contrary to the competitive exclusion principle in systems with competitors for a single resource stability stems from commensalism @xcite .
hereby one consumer can in some way alter the habitat to benefit the other .
recently such an interaction was observed in experiments with a microzooplankton food web community @xcite .
the experimental system included two predators : a tintinnid species _
favella ehrenbergii _ and a heterotrophic dinoflagellate species _
gyrodinium dominans_. they are both grazing on a phototrophic dinoflagellate _
scrippsiella trochoidea_. the authors showed that the ig predator _
f. ehrenbergii _ can precondition a substantial part of the common resource _
s. trochoidea _ during its feeding procedure by immobilizing the common prey without ingestion .
such preconditioned individuals can be captured more easily by the ig prey _
g. dominans _ than the mobile individuals of the same resource species .
this mutualistic interaction leads to higher growth rates of the ig prey in the presence of the ig predator .
the authors characterized their experimental observations as a facilitative igp relationship with a commensalistic pattern .
our motivation for this modeling study was to investigate if such commensalistic patterns can create loopholes for a stable coexistence of all species in the investigated system .
of our major interest was if in the igp system an immobilization @xcite or the partitioning of prey populations into distinct groups of individuals offers opportunities for competition avoidance among both consumer species .
we reformulated the 3-species igp model proposed in @xcite to include multiple subpopulations of prey .
furthermore , we explored the effect of diversification of the resource available to higher level consumers on the species persistence by numerical simulations of an extended igp module .
specifically , we investigate how the addition of new links to a focal igp module enhances stability of population dynamics by reducing the competitive interactions of predators for their shared resource . in order to explain the results of the experimental findings of leder et al .
we investigated the influence of multiple traits of the resource community on a stable coexistence in the 3-species model with different types of resource .
for this purpose we adapted and reformulated the original model by holt and polis ( 1997 ) and added a new type of interaction .
this link specifies the immobilization mechanism that depends on the densities of mobile and immobile resource items and the top predator which creates the immobile resource fraction during feeding .
the immobilization term is used to model the interactions between the ig predator and the resource .
another type of interaction considered in this paper is a resource turnover mechanism .
this mechanism describes mutual interactions between species from distinct resource subpopulations .
the interaction term depends exclusively on the resource subpopulation densities .
the rate of turnover is constant . if no turnover or immobilization of individuals from one group to another occurs then the basic igp model with a single population of resource is recovered .
we discuss the influence of immobilization and transfer of species on the coexistence patterns in a system with different subpopulations of the resource and compare the results with the basic 3-species igp .
this paper is organized as follows : in the first section we introduce a general 3-species igp model with a new type of interaction that links the resource pools to the top consumers . in the following sections two distinct igp formulations with @xmath0 resource subpopulations are discussed .
both models are derived from the basic igp module by including additional links : @xmath1 the immobilization by the predator and @xmath2 the resource turnover . in the results section
we numerically investigate stability of equilibrium densities for various trophic configurations .
data from numerical analysis are presented for the igp model with the immobilization and for the model with the resource turnover .
at last we discuss results for a general igp model with the resource turnover mechanism and @xmath3 subpopulations of the resource . after sketching the main conclusions we review the model predictions and compare their relevance to the immobilization experiment @xcite .
furthermore we discuss possible alternative reformulations of the model . in the appendix explicit forms for the steady states for two simple analytical cases and multidimensional system are specified . as a part of a linear stability routine the jacobian matrices for two types of formulations
are given .
finally , we carry over to a higher dimensional formulation and describe the parameters choice and the equilibrium densities .
we introduce an omnivory model with an igp unit derived from a simple non - spatial lotka - volterra system with the linear functional responses adapted from holt and polis ( 1997 ) .
the original model consists of populations of two predators ( ig predator and ig prey ) and a common resource . here
, we include new features such as a resource differentiation mechanism which affects palatability of a fraction of resource for the predators .
specifically , the entire resource population is subdivided into distinct groups under the assumption that the groups differ from each other by the quality and fitness of the individuals .
they are consumed by the predators at different group - specific grazing rates .
the differentiation of the resource could be due to damage by the predator or initial inhomogeneous distribution of the resource quality .
afterwards , we generalize our model to the case of the multiple resources . the food web model for a multiple number of prey subpopulations @xmath4 is sketched in fig .
[ fig1 ] a. the top predator @xmath5 and the intermediate predator @xmath6 are engaged in the igp and share a common resource @xmath7 .
the resource pools are not independent because there is an exchange of individuals among different subpopulations @xmath8 following the links in fig .
another special case of the igp with two distinct populations of resource @xmath9 and @xmath10 is presented in fig .
shown is a schematic view of trophic interactions including intraguild predation and two populations @xmath10 and @xmath9 of immobilized and mobile resources respectively .
the ig prey @xmath6 competes with the ig predator @xmath5 for both resource types and is also an additional resource for the ig predator .
the size of the population @xmath10 increases due to immobilization of individuals from the population @xmath9 by the ig predator @xmath5 .
we begin with an overview of a general igp model and all the important trophic links and parameters that are used to define it .
later we focus specifically on two different formulations of the general igp model .
the general model for a food web with an inhomogeneous resource is derived from the lotka - volterra omnivory model @xcite with the interaction term that accounts for the transitions among different pools .
the lotka - volterra omnivory model consists of @xmath11 equations .
it is used as an approximation for the food web community with the igp and @xmath12 mutually interacting subpopulations of the resources . in the absence of predation a basal population @xmath7 develops according to logistic growth @xcite .
the set of equations for the population densities are written as follows : _ shared @xmath13st resource _ : @xmath14 s_1\nonumber\\ & & -z_1(s_1,s_2,\ldots , s_n , g , f),\nonumber\end{aligned}\ ] ] _ shared @xmath15th resource ( @xmath16 ) _ : @xmath17 s_k,\nonumber\end{aligned}\ ] ] _ intermediate predator ( ig prey ) _ : @xmath18 _ top predator ( ig predator ) _ : @xmath19 the parameters of the model and main populations are described in details in table [ tab1 ] . here
@xmath20 is the maximum specific growth rate of the resource population @xmath7 , @xmath21 is the carrying capacity of the resource defined as enrichment factor in the previous models @xcite .
the subpopulations @xmath8 are derived from the basal resource @xmath7 via immobilization or via individual - to - individual turnover .
species from @xmath7 and @xmath22 are consumed by the ig predator at potentially different rates @xmath23 and @xmath24 and by the ig prey at rates @xmath25 and @xmath26 respectively .
the differentiation among subpopulations @xmath8 is preserved by a choice of distinct predation pressures @xmath27 , feeding rates @xmath28 and mortality coefficients @xmath29 .
the density - independent mortality rates for @xmath30 and @xmath5 are @xmath31 and @xmath32 correspondingly .
they are used as factors limiting the growth of the populations in ( [ eq.1 ] ) .
a key assumption of the model is that there is only one - directional movement between the basal resource @xmath7 and its fractions @xmath22 .
the local interactions among individuals from alternative pools are embedded via functional terms @xmath33 provided in table [ tab2 ] for each type of the igp formulation .
these terms account for transitions among the resource items @xmath22 .
the general omnivory model ( [ eq.1 ] ) can be reduced to three types of igp formulations : system with immobilization and systems with the resource turnover for @xmath0 subpopulations and for @xmath3 pools . for each of the formulations
specific expressions of functional forms @xmath34 and @xmath35 are provided in the table [ tab2 ] .
the term @xmath36 is responsible for the exchange of individuals among subpopulations @xmath37 due to the species turnover or the immobilization mechanism .
the transfer of individuals from the population @xmath7 to @xmath8 happens instantaneously at constant rates @xmath38 correspondingly .
analogously @xmath39 are defined as instantaneous migration rates among subpopulations @xmath8 .
the terms @xmath40 and @xmath41 are used to evaluate the total predation of the ig prey and the ig predator on the resource . to achieve a stable persistence of all species
the ig prey should benefit more from an alternative resource than the ig predator .
for this reason , whereas the attack rates of the ig predator are equal for different resource pools , the ig prey establishes a higher predation pressure on subpopulations @xmath22 than on the basal pool @xmath7 .
the numerical values for the attack rates are chosen to be close to the experimentally observed values @xcite .
holt and huxel ( 2007 ) used an extended igp module with alternative resources that are defined independently .
they evolve according to their own intrinsic growth rates .
as opposed to the formulation given by holt and huxel ( 2007 ) and to a model with trophic supplements @xcite here we do not consider external alternative resources . in our model with immobilization the population density in every resource pool varies due to immobilization by the ig predator and consumption by the predators .
similarly in the formulation with the resource turnover the transfer mechanism between resource subpopulations plays a role in exchange among the distinct resource pools .
alternative pools grow due to the influx of species from the basal resource or the other pools .
therefore the sizes of subpopulations are controlled mainly by a number of direct encounters with the ig predator ( immobilization ) or by a species turnover from one resource subpopulation to another .
in addition , the individuals in the different pools of the basal resource are distinguished by group - specific predation pressures that establish a top - down regulation of densities of each subpopulation . in the following sections we present an explicit formulation of the model with immobilization and of the model with the resource turnover for @xmath0 subpopulations . ' '' '' ' '' '' .[tab1 ] the variables and parameters for the general model ( [ eq.1 ] ) .
[ cols= " < , < " , ] ' '' '' ' '' '' 2 the system with the immobilization illustrated in fig .
[ fig1 ] b is derived from the equations ( [ eq.1 ] ) for two resource subpopulations by substituting the interaction terms @xmath42 and @xmath43 from table [ tab1 ] .
after the substitution the set of equations for the igp model with immobilization yields : + _ mobile resource : _
@xmath44s_m\nonumber\\&&-(f+i_m ) f s_m,\nonumber\end{aligned}\ ] ] _ immobilized resource : _
@xmath45 s_i , \nonumber\end{aligned}\ ] ] _ ig prey : _ @xmath46 g,\nonumber\end{aligned}\ ] ] _ ig predator : _ @xmath47 f,\label{eq.2}\end{aligned}\ ] ] where the state variables @xmath9 and @xmath10 are the densities of mobile and immobilized species .
note that the feeding rates of the top predator @xmath5 on both populations @xmath10 and @xmath9 are equal .
by contrast , the attack rate of the ig prey on immobilized subpopulation is higher than on mobile species .
the relation @xmath48 holds in the presence and in the absence of the predator @xmath5 .
this assumption is well justified by the observations of an experiment with artificial immobilization @xcite . _ g. dominans _ demonstrate a strongly selective behaviour towards immobilized species when offered in a mixture with mobile cells of _ s. trochoidea_. it was measured that ingestion rates of the predator in the immobilized prey treatment were by a factor of @xmath49 greater than those in the control treatment .
the stability of equilibrium densities and the persistence zones of the system ( [ eq.2 ] ) with a non - zero immobilization rate are discussed in section [ sec.3.1 ] . the model with the resource turnover
is derived from the general case ( [ eq.1 ] ) by substituting the functional forms from table [ tab2 ] .
it is written as follows : + _ @xmath13st resource _ : @xmath50 s_1\nonumber\\ & & -[f f+t_r s_2 ] s_1\nonumber,\end{aligned}\ ] ] _ @xmath51nd resource _ : @xmath52 s_2,\nonumber\end{aligned}\ ] ] _ ig prey _ : @xmath53 g,\nonumber\end{aligned}\ ] ] _ ig predator _ : @xmath54 f.\label{eq.3}\end{aligned}\ ] ] all the parameters are chosen the same as for the system with immobilization ( [ eq.2 ] ) .
note that the evolution equations are written as in ( [ eq.2 ] ) but immobilization term is replaced with the transfer term that is dependent on the population densities .
the transfer between the two subpopulations occurs each time whenever species from two different pools encounter each other . in the simplest case
the number of encounters is proportional to the population densities of @xmath7 and @xmath55 .
if the density of second subpopulation is zero and no differentiation in the resource takes place at @xmath56 than the top predator @xmath5 outcompetes the predator @xmath6 due to a higher predation rate ( @xmath57 ) .
this outcome is predicted by the basic igp model @xcite . by contrast ,
whenever the turnover of species takes place and non - zero densities are produced in the resource pool @xmath55 the intraguild predation introduces a higher pressure on the second subpopulation @xmath55 .
this will potentially lead to a negative effect on the population density in @xmath55 and to higher levels of subpopulation @xmath7 .
the result of this interaction is that the 3-species coexistence is reached via the igp competition trade - off .
we illustrate an emergent dynamical behaviour for the three formulations provided in table [ tab3 ] with stability diagrams . due to high dimensionality of the models ( [ eq.1])-([eq.3 ] )
the analysis of an entire parameter space is intractable .
only several illustrative examples for every formulation will be shown here . in fig .
[ fig2 ] the regions of stable positive equilibrium solutions versus immobilization and enrichment are shown .
the parameter space is partitioned into several stability zones associated with the regions of coexistence , exclusion of both predators and exclusion of the ig prey at @xmath58 .
the boundaries defined for partitioning of the diagram are found from the eigenvalue analysis ( see appendix ) . as shown in fig .
[ fig2 ] at low enrichment the densities of both predators decay to zero and the summed abundance of the resource reaches steady state at @xmath59 .
the case of zero immobilization has been already considered in previous studies @xcite . at low immobilization and at high enrichment
only the top predator and resource are stable and positive , just as in the 3-species igp model @xcite , whereas the coexistence between both predators and common resource is possible only in the regions of intermediate enrichment . a higher mortality rate for the predator @xmath6 results in its extinction in the region of low immobilization in fig . [ fig2]b where an extra resource can no longer support its persistence .
only the ig predator and resource persist in this region of parameters .
situation is different for higher immobilization where a large region of coexistence for both predators exists .
the equilibrium densities shown on the diagrams are defined in the eq .
( [ eq.7 ] ) in appendix .
[ fig3 ] shows equilibrium densities of the four components of the food web and their dependence on enrichment and immobilization rates .
for a high immobilization rate the resource population is dominated by immobilized individuals . meanwhile at low immobilization mobile and immobilized populations increase along the gradient of enrichment
an adverse pattern occurs at high immobilization .
the growth rate of the ig predator is noticeably reduced at @xmath60 due to an increase of the competitive trade - off with the ig prey .
the dependence of the population densities on the enrichment of resource is shown in fig .
[ fig4 ] for the immobilization @xmath61 .
meanwhile as predicted from the standard igp model @xcite the ig prey is excluded at high enrichment in the model with immobilization at @xmath61 the ig prey benefits from immobilized resource and its persistence is increased at a broader range of carrying capacities .
the density of the mobile ( immobilized ) resource subpopulation reach saturation threshold at a higher enrichment ( see fig .
[ fig4 ] ) . how sensitive is a stable coexistence to small variations of the attack rates of the intermediate predator @xmath6 ?
will our predictions be still valid ? to examine the system behaviour for different attack rates of @xmath6 the coexistence zones are exemplified for different values of @xmath25 in fig .
colorcode is assigned according to grazing rate @xmath25 .
overall the stability diagram exhibits similar pattern as in fig .
specifically , the region of stable coexistence enlarges for higher immobilization . as it seems reasonable the number of stable solutions and the 3-species permanence zone in fig .
[ fig5 ] gradually broadens with the increase of predation pressure from predator @xmath6 .
simultaneously fewer exclusion steady states for the predator @xmath6 are discovered . in this section
the equilibrium solutions and stability of the equilibria are discussed for the system ( [ eq.2 ] ) with the mechanism of species turnover . in fig .
[ fig6 ] the regions of stable ( unstable ) equilibria are plotted versus the enrichment and the transfer rate .
the results are contrasted on the stability diagrams in figs .
[ fig6 ] @xmath62 for different predation rates of the ig predators .
four different states are localized in the parameter space that corresponds to stable ( unstable ) persistence and exclusion of the ig prey ( ig predator ) .
we investigate how the dynamics in the extended igp system responded to variation of enrichment levels . for each case shown in fig .
[ fig6 ] an increase of enrichment is accompanied with a series of bifurcations in the system manifested by an invasion of higher trophic levels similar to predictions from the linear food chain theory @xcite .
for instance , at low enrichment both regimes @xmath63 and @xmath13 are stable .
further increase of @xmath21 at fixed @xmath64 results in a chain of bifurcations from a stable regime @xmath63 to an unstable @xmath51 and subsequently to a stable coexistence regime @xmath13 .
a further increase of carrying capacity favours an exclusion of @xmath6 and shifts the population densities towards the dominance of the ig predator .
an interesting feature is that at low transfer rate only a coexistence of the ig predator and the resource is found .
the second subpopulation @xmath55 is extinguished fast due to predation and low transfer rate .
the steady states found for low enrichment are similar to the case of a single prey population without transfer mechanism at @xmath56 and @xmath65 .
as typified on the diagram in fig .
[ fig6 ] d the ig prey levels remain positive . since the ig prey has an advantage as a competitor for the shared resource only the ig predator gets excluded from the system .
the stability behaviour of the system ( [ eq.3 ] ) is highly sensitive to the alternations of attack rates of @xmath6 and the productivity of resource @xmath20 .
changes of these parameters produce different emergent patterns as shown in fig .
the location of states of stable ( unstable ) permanence and the exclusion zone of @xmath6 is still comparable to the patterns shown in fig .
[ fig6 ] , however the region of 3-species coexistence gets visibly reduced .
the reduction is more evident on the plots fig .
[ fig7 ] @xmath66 and @xmath67 . at higher transfer rates
the coexistence of all 3-species is no longer observed and only the population of ig prey and resource persist .
due to low productivity the densities of the basal resource @xmath7 are quickly depleted and the ig predator is driven to extinction . on the contrary ,
conditions become more profitable for the ig prey that is released from the igp pressure and simultaneously obtains more benefits by predation on the extra resource @xmath55 . at low transfer rates ( fig .
[ fig7]c and d ) the ig predator is excluded independently on carrying capacity of the resource .
as expected , with increase of the attack rate of @xmath6 the population of the ig predator is driven to extinction due competition with ig prey .
however , situation becomes more favourable for the ig predator at higher values of the transfer coefficient @xmath68 . for high enrichment and intermediate transfer the ig prey
is excluded from the system . at a fixed enrichment
several alternating states are found along the gradient of @xmath68 ( see fig .
[ fig7 ] d ) . for example , at @xmath69 the behaviour of the food web is very sensitive even to a small alternations of @xmath68 .
indeed , the system passes through distinct steady states just within a small increment of transfer rate .
the exclusion of the ig predator is observed at @xmath70 , the coexistence is found at @xmath71 and the exclusion of the ig prey is achieved at @xmath72 . finally at a higher transfer values ( @xmath73 )
both predators enter the system and persistence is reached . after presenting the results for the systems ( [ eq.2 ] ) and ( [ eq.3 ] ) we proceed to a more complex situation with @xmath3 of distinct subpopulations of the resource . for a multipopulation model
the choice of parameters including predation rates can be enormously large .
as a consequence more freedom is provided for choosing equilibrium densities that can fit the model ( [ eq.1 ] ) .
since it is impossible to investigate the entire range of biologically plausible parameters we make a particular choice of parameters that allow an easier comparison of the case @xmath3 in ( [ eq.1 ] ) with the model ( [ eq.2 ] ) .
the details of the procedure are provided in appendix . in this section
we show the results of the numerical simulation for the model with @xmath3 prey subpopulations .
the system ( [ eq.1 ] ) for the case @xmath3 is integrated numerically .
for the calculation of the stability diagrams at different fixed values of enrichment and transfer rate we perform @xmath74 simulations .
the results of the simulations for @xmath75 and @xmath76 subpopulations are illustrated in fig .
the percentage of stable 3-species coexistence is calculated for every point in the parameter space with fixed enrichment @xmath21 and limiting value @xmath68 .
the colorcode is assigned according to the percentage of stable coexistence solutions found for @xmath74 food webs . in all the replicas of the simulated system the steady state densities for @xmath77 and @xmath7
are fixed ( see ( [ eq.12 ] ) in the appendix ) .
thus only the variations among possible equilibrium densities @xmath78 are examined .
the constraints for the parameters of high dimensional system ( [ eq.1 ] ) are given in eqs .
( [ eq.14])-([eq.16 ] ) in the appendix . the stability diagrams in fig .
[ eq.8 ] show some similarities to the regions of coexistence in figs .
[ fig6 ] and [ fig7 ] found for the @xmath0 subpopulation model ( [ eq.3 ] ) .
the size of the stability zone expands with the increase of the transfer coefficient
. at low transfer rates no stable persistence is found , but different alternative traits .
the percentage of stable food webs with 3-species is substantially lower for a large system with @xmath79 subpopulations than for @xmath80 .
this reduction in stability is independent on the number of simulated food webs and a choice of main parameters of the system .
it is possible that an increase of food web connectivity in this case impacts negatively the system ( [ eq.1 ] ) stability .
another feature is that for @xmath79 the percentage of stable equilibria at a fixed enrichment value decreases for large values of @xmath68 unlike in previous cases in fig .
[ fig8 ] a - c .
the results of the numerical simulation demonstrate that for @xmath0 subpopulation up to @xmath81 of stable systems are found at a higher transfer rate and an intermediate enrichment .
second , for a larger food web with @xmath79 subpopulations a higher percentage of stable steady states ( up to @xmath82 ) are identified at low transfer rate and at high enrichment .
we compare the results of simulation for four cases @xmath83 at fixed enrichment @xmath84 and variable transfer coefficient in fig .
the yields are derived for @xmath85 simulations of food webs .
the estimations of the number of steady states show that for food webs with @xmath86 a higher percentage of solutions with a stable coexistence are identified than for food web with @xmath79 pools .
indeed , the yield for @xmath0 reaches almost @xmath81 meanwhile the percentage of stable food webs found for @xmath79 saturates at @xmath87 for large @xmath68 .
the non - monotonic variations of the yields in fig .
[ fig9 ] reveal a highly sensitive behaviour of the igp model ( [ eq.1 ] ) to a change in transfer rate in all cases .
for @xmath79 the percentage of stable food webs reaches @xmath88 at a low transfer rate .
it decreases substantially for higher values of the transfer rate . for @xmath86
there is an overall incline from @xmath89 at @xmath90 to @xmath81 at @xmath91 of stable configurations .
two types of stable equilibrium solutions are illustrated in fig .
[ fig10 ] .
both solutions are obtained inside the stable coexistence region as indicated in fig .
the system ( [ eq.1 ] ) is simulated with @xmath92 number of subpopulations and initial conditions as defined in the appendix . for the steady state in fig .
[ fig10 ] a and the oscillatory state in fig .
[ fig10 ] b most of resource subpopulations are unstable and their densities rapidly decline to zero after some initial transient .
nevertheless , coexistence in the system is typically supported by one or two resource pools with non - zero densities .
+ + +
there is growing evidence from theoretical and empirical studies that creating additional trophic links have a stabilizing effect on food webs @xcite .
generalized models reveal that the stability of food webs can be enhanced when species at higher trophic levels graze upon multiple prey species @xcite .
in particular , for low dimensional food webs it is demonstrated that an addition of alternative food resources can stabilize the interactions @xcite and open up a possibility for feedbacks on population dynamics due to apparent competition .
the predictions of our model confirm the main conclusions given in a theoretical study of an extended igp model @xcite . in the alternative formulations used here the ig prey
has the access to an extra resource beyond the shared resource for which both predators compete .
this extra resource is a more attractive resource item for the ig prey and is thus attacked at higher rates by the ig prey whereas the attack rate of the ig predator stays the same .
moreover the ig predator indirectly stimulated the growth of the ig prey population by providing this extra resource .
our predictions tested by the application of a stability analysis are robust in the sense that they are independent of the form of the interaction term that is responsible for the availability of an additional resource .
we demonstrate that for different formulations of the basic igp model with the embedded interactions a stable 3-species coexistence is ultimately reached whenever a moderate strength of the omnivorous links is used .
however , the problem to relate the experimental findings @xcite to the theoretical predictions of our model ( [ eq.2 ] ) still remains open .
since the experiment is aimed to observe a short term populations development it is not easy to find a direct correspondence between empirical population dynamics and theoretically predicted behaviour . in the @xmath93 day batch culture experiment of lder
et al . with all 3 species present both predators _ g. dominans _ and _ f. ehrenbergii _ displayed positive growth while the prey population _ s. trochoidea _ displayed simultaneously a sharp decline to almost zero .
how can this behaviour be classified according to our theoretical model ? could it be a part of an oscillatory cycle or an unstable state ?
it is not easy to answer these questions , however , we can make a guess that the short term evolution observed in the experiment recasts as a part of an oscillatory cycle for a periodic equilibrium state found at intermediate immobilization . a similar type of experiments performed for various initial species densities could furnish a justification of this hypothesis .
the above results demonstrate that a persistence of ig predator , ig prey and resource is achieved even at a low value of immobilization rate .
moreover , a significantly higher percentage of observed stable configurations is found when the immobilization and transfer links in ( [ eq.1 ] ) and ( [ eq.2 ] ) are strengthened . because our model is an oversimplification of the experimental behaviour the partitioning of the parameter space according to stable versus unstable coexistence could be used as an approximation of the population dynamics found in a real experimental situation .
firstly , the conditions for long term stable coexistence found by numerical simulations are not so easy to examine experimentally because of technical and temporal restrictions .
experimental samples in ref .
@xcite are taken during @xmath93 days of incubation due to a decline of the prey population . secondly , due to the existence of stable limit cycles as predicted by our linear stability analysis ( see fig .
[ fig10 ] ) the oscillatory solutions go through a period of very low densities and might be driven to extinction in the presence of random fluctuations of the environment .
we point out that our numerical simulations of the extended 3-species igp module ( [ eq.1 ] ) do not explore the entire swath of parameters and configurations for the steady states .
rather our analysis focuses on explaining conceptual features of the igp model with diverse prey populations .
earlier studies focussing on dynamics of complex ecological communities @xcite demonstrated the importance of multiple prey traits in mitigating predator selection pressures and altering predator - induced behavioural shifts in natural environments @xcite .
our model can also be adapted to food web communities in which differentiation among prey individuals , namely , variation in individual traits such as fitness and mobility , is a result of heterogeneities in their natural habitat and/or adaptation of the species to the local conditions of the habitat .
we show that an existing diversity of resource items traits can significantly alter the emergent community patterns . adding new subpopulations of resources with distinct traits that are more vulnerable to an attack from the ig predator
facilitates the coexistence of both igp - related predators which compete for the common food resource .
thus , the presence of an alternative resource indirectly induces shifts in exploitative competition .
it is important to note that a general mathematical model with density- dependent interactions and immobilization do not render a unique theoretical description for the results of the experiment @xcite .
our model predictions can be tested against alternative formulations .
indeed , the main features of the experimental system can be examined by the inclusion of predation rates that are dependent on the mobility of the resource species . since slow and immobile individuals
can also be found among mobile species one can use an inhomogeneous distribution of velocities of the resource species in a theoretical model . to guarantee more benefit for an intermediate consumer in catching a certain type of individual distinct predation rates should be assigned according to different velocities of resource species .
another question is if the growth rate of the ig predator will be affected by the inclusion of the time of resource capture . how will the inclusion of the time lag change the predictions of our immobilization model ?
these extensions of a general igp model will be a topic for our future investigations .
finally , we point out that it is of potential interest for biological control and conservation management to understand functioning of omnivory and igp systems in relation to global changes of the environment .
since igp food webs are widespread in natural communities their adaptation and resilience behaviour is principal for understanding the restructuring of natural communities .
we have used three formulations of a general igp model to explore the effects of increasing diversity in the prey population on higher trophic levels .
the reformulated igp model alters the results from the basic igp theory @xcite .
we show that an increase of a number of trophic interactions in the system via differentiation of resource can stabilize the population dynamics of the igp module .
this conclusion holds for the densities of the ig prey that level up even when the ig predator is a superior competitor for the common basal resource .
first , we show that for the system with the immobilization term up to three regions of stable trophic configurations are observed along the enrichment gradient .
meanwhile at low enrichment both ig prey and ig predator are excluded , at high enrichment the presence of only small concentration of immobilized cells is sufficient to facilitate the coexistence of the competitors in the igp relationship .
moreover the percentage of all admissible trophic configurations for the 3-species persistence inclines substantially for higher immobilization .
second , given that immobilization is high enough it prompts the exchange between pools of mobile and immobilized resource and facilitates fast decline of the mobile population and a growth of immobilized subpopulation .
meanwhile the exchange between the basal mobile resource and the predators gets weaker due to the low density of mobile species the immobilized individuals become a major food resource for the predators . because the ig prey is a superior competitor for immobilized resource
a robust coexistence of both predators will be easily supported .
in addition , along an increasing gradient of immobilization the relative abundance of ig prey becomes higher than the abundance of ig predator .
restructuring of the basic igp module by adding individual - to - individual turnover facilitates the coexistence and stabilizes the otherwise unstable system .
moreover a strengthening of the interaction link leads to a significantly broader range of enrichment values at which stable coexistence could be found . at low transfer rate
two types of equilibria are observed : @xmath1 if the ig predator is a superior competitor for the resources than at low enrichment both predators are excluded and at high enrichment only the ig predator stays in the system ; @xmath2 an increase of the attack rates of the ig prey depresses the population of the ig predator until it is completely excluded .
numerical simulations of food web ( [ eq.1 ] ) with @xmath75 and @xmath76 distinct pools demonstrate that the high dimensional food webs overall manifest far less stable behaviour than the food webs with only two distinct subpopulations .
an interesting feature is that the percentage of stable states for @xmath79 substantially decreases from @xmath94 to @xmath87 with an apparent increase of the value of transfer rate .
by contrast , for food webs with @xmath86 an increase in transfer rate leads to the growth of the percentage of stable coexistence solutions from about @xmath89 to @xmath81 . in the appendix
we review the steady state solutions for the lotka - volterra models ( [ eq.2 ] ) and ( [ eq.3 ] ) and provide jacobian matrices to examine their local stability for the coexistence of both predators and the resource .
first , equilibrium solutions are derived for the 3-species model ( [ eq.2 ] ) with zero immobilization ( @xmath95 ) and zero initial size of immobilized population ( @xmath96 ) .
second , the steady states are given for the model ( [ eq.2 ] ) with immobilization ( @xmath97 ) .
at last , the equilibrium solutions are presented for the system with the resource turnover ( [ eq.3 ] ) .
for every case various trophic configurations are considered : @xmath1 exclusion of both predators , @xmath2 exclusion of ig prey or ig predator and @xmath98 the 3-species coexistence .
the equilibrium solution of ( [ eq.2 ] ) for the 3-species coexistence without immobilization is stated as follows : @xmath99 where @xmath100 .
the necessary condition for the coexistence requires that the right hand side is positive in ( [ eq.4 ] ) .
the expressions for the equilibrium densities for the survival of the ig prey and the resource with exclusion of the ig predator at @xmath101 yield : @xmath102 the condition for persistence of the ig prey and the resource reads : @xmath103 . at zero density of the intermediate consumer ( @xmath104 ) one yields the steady states of the resource @xmath105 and the ig predator @xmath106 : @xmath107 the densities are positive if and only if the condition @xmath108 holds true . for the model ( [ eq.2 ] ) with immobilization @xmath109 we define a set of equilibrium densities to satisfy the equalities below : @xmath110s_m,\nonumber\\ q_2&=&i_m f s_m - s_i ( b g+f f ) , \\
q_3&=&(a'a s_m+a'b s_i - g f - m_g ) g,\nonumber\\ q_4&=&(f'f ( s_i+s_m ) + g g ' g - m_f ) f.\nonumber\label{eq.7}\end{aligned}\ ] ] the system ( [ eq.7 ] ) has four alternative solutions : @xmath1 exclusion of both predators at @xmath111 ; @xmath2 exclusion of the ig predator @xmath112 at @xmath113 ; @xmath98 the coexistence of resource and the ig predator @xmath114 at @xmath58 ; @xmath115 the 3-species coexistence @xmath116 . in the absence of @xmath5
the immobilization mechanism is not active and the model ( [ eq.2 ] ) reduces to the system without immobilization @xcite where the equilibrium solutions written as ( [ eq.5 ] ) . upon exclusion of the ig prey in ( [ eq.7 ] ) one obtains expression for the equilibrium densities of resource and the ig predator : @xmath117 note that the size of mobile population @xmath118 is proportional to the size of immobilized population @xmath119 . as is expected the population of immobilized preys is impacted positively by the increase of immobilization .
although the predation pressures are equal for both resource subpopulations the immobile population @xmath119 extinguishes faster than the mobile population @xmath118 . indeed , with the increase of predation rate @xmath23 the following approximations hold : @xmath120 and @xmath121 .
the matrix diagonal is written in terms of the equilibrium densities @xmath125 and @xmath123 as follows : @xmath126 the solution ( @xmath127 ) is globally asymptotically stable in the phase space @xcite if the condition for stability is satisfied . for the stable coexistence
it is necessary that the real parts of all four eigenvalues @xmath128 of the stability matrix are non - positive . to obtain the boundary for stability regions in the parameter space the eigenvalues of the above stability matrix
are evaluated numerically at different parameters combinations .
the resulting stability diagrams are presented in fig .
[ fig2 ] and fig .
[ fig5 ] . as in the previous case
the system [ eq.3 ] for @xmath0 pools permits four steady states : @xmath1 the exclusion of the predators at @xmath129 ; @xmath2 the exclusion of the ig predator @xmath130 at @xmath113 ; @xmath98 the exclusion of the ig prey @xmath131 at @xmath58 ; @xmath115 the coexistence of 3-species @xmath132 .
the solution for the coexistence of the resource and the ig prey in the absence of the ig predator is expressed as follows : @xmath133 note that at @xmath56 the ig prey is excluded and steady state density for the resource approach the carrying capacity limit @xmath21 .
the positive solution of ( [ eq.10 ] ) exists if the parameters satisfy the inequality : @xmath134 .
the steady state for the resource and the ig predator in the absence of ig prey yields : @xmath135 a nontrivial solution for the 3-species coexistence is found by solving for the equilibrium @xmath136 of the following system : @xmath137 finally , the condition for the stable coexistence is provided by solving for the eigenvalues of the stability matrix : @xmath138 $ ] + the eigenvalues of the stability matrix are functions of the constant rates ( see table . [ tab2 ] ) and the equilibrium densities @xmath139 and @xmath123 .
the condition for the stable coexistence requires that the real parts of the eigenvalues of the stability matrix are non - positive .
here we define initial conditions that are used for the numerical simulation of the model ( [ eq.1 ] ) with @xmath3 number of subpopulations . unlike for the system ( [ eq.3 ] ) with @xmath0 subpopulations
the equilibrium densities can no longer be determined analytically . to simplify the search for the equilibria in ( [ eq.1 ] ) we implement several assumptions .
we chose the parameters and equilibrium densities to fulfil several constraints provided below .
initial densities are equal to the steady states ( [ eq.12 ] ) for @xmath0 subpopulations , namely : @xmath140 and @xmath141 . for the sake of simplicity the values for the remaining densities @xmath142
are defined to be less than the equilibrium density @xmath143 .
provided that the interaction rates @xmath144 are randomly assigned values not exceeding @xmath145 the equilibrium density @xmath146 can be found from the constraint : @xmath147 we chose the zero rates of decline @xmath148 and @xmath149 for the first two subpopulations @xmath7 and @xmath55 .
the values for the grazing rates @xmath150 are randomly assigned from the interval @xmath151 $ ] . in the next step an antisymmetric matrix @xmath152 of individual - to - individual interactions
is defined with the upper diagonal coefficients @xmath153 that obey the inequalities : @xmath154 . the remaining lower diagonal coefficients @xmath155 should satisfy the antisymmetry relation : @xmath156 .
the set of equations in ( [ eq.1 ] ) for @xmath157 holds true if the positive mortality rates @xmath158 are expressed from the equations as follows : @xmath159 the equation ( [ eq.15 ] ) for @xmath160 is used to solve for the attack rate @xmath161 .
finally , the attack rates @xmath162 and the feeding rates @xmath163 are randomly assigned from the intervals @xmath164 $ ] and @xmath165 $ ] correspondingly .
the choice of the feeding rates enables to reduce the predation pressure on populations @xmath142 by a factor of @xmath166 .
the remaining grazing rates @xmath167 and @xmath168 are defined from the relations : @xmath169 and @xmath170 . finally , two constraints hold to solve for the coefficients @xmath171 and @xmath172 : @xmath173 we want to emphasize that the attack rate @xmath168 should be higher than any of the rates @xmath174 . nevertheless since @xmath175 relation holds the total predation pressure of @xmath6 summed over alternative pools exceeds the predation exclusively on @xmath7 .
this far the positive population density of the ig prey is maintained due to the consumption of alternative resources @xmath176 .
we provided a special assignment of parameters for the predation , feeding and mortality that enables to fulfil the condition of positive equilibria that are comparable to the realistic biodensities . moreover the stability results can be conclusively compared with two different formulations ( [ eq.1 ] ) and ( [ eq.3 ] ) .
00 abrams , p. a. , roth j. , 1994 .
the responses of unstable food chains to enrichment .
ecol . 8 , 150 - 171 .
abrams , p. a. , roth , j. d. , 1994 .
the effects of enrichment of three - species food chains with nonlinear functional responses .
ecology 75(4 ) , 1118 - 1130 .
abrams , p. a. , fung , s. r. , 2010 .
prey persistence and abundance in systems with intraguild predation and type-2 functional responses .
264 , 1033 - 1042 .
amarasekare , p. , 2000 .
coexistence of competing parasitoids on a patchily distributed host : local vs spatial mechanisms .
ecology 81 , 1286 - 1296 .
amarasekare , p. , 2006 .
productivity , dispersal and the coexistence of intraguild predators and prey .
234 , 121 - 133 .
amarasekare , p. , 2007 .
spatial dynamics of communities with intraguild predation : the role of dispersal strategies .
naturalist 170(6 ) , 819 - 831 .
arim , m. , marquet , p. a. , 2004 .
intraguild predation : a widespread interaction related to species biology .
lett . 7 , 557 - 564 .
bonsall , m. b. , holt , r. d. , 2003 .
the effects of enrichment on the dynamics of apparent competitive interactions in stage - structured systems .
naturalist 162 , 780 - 795 .
borer , e. t. , briggs , c. j. , murdoch , w. w. , swarbrick , s. l. , 2003 .
testing intraguild predation in field system : does numerical dominance shift along a gradient of productivity ?
6 , 929 - 935 .
brodeur , j. , rosenheim , j. a. , 2000 .
intraguild interactions in aphid parasitoids .
97 , 93 - 108 .
brose , u. , berlow , e. l. , martinez , n. d. , 2005 . from food webs
to ecological networks : linking non - linear trophic interactions with nutrient competition . in : de ruiter , p. c. , moore , j. c. , wolters , v.(eds . ) , dynamic food webs .
elsevier , pp 27 - 36 .
bruno , j. f. , stachowicz , j. j. , bertness , m. d. , 2003 .
trends evol .
18 ( 3),119 - 125 .
crowley , p. h.,cox , j. j. , 2011 .
intraguild mutualism , trends evol .
26(12 ) , 627 - 633 . daugherty , m. p. , harmon , j. p. , briggs , c. j. , 2007 .
trophic supplements to intraguild predation .
oikos 116 , 662 - 677 .
denno , r. f. , fagan , w. f. , 2003 . a stoichiometric perspective on omnivory in arthropod - dominated ecosystems : the importance of nitrogen limitation .
ecology 84 , 2522 - 2531 .
dewitt , t. j. , langerhans , r. b. , 2003 .
multiple prey traits , multiple predators : keys to understanding complex community dynamics .
j. sea res .
49 , 143 - 155 .
diehl , s. , feiel , m. , 2000 .
effects of enrichment on three - level food chains with omnivory .
naturalist 155 , 200 - 218 .
diehl , s. , feissel , m. , 2001 .
inraguild prey suffer from enrichment of their resources : a microcosm experiment with ciliates .
ecology 82(11 ) , 2977 - 2983 .
finke , d. l. , denno , r. f. , 2002 .
intraguild predation diminishes in complex - structured vegetation : implication for prey suppression .
ecology 83 , 643 - 652 .
gardner , m. r. , ashby , w. r. , 1970 .
connectance of large dynamic ( cybernetic ) systems : critical values for stability .
nature 228 , 784 .
gross , t. , rudolf , l.,levin , s. a. , dieckmann , u. , 2009 .
generalized models reveal stabilizing factors in food webs . science .
325 , 747 - 750 .
holt , r. d. , grover , j. , tilman , d.,1994 .
simple rules for interspecific dominance in systems with exploitative and apparent competiton .
naturalist 144 ( 5 ) , 741 - 77 .
holt , r. d. , polis , g. a. , 1997 .
a theoretical framework for intraguild predation .
naturalist 149 , 745 - 764 .
holt , r.d . ,
huxel , g. r. , 2007 .
alternative prey and the dynamics of intraguild predation : theoretical perspectives .
ecology 88 , 2706 - 2712 .
hosack , g. r. , li , h. w. , rossignol , p. a. , 2009 .
sensitivity of system stability to model structure .
modelling 220 , 1054 - 1062 .
hutchinson , g. e. , 1961 .
the paradox of the plankton .
naturalist 95 ( 882 ) , 137 - 145 .
ives , a. r.,carpenter , s. r. , 2007 .
stability and diversity of ecosystems .
science 317 , 58 - 62 .
janssen , a. , sabels , m. w. , magalhaes , s.,montserrat , m.,van der hammen , t.,2007 .
habitat structure affects intraguild predation , ecology 88(11 ) , 2713 - 2719 .
kivan , v. , 1996 . optimal foraging and predator - prey dynamics .
49 , 265 - 290 .
kivan , v. , diehl , s. , 2005 .
adaptive omnivory and species coexistence in tri - trophic food webs .
67 , 85 - 99 .
leibold , m. a. , 1996 . a graphical model of keystone predators in food webs : trophic regulation of abundance , incidence and diversity patterns in communities .
naturalist 147(5 ) , 784 - 812 .
leibold , m. a. , hall s. r. , bjornstad o. , 2005 . food web architecture and its effects on consumer resource oscillations in experimental pond ecosystems . in : de
ruiter , p. c. , moore , j. c.,wolters , v.(eds . ) , dynamic food webs .
elsevier , pp .
lima , s. l. , dill , l. m. , 1990 .
behavioral decisions made under the risk of predation : a review and prospectus . can .
68 , 619 - 640 .
lder , m. g. j. , boersma , m. , kraberg , a. c. , aberle , n. , shchekinova , e. , wiltshire , k. h. , xxxx .
even smallest competitors can promote each other : commensalism between marine microzooplankton predators ( submitted ) .
moore , j. c. , 2005 .
variations in community architecture as stabilizing mechanisms of food webs . in : de ruiter , p. c. , moore , j. c. , wolters , v. , ( eds . ) , dynamic food webs .
elsevier , pp .
24 - 26 mylius , s. d. , klumpers , k. , de roos , a. m. , persson , l. , 2001 .
impact of intraguild predation and stage structure on simple communities along a productivity gradient .
naturalist 158 , 259 - 276 .
namba , t. , tanabe , t. , maeda , n. , 2008 .
omnivory and stability of food webs .
complexity 5 , 73 - 85 .
oksansen , l. , fretwell , s. d. , arruda , j. , niemela , p. , 1981
exploitation ecosystems in gradients of primary productivity .
amer . naturalist .
118 , 240 - 261 .
pimm , s. l.,lawton , j. h. , 1978 . on feeding on more than one trophic level .
nature 275 ( 12 ) , 542 - 544 .
polis , g. a. , myers , c. a. , holt , r. d. , 1989 . the ecology and evolution of intraguild predation : potential competitors that eat each other .
ecol . system .
20 , 297 - 330 .
polis , g. a. , holt , r. d. , 1992 .
intraguild predation : the dynamics and complex trophic interactions , tree 7(5 ) , 151 - 154 .
rosenheim , j. a. , wilhoit , r. , armer , c. a. , 1993 .
influence of intraguild predation among generalist insect predators on the suppression of an herbivore population .
oecologia 96 , 439 - 449 .
stoecker , d. k. , evans , g. t. , 1985 .
effect of protozoan herbivory and carnivory in a microplankton food web , mar .
25 , 159 - 167 .
svirezhev , v. m. , logofet , d. o. , 1983 .
stability of biological communities .
mir publishers , moscow , ussr .
thomson , r. m. , hemberg , m. , starzomski , b. m. , shurin , j. b. , 2007 .
trophic levels and trophic tangles : the prevalence of omnivory in real food webs .
ecology 88(3 ) , 612 - 617 .
trussell , g. c. , ewanchuk , p. j. , bertness , m. d. , 2002 .
field evidence of trait - mediated indirect interactions in a pocky intertidal food web .
lett . 5 , 747 - 756 .
vadeboncoeur , v. , mccann , k. s. , vander zanden , m. j. , rasmussen , j. b. , 2005 .
effects of multi - chain omnivory on the strength of trophic control in lakes , ecosystems 8 , 682 - 693 .
vandermeer , j. , 2006 . omnivory and the stability of food webs .
238 , 497 - 504 .
woodward , g. , thompson , r. , townsend , c. r. , hildrew , a. c. , 2005 . pattern and process in food webs : evidence from running waters . in : belgrano ,
a. , scharler , u. m. , dunne , j. , ulanowicz , r. e. ( eds . ) , aquatic food webs : an ecosystem approach .
oxford university press , oxford , pp . | food webs with intraguild predation ( igp ) are widespread in natural habitats .
their adaptation and resilience behaviour is principal for understanding restructuring of ecological communities . in spite of the importance of igp
food webs their behaviour even for the simplest 3-species systems has not been fully explored .
one fundamental question is how an increase of diversity of the lowest trophic level impacts the persistence of higher trophic levels in igp relationships .
we analyze a 3-species food web model with a heterogeneous resources and igp .
the model consists of two predators directly coupled via igp relation and indirectly via competition for resource .
the resource is subdivided into distinct subpopulations .
individuals in the subpopulations are grazed at different rates by the predators .
we consider two models : an igp module with immobilization by the top predator and an igp module with species turnover .
we examine the effect of increasing enrichment and varying immobilization ( resource transfer ) rate on a stable coexistence of predators and resources .
we explore how the predictions from the basic 3-species model are altered when the igp module is extended to multiple resource subpopulations .
we investigate which parameters support a robust coexistence in the igp system . for the case of multiple subpopulations of the resource we present a numerical comparison of the percentage of food webs with stable coexistence for different dimensionalities of the resource community . at low immobilization ( transfer )
rates our model predicts a stable 3-species coexistence only for intermediate enrichment meanwhile at high rates a large set of stable equilibrium configurations is found for high enrichment as well .
intraguild predation , immobilization , alternative resource , multiple resource traits , stable coexistence 2 |
considerable experimental effort has been made in measuring proton - deuteron ( pd ) scattering @xcite cross sections at intermediate energies . for up to 300 mev proton energy those data
have been analyzed with rigorous three - nucleon ( 3n ) faddeev calculations @xcite based on the cd - bonn potential @xcite and the tucson - melbourne 3n force ( 3nf ) @xcite .
theoretical predictions based on 2n forces alone are not sufficient to describe the data above about 100 mev .
some of those defects are known as the sagara discrepancy @xcite .
though 3nf effects are already seen below 100 mev , they increase significantly above that energy .
however , presently available 3nf s only partially improve the description of cross section data and spin observables .
since most of the cited calculations are based on the non - relativistic formulation of the faddeev equations @xcite , one needs to question if in the intermediate energy regime a poincar invariant formulation is more adequate .
there are two basic approaches to a relativistic formulation of the 3n problem .
one is a manifestly covariant scheme linked to a field theoretical approach @xcite , the other is based on an exact realization of the symmetry of the poincar group in three nucleon quantum mechanics @xcite .
we employ the second approach , where the mass operator ( rest energy operator ) consists of relativistic kinetic energies together with two- and many - body interactions including their boost corrections @xcite .
the first attempt in solving the relativistic faddeev equation for the 3n bound state based on second approach has been carried out in @xcite , resulting in a decrease of the binding energy compared to the nonrelativistic result . on the other hand , similar calculations based on the field theory approach @xcite
increase it .
these contradictory results require more investigation . in the following we
summarize the results of our calculations based on the second approach : in section 2 we introduce the relativistic 2n potential , in section 3 we present the 2n t - matrix , which fulfills the relativistic boosted lippmann - schwinger ( ls ) equation , and in section 4 we give numerical results for the triton binding energy based on the poincar invariant faddeev equation .
modern meson theoretical nn potentials , e.g. charge dependent bonn potential ( cd - bonn ) @xcite , are derived from a relativistic lagrangian , then cast into a three - dimensional form using the blankenbeclar - sugar equation , which by kinematical redefinitions can be written in the form of a standard nonrelativistic ls equation , which in partial wave decomposed form reads @xmath1 the corresponding relativistic ls equation is given as @xmath2 where @xmath3 in the relativistic faddeev equation one needs @xmath4 off - the - energy - shell . according to @xcite there is a direct operator relation between the nonrelativistic @xmath5 and the relativistic @xmath6 : @xmath7 in a momentum representation this leads to @xmath8 this is the nonlinear relation between the relativistic potential @xmath6 and the nonrelativistic potential @xmath5 from eq .
( [ eq1 ] ) , which has recently been solved @xcite .
the resulting on - shell - t - matrix @xmath4 is on - shell identical to the t - matrix @xmath9 from eq .
( [ eq1 ] ) .
cluster properties require that the energy is additive . because of the non - linear relations between the mass and energy in special relativity
, the additivity of energies in the rest frame implies a non - linear relation between the two - body interactions in the two and three - body mass operators @xcite .
we call the two - body interaction in the three - body mass operator the `` boosted potential '' .
@xmath10 ^ 2 + q^2 } -\sqrt{\left [ 2 \sqrt{{\hat p}^2 + m^2}\right]^2 + q^2 } , \label{eq8}\end{aligned}\ ] ] where the spectator momentum @xmath11 in the 3-body center of mass is simultaneously the negative total momentum of the pair . using eq .
( [ eq6 ] ) this can be rewritten as @xcite @xmath12 thus one can obtain @xmath13 by the same technique @xcite as @xmath14 .
the boosted off - shell t - matrix is the solution of the ls equation @xmath15 with @xmath16 . in fig .
[ figtmat ] we display the boosted half - shell ( @xmath17 ) t - matrix of the cd - bonn @xcite potential at @xmath18=350 mev for three different spectator momenta @xmath11 .
the magnitude of the t - matrix gradually decreases with increasing the boost momentum @xmath11 .
it can be shown @xcite that the half - shell t - matrices @xmath19 and @xmath20 are related by simple factors @xmath21 solving eqs .
( [ eq5 ] ) and ( [ eq10 ] ) to obtain @xmath4 and @xmath22 independently we numerically confirmed the relation ( [ eq11 ] ) with high precision .
it is the factor on front of @xmath4 in the right hand side of eq .
( [ eq11 ] ) which attenuates the amplitude of the t - matrix with increasing @xmath11 .
it can also be shown @xcite that the relativistic half - shell t - matrix @xmath4 is related to the corresponding nonrelativistic one via @xmath23 the explicit construction of first @xmath6 and then @xmath4 is equivalent to obtaining @xmath4 via resolvent equations as suggested in @xcite and carried out in @xcite .
the relativistic bound state faddeev equation was solved using the boosted t - matrix @xmath24 . in table
[ table1 ] the results for the triton binding energy using the cd - bonn potential as input are shown .
the triton binding energy obtained from the relativistic calculation is about 100 kev smaller compared to the one calculated nonrelativistically .
this value is significantly smaller than a previously published result @xcite in which a reduction of the triton binding energy by about 400 kev was given .
the reason for this overestimation of a relativistic effect on the binding energy can be attributed to a different construction of the relativistic off - shell t - matrix @xmath4 .
the scaling transformation employed in @xcite does not keep the 2n scattering data invariant as function of the 2n c.m .
momentum .
[ table1 ] we also included the wigner spin rotation as outlined in @xcite .
thereby the the balian - brezin method@xcite in handling the permutations is quite useful . in table
[ table2 ] the triton binding energies are shown allowing charge independence breaking ( cib)@xcite * ? ? ?
* and wigner spin rotations .
wigner spin rotation effects reduce the binding energy by only about 2 kev .
a phase - shift equivalent 2n potential @xmath25 in the relativistic 2n schrdinger equation is related to the potential @xmath5 in the nonrelativistic schrdinger equation by the nonlinear relation given in eq .
( [ eq6 ] ) .
the boosted potential @xmath26 is related to @xmath25 by a similar expression , eq .
( [ eq8 ] ) . with these potentials
we generate the relativistic fully - off - shell t - matrix @xmath27 , which enters into the relativistic faddeev equation .
we solve the relativistic bound state faddeev equation and compare the binding energy for the triton with the one obtained from a nonrelativistic calculation with the same input interaction .
we find that the difference between the two calculations is only about 90 kev including cib , where the relativistic calculation gives slightly less binding . taking wigner spin rotations into account in the relativistic calculation reduces the binding energy by a very small amount , @xmath28 2 kev , indicating that wigner rotations of the spin have essentially no effect on the predicted value of the binding energy .
applications to the 3-body continuum are in progress .
recently @xcite the formulation lined out above has been used to study the low energy @xmath29 puzzle in neutron - deuteron scattering .
details are presented by witaa in this conference . in the intermediate energy regime
the formulation has been applied to exclusive proton - deuteron scattering cross sections at 508 mev @xcite based on a formulation of of the faddeev equations which does not employ a partial wave decomposition .
the approach can also be extended and applied to electromagnetic processes@xcite .
this work was partially supported by the 2008 - 2011 polish science funds as a research project no .
n n202 077435 .
it was also partially supported by the helmholtz association through funds provided to the virtual institute `` spin and strong qcd''(vh - vi-231 ) .
the numerical calculations were performed on the ibm regatta p690 + of the nic in jlich , germany .
r. machleidt , f. sammarruca , y. song , _ phys .
c _ * 53 * , r1483 ( 1996 ) .
s. a. coon , w. glckle , _ phys .
c _ * 23 * , 1790 ( 1981 ) .
k. sagara _
c _ * 50 * , 576 ( 1994 ) .
s. nemoto _
c _ * 58 * , 2599 ( 1998 ) .
h. witaa _
lett . _ * 81 * , 1183 ( 1998 ) . | the phase equivalent relativistic nn potential , which is related by a nonlinear equation to the original nonrelativistic potential , is used to construct the mass operator ( rest hamiltonian ) of the 3-nucleon system . employing the cd bonn nn potential , the solution of the relativistic 3n faddeev equation for @xmath0h shows slightly less binding energy than the corresponding nonrelativistic result .
the effect of the wigner spin rotation on the binding is very small . |
first described nearly fifty years ago @xcite , the phenomenon of hidden mechanical momentum continues to generate interest @xcite .
the basic idea is simple . in flat spacetime ,
any time - independent , conserved stress - energy @xmath1 must have zero total momentum @xmath2 ( when this integral is defined ) .
for matter coupled to electromagnetic fields , we then have for stationary configurations that the total mechanical plus electromagnetic field momentum is zero , @xmath3 ( here the total stress energy @xmath4 is given by a matter component @xmath5 and the field component @xmath6 , and the momenta @xmath7 and @xmath8 are defined in terms of their respective stress tensors . ) however , using stationary sources it is easy to create stationary field configurations whose field momentum is non - zero .
thus , the sources contain non - zero momentum
there is mechanical momentum in bodies at rest !
this is the hidden momentum .
the original and simplest example is a magnetic dipole immersed in an electric field .
consider a point dipole of strength @xmath9 located at a position @xmath10 in a stationary electric field @xmath11 assumed to vanish at infinity .
then the field momentum is @xmath12 the first equality is the definition of @xmath13 , the second uses @xmath11 , the third integrates by parts , the fourth uses the maxwell equation @xmath14 , and the fifth uses the point - dipole current @xmath15 . by equation ,
the hidden mechanical momentum is then @xmath16 equation gives the mechanical momentum present in a stationary dipole immersed in an electromagnetic field .
how is this momentum stored ? for a simple model of a current loop dipole , the electric field modifies the flow of current in such a way that the ( relativistic ) particle momenta of the charge carriers sums precisely to the simple formula above @xcite .
the above discussion applies to stationary situations only , and it is natural to ask the fate of the hidden momentum if the fields are turned off or the configuration is otherwise destroyed by a time - dependent process . is some ( or all ) of the hidden mechanical momentum converted into ordinary momentum , resulting in a kick ( balanced by emitted radiation ) ? or does the system just relax to the trivial configuration containing a body at rest ?
the process is difficult to model , since one must include the agent turning off the fields as part of the calculation , and the result will in general depend on the details of the turning off process . to our knowledge
, no such calculation has yet been carried out .
a recent study of hidden momentum @xcite concluded that while some impulse may be delivered , `` there is no obvious reason why this impulse should equal the momentum originally stored in the fields all we can say in general is that the total momentum afterward , like the total momentum before , is zero . '' in this paper
we will determine the fraction of momentum that is converted into impulse in one particular scenario collapse to a black hole .
specifically , we will consider a point electric charge placed near a point magnetic dipole , both surrounded by a shell of matter assumed large and light enough that its gravity does not ( initially ) distort the electromagnetic field configuration .
thus , initially , we have a perfectly ordinary hidden momentum configuration .
however , at a time @xmath17 the shell begins to collapse at the speed of light , beginning the process of black hole formation . by the `` no hair theorem ''
the final black hole can only support a simple monopolar electric field , so the final configuration has no hidden momentum and no field momentum . by monitoring the momentum radiated ( relative to the rest frame of the black hole ) we determine that a kick corresponding to @xmath0 of the initial stored momentum is delivered to the black hole . to perform these calculations
we adapt a model due to @xcite , which treats the electromagnetic fields as small perturbations to the collapse spacetime and exploits causality to simplify the calculations .
although the `` form a black hole '' method of turning off the fields may seem unduly complicated ( or completely out of the blue ) , this type of trick makes it likely to be one of the _ simplest _ situations to model .
a completely separate motivation for studying black hole formation in the presence of hidden momentum comes from numerical simulations of black hole mergers .
a surprising discovery @xcite was that for certain configurations of black hole binaries containing spinning black holes , there is a bobbing up and down of the entire orbital plane that correlates to a large kick velocity obtained at merger .
in previous work @xcite , this behavior was explained using the concept of hidden momentum . by revealing the bobbing to be associated with a kinematical effect of spin that is present even for systems for which there can be no kick ( an example is spinning balls connected by a string )
, the authors of @xcite concluded that the kick should not be viewed as an inertial continuation of the bobbing .
but what , then , causes the kick ? it was noted that , at least in an electromagnetic analog of orbiting charged magnetic dipoles , there is an orbitally modulated field momentum present outside the bodies , along with corresponding hidden momentum within the bodies .
the authors suggested that a rapid merger event allows the release of some of this field momentum , leading to a kick that will depend sinusoidally on the phase at merger , as observed in numerical simulations . while this picture accounts qualitatively for the observed behavior
, there was no quantitative estimate of the percentage of the field momentum that a merger could release .
furthermore , there was no means of distinguishing the proposal from a competing scenario also qualitatively successful in which the kick is caused by the bobbing motion s modulation of the dominant orbital emission @xcite .
the calculations of this paper provide an estimate of the efficiency of the momentum release mechanism in a context where the competing mechanism is absent , thereby helping to distinguish the two scenarios .
the implications for our understanding of black hole merger kicks are discussed in section [ sec : discussion ] .
r0.4 in their early investigation of the fate of perturbing fields during gravitational collapse , de la cruz , chase , and israel @xcite introduced the null shell collapse spacetime pictured in figure [ fig : collapse ] .
the spacetime is constructed by joining flat spacetime ( region i ) to schwarzschild spacetime ( regions ii , iii , iv ) on a spherical hypersurface ( blue line ) that is initially tangent to the direction of time symmetry ( i.e. , stationary ) , but at a later time suddenly becomes tangent to the ingoing radial null direction ( collapsing at the speed of light ) .
following the standard junction conditions @xcite , the induced metric is required to match on the shell , and the jump in extrinsic curvature is interpreted as the stress - energy of the shell .
this spacetime models a sudden and rapid gravitational collapse of a shell of matter . in the words of the authors of @xcite ,
the model is `` highly artificial from an astrophysicist s point of view , but does not violate any of the principles of relativity theory . ''
we work in coordinates @xmath18 with @xmath19 the direction of time symmetry , @xmath20 the areal coordinate for symmetry two - spheres , and @xmath21 spherical coordinates on those spheres .
we take the collapse to begin at coordinate time @xmath22 . for @xmath23 ,
the metric is that of the well - studied static spherical shell , given in our coordinates by @xmath24 where @xmath25 is the metric of the two - sphere . here
@xmath26 is the initial coordinate radius of the shell .
however , we will take @xmath27 to be very large , @xmath28 so that for @xmath23 the spacetime is simply minkowski spacetime everywhere .
on this minkowski spacetime we place a point charge @xmath29 and a point magnetic dipole @xmath30 .
the magnetic dipole is located at the origin and is oriented in the @xmath31 ( @xmath32 ) direction , while the charge is located on the @xmath33 axis ( @xmath34 ) at a distance @xmath35 from the origin .
( we require @xmath35 so that the charge ends up inside the black hole . ) in a suitable gauge the non - zero components of the vector potential are @xmath36 where we have expanded the point charge field @xmath37 in spherical harmonics @xmath38 ( with the conventions of @xcite ) . here
@xmath39 denotes the lesser ( greater ) of @xmath20 and @xmath40 , and an overbar denotes complex conjugation . by assumption , equations and give the fields for @xmath23 .
however , by causality they must also give the fields in the entirety of regions i and ii , and , by continuity of the electromagnetic field , can therefore provide characteristic and @xmath41 . ] initial data for an evolution into the remaining portions of the spacetime . since our main concern is with the momentum radiated to infinity
, we confine the evolution to region iii , the schwarzschild exterior .
we use double - null coordinates , related to schwarzschild coordinates @xmath42 by @xmath43 where @xmath44 is the tortoise coordinate . the two - sphere at the onset of collapse ( meeting point of regions i , ii , iii ) is described by @xmath45 and @xmath46 . for @xmath47
, @xmath41 is the collapsing shell s surface , while @xmath48 represents a spherical shock front propagating outward at the speed of light , bringing news of the sudden onset of collapse .
the ( future ) event horizon is located at @xmath49 ( fix @xmath50 ) , while @xmath51 ( fix @xmath52 ) represents future null infinity @xmath53 , the vantage - point of a distant observer receiving the radiation .
the shell , shock front , horizon , and null infinity bound our evolution region iii , @xmath54 and @xmath55 .
the initial data is given on the shell and shock front , @xmath41 and @xmath48 . to solve the time - dependent maxwell equation on the schwarzschild background
, we will use the master function approach @xcite as formulated in @xcite .
the vector potential is expanded into four scalar amplitudes , @xmath56 , @xmath57 and @xmath58 , @xmath59 where indices @xmath60 run over the coordinates @xmath61 and indices @xmath62 run over the coordinates @xmath63 . here
@xmath64 is the even parity vector spherical harmonic , while @xmath65 is the odd parity vector spherical harmonic .
( the levi - civita tensor @xmath66 is that of the unit two - sphere ( @xmath67 ) , and the index is raised with the metric of the unit two - sphere . )
one then defines master functions @xmath68 in terms of the amplitudes by @xmath69 and , using maxwell s equations , it turns out that these functions entirely determine the field strength by @xmath70 back in terms of double - null coordinates , maxwell s equations imply that the master functions satisfy @xmath71 equation applies only to the modes with @xmath72 ; the @xmath73 mode is non - radiative and requires a separate treatment .
( the non - radiative character follows immediately from gauss law .
since our interest is in the radiation , we do not consider this mode .
, @xmath74 , is also a solution in schwarzschild spacetime ( when areal coordinates are identified , as they are in our shell metric ) .
this observation allows one to see explicitly that the @xmath73 mode does not evolve when the shell begins to collapse . ] ) note that equation is the same for all @xmath75 and each choice of @xmath76 or @xmath77 .
the initial data for our evolution is provided by the static solutions , equations - . since @xmath35 ,
the initial data surfaces @xmath48 and @xmath41 satisfy @xmath78 , and we require the point charge solution only for @xmath78 ( i.e. , @xmath79 ) .
computing the master functions associated with the initial data then gives @xmath80 here we have used the fact that @xmath81 .
equations - provide initial data on @xmath48 and @xmath41 for the evolution of equation . in order to avoid redundant integrations over different @xmath75-values ( which satisfy the same wave equation ) , it is convenient to introduce real amplitudes @xmath82 such that @xmath83 the amplitudes satisfy the same wave equation , @xmath84 with initial values given by @xmath85 our main interest in this evolution is the total energy and ( especially ) momentum radiated away . at future null infinity ,
@xmath51 , the spacetime is flat and we may compute the total flux of energy and momentum using expressions valid in minkowski spacetime .
let @xmath86 and @xmath87 ( @xmath88 ) be the time and space translation killing fields of flat spacetime .
( in cartesian minkowski coordinates , we have @xmath89 and @xmath90 . ) these vector fields , when dotted into a conserved stress - energy tensor , give the associated current of energy and momentum , which can be integrated over the surface @xmath51 to give the total radiated energy and ( @xmath91 component of ) momentum . for the total electromagnetic energy and momentum lost through future null infinity in our spacetime
, we have @xmath92 where @xmath63 range over the two - sphere ( area element @xmath93 ) and @xmath52 runs from @xmath94 to @xmath95 .
( however , for our field configuration the flux vanishes for @xmath96 and one may begin the integral at @xmath48 . ) here @xmath1 is the stress - energy tensor of the electromagnetic field , @xmath97 the radiated energy takes an elegant form in terms of the master functions ; using the orthogonality of the spherical harmonics it is straightforward to show that @xmath98 where the vertical bars denote complex norm . in deriving this equation and analogous flux equations below ,
we have used the fact that as @xmath51 the general solution of equation is the sum of a pure function of @xmath52 and a pure function of @xmath50 . in terms of our real amplitudes , we have @xmath99 the general expression for the radiated momentum ( analogous to equation for the energy ) will be less elegant and more laborious to derive , since it will involve clebsch - gordon - type coupling coefficients .
( each spatial killing vector @xmath100 is @xmath101 , so that triple products of spherical harmonics appear under the integral . ) however , from the symmetry of the problem and the `` addition of angular momentum '' rules , it is straightforward to establish the following : ( a ) the radiated momentum in the @xmath33 direction takes the form of a coupling between even parity modes with @xmath102-values separated by 1 ; ( b ) the radiated momentum in the @xmath103 direction will come entirely from the @xmath101 even - parity mode multiplied by the @xmath101 odd - parity mode ; and ( c ) the radiated momentum in the @xmath31 direction is vanishing . evaluating the angular integrals explicitly for low values of @xmath102 , we have @xmath104 the radiated momentum in the @xmath33 direction would persist in the absence of the magnetic dipole and can be thought of as resulting from the asphericity of the collapse of a shell onto an off - center point charge .
this asphericity will result in a slight asphericity of the emitted radiation , leading to a kick . on the other hand ,
the radiation in the @xmath103 direction requires both the dipole and the charge and may be thought of as due to a release of the initial field momentum ( entirely in the @xmath103 direction ) during the rapid collapse event . since the initial data for each real amplitude @xmath82 scales as @xmath105
, we can anticipate that this will be the dominant contribution , i.e. , @xmath106 .
( we find that @xmath107 . )
furthermore , it is clear that we lose little accuracy by dropping terms higher order in @xmath102 , represented by the @xmath108 in equation .
( we find that the second term is smaller than the first by a factor of about @xmath109 for order unity parameters . )
it is clear that we may therefore neglect the third and higher terms , already not displayed in equation .
some further simplification is helpful for numerical purposes and for revealing the scaling of the results with the parameters of the model .
since @xmath110 and @xmath111 are proportional initially ( see equations and ) and satisfy the same linear wave equation , they are in fact proportional for all time .
we may thus keep track of both @xmath101 amplitudes by a single fiducial amplitude @xmath112 .
we will define @xmath112 to be the numerical function obtained by integrating the wave equation with @xmath101 and @xmath113 and with initial data given by @xmath114 ( on @xmath48 and @xmath41 in the limit of a large shell , @xmath115 ) .
using equations - , the momentum flux formula becomes @xmath116 where we define the fiducial integral @xmath117 by @xmath118 with the factor @xmath119 introduced for convenience .
the scaling of @xmath120 with @xmath30 , @xmath29 , and @xmath40 is evident from equations and , while the factor of @xmath121 could have been guessed on dimensional grounds . and
@xmath122 , which agree with @xmath52 and @xmath50 in the fiducial case @xmath113 . ]
the integral @xmath117 is simply a number that may be computed numerically .
from equation we have that @xmath117 is manifestly positive , so that the kick is guaranteed to be in the direction of the initial hidden momentum , as expected .
we find numerically that @xmath123 ( see section [ sec : numerical ] ) . by conservation of momentum ,
the net kick delivered to the final black hole is @xmath124 from equations - , the initial field configuration has momentum @xmath125 ( counterbalanced by hidden mechanical momentum @xmath126 ) .
we may define the efficiency @xmath127 of the hidden momentum kick as the fraction of this field momentum that is radiated away during the collapse . from equation
we have @xmath128 thus for separation distances @xmath40 of order the mass @xmath129 of the shell , the efficiency is about @xmath0 . for the maximum
allowed separation distance , @xmath130 , the efficiency is @xmath131 .
if we keep only the @xmath101 terms in the energy flux , then equation shows that the fiducial integral @xmath117 also controls the energy flux , @xmath132 we find numerically that the @xmath133 term is a correction of fractional order @xmath134 , while the @xmath135 term is a correction of fractional order @xmath136 .
thus equation should give an excellent approximation to the energy flux .
finally note that we may divide equations and to obtain @xmath137 the fiducial integral has canceled out , and we have a fully analytic result for the ratio of the @xmath103-momentum flux to the dominant energy flux .
we see that the ratio is controlled by a symmetric combination of the initial magnetic dipole moment @xmath30 and electric dipole moment @xmath138 , reaching a maximum of @xmath139 when the moments are equal .
equation could also have been derived directly from equations and using the fact that @xmath140 . to summarize , we first presented our problem in terms of a general master - function formulation of maxwell s equations on a schwarzschild background .
then we reduced that prescription to one involving real amplitudes @xmath82 , whose initial values are given in equations and . once equation is integrated with these initial values
, the fluxes of energy and momentum are given by equations - .
finally , we introduced a fiducial @xmath112 , corresponding to integrating equation with @xmath113 and initial data @xmath141 .
this integration should be carried out with initial shell radius taken to be very large , @xmath115 .
the integral @xmath117 , defined by equation in terms of this @xmath112 , then contains all the information about the dominant energy flux @xmath142 ( equation ) and momentum flux @xmath120 ( equation ) for any choice of model parameters @xmath143 .
we evolve equation using a 2nd order finite difference discretization , @xmath144 where the indices @xmath145 refer to north , south , west , east on the grid as in reference @xcite , @xmath146 is the grid spacing , and @xmath147 is approximated by @xmath148 and @xmath149 . we evolve equation with @xmath113 on a grid of @xmath150 $ ] and @xmath151 $ ] with initial data @xmath152 for @xmath153 . in principle
our model requires @xmath154 , @xmath155 and @xmath156 , and we settle on values @xmath157 , @xmath158 and @xmath159 as sufficiently large ( see below for error discussion ) . figure [ fig : field ] shows the result of these evolutions with @xmath112 evaluated on @xmath160 , approximating future null infinity @xmath53 .
we performed each simulation for three different resolutions with grid spacing of @xmath161 .
the dashed line shows the convergence factor @xmath162 , which should be @xmath163 for 2nd order convergence .
notice that the waves reach null infinity ( @xmath164 ) around @xmath165 , peak around @xmath166 , and decay by about @xmath167 . since
our shell is at @xmath168 , these retarded times correspond ( respectively ) to emission from @xmath169 , or @xmath170 ( initial emission ) , @xmath171 ( peak of emission ) and @xmath172 ( end of emission ) .
thus the vast majority of the emission comes from the strong - field regime . as a function of @xmath52 on the grid boundary @xmath173 which approximates @xmath53 .
the field ( left scale ) is evolved for the three different cases @xmath153 .
the dashed line shows the convergence factor @xmath174 ( right scale ) and demonstrates good 2nd order convergence . ] the @xmath103 momentum flux and the dominant energy flux are both captured by the fiducial integral @xmath117 , equation , which is computed from the @xmath101 evolution .
we find @xmath175 in establishing this result we have quantified three sources of error : finite resolution ( @xmath176 ) , finite extraction radius ( @xmath177 ) , and finite initial shell radius ( @xmath178 ) .
our resolution error estimate is done via richardson extrapolation . for the finite extraction radius error , we vary @xmath160 and fit a function of the form @xmath179 in @xmath180 .
( the grid is kept square , so that @xmath181 varies as well . ) for the initial shell radius error we similarly fit a function of the form @xmath182 in @xmath183 .
the value @xmath117 quoted above is extrapolated in @xmath160 , but not in resolution @xmath146 or @xmath184 . from the @xmath133 and @xmath135 evolutions
we also compute the higher order corrections to the radiated energy , equation , and the radiated @xmath33 momentum , equation . for order unity parameters ,
the subleading energy term is a fractional correction of order @xmath185 , while the @xmath33 momentum flux is smaller than the @xmath103 momentum flux by a factor of order @xmath186 .
the text below equation gives the precise numbers , along with parameter scalings . as a further check of our code
we reproduce the results of the numerical run done in reference @xcite .
in particular , for their parameter choices of @xmath187 and @xmath188 , we are able to reproduce their integral @xmath189 ( see text below their equation ( 2 ) ) to the accuracy given in the paper .
in this paper we have explored the effect of dynamical processes on hidden momentum in one particular scenario involving gravitational collapse .
the main result may be presented as either a kick velocity of the final black hole or an efficiency of the conversion of hidden momentum to ordinary momentum by the collapse ( see equations - ) , @xmath190 in natural units .
the efficiency of the process scales with the ratio of the size of the charge - and - dipole configuration , @xmath40 , to the final black hole size @xmath191 .
if the mass @xmath129 of the shell is very large , so that the eventual black hole is much larger than the charge - and - dipole , then the kick becomes small .
this stands to reason , since the black hole engulfs both the hidden momentum within the dipole and the ( equal and opposite ) field momentum stored outside , leading to zero net kick . on the other hand , when the size of the charge - and - dipole approaches the size of the final black hole , the black hole swallows only a portion of the field momentum , leading to a net kick .
we see no a priori way to estimate the percentage swallowed , which in principle could have been negligible .
our numerical calculations have determined it to be about @xmath0 , which leads to the prefactor of 400 km / s when the model parameters are expressed in gaussian geometrized units .
the @xmath192 factor will be very small for ordinary laboratory and astrophysical systems ( see section [ sec : gen ] for some discussion ) , and we are not attempting to apply our model to presently observable electromagnetic phenomena .
rather , we are primarily concerned with exploring hidden momentum in a dynamical context , and our particular calculations show that its sudden destruction can lead to a kick .
a second goal of the paper is to use the electromagnetic problem to help elucidate the mechanism behind the bobbing - and - kicks behavior seen in binary black hole mergers . to compare directly to the kicks observed in binary black hole simulations we must make several extrapolations .
first , we must use the analogy between gravitation and electromagnetism , involving @xmath193 and @xmath194 , where @xmath75 and @xmath195 are the mass and spin of a black hole in the ( comparable mass , comparable spin ) binary .
second , since the shell of mass @xmath129 enforces our merger and produces a final black hole of mass @xmath196 , we should also take @xmath197 to approximate the merger of black holes of mass @xmath198 .
third , to approximate the superkicks premerger configuration , we should take the distance between the particles to also be of order the mass / size of the black holes , @xmath199 , and set the spin to be of order the maximal value , @xmath200 .
finally , we must extrapolate the kick formula beyond the test - maxwell - field regime where it is strictly valid , since the above parameter identifications entail @xmath201 and @xmath202 .
these extrapolations are crude , and a combination of order - unity errors could easily migrate the answer by an order of magnitude .
we therefore take the numerical value of 400 km / s only as a rough ballpark for the size of kicks that a hidden momentum kick mechanism could generate .
no such estimate was made in the original work proposing this mechanism @xcite , and our investigation could in principle have revealed the effect to be negligible . that our figure is within an order of magnitude of the observed superkicks of 4000 km / s gives plausibility but by no means definitiveness to the hidden momentum kick mechanism .
we therefore believe that the hidden momentum kick mechanism may have a role to play in explaining the superkick phenomenon , and could in principle account for the entire effect .
however , an alternative proposal @xcite also predicts kicks of the correct qualitative and roughly the correct quantitative character . in this scenario
one imagines that the bobbing motion of the black holes provides a small modulation to the dominant emission ( due to the quasi - circular motion ) , which induces an asymmetry in that emission , resulting in the kick .
since the bobbing motion is of direct relevance in producing the kick in this scenario , we will call this the `` motion scenario '' .
this scenario does not rely on the presence of a merger to generate the kick , and a number of simulations of grazing collisions of black holes ( i.e. , near the threshold between capture / merging and scattering ) @xcite have indicated that large kicks can be obtained even in the non - merging case . however , in these simulations the black holes complete at least a full orbit before separating @xcite , and one could therefore still view the kick as a consequence of release of quasi - static field momentum due to the sudden separation of the bodies .
thus the simulations significantly inform the discussion but do not conclusively distinguish between the two scenarios .
the work reported here helps further distinguish the scenarios by isolating the relevant physics of our scenario ( a rapid event destroying the quasi - static configuration ) and creating a situation where the relevant physics of the motion scenario is entirely absent .
indeed , our `` binary '' consists of particles at _ rest _ , with the emission entirely determined by the merger s destruction of the initially static field configuration . that we can recover kicks in the ballpark of those seen in numerical simulations _ without any orbital motion at all _
suggests that the rapid destruction of the quasi - static field configuration may indeed be playing an important role in producing the superkicks .
however , the uncertainty inherent in the extrapolation of our results to the gravitational case leaves plenty of room for the effect to be too small to fully account for the superkick .
the motion effect could be dominant , or both effects could be playing an important role .
the original paper on null - shell collapse in the presence of perturbing fields @xcite treated a gravitational quadrupole in addition to an electromagnetic dipole .
this raises the question of why a purely gravitational calculation was not attempted here .
the answer is that the equivalence principle forbids a static binary in general relativity all gravitating bodies attract . a gravitational binary
can only be maintained with orbital velocity , and so it is difficult to cleanly separate effects due to motion from those due to merger or collapse .
studying an electromagnetic analog enables us to completely eliminate motion while preserving the other essential features of the pre - merger superkicks configuration .
we conclude with a brief discussion of other physical and astrophysical systems in which hidden momentum may have a role to play .
first , since the phenomenon is basically independent of the internal structure of the body ( being enforced by the presence of quasi - static field momentum and the conservation of total momentum ) , similar kick phenomena are to be expected for mergers and near - merger scattering of neutron stars , boson stars , or any other object that could be simulated . however , the kick will be smaller for bodies that are less compact .
returning to the electromagnetic case where the notion of hidden momentum is most precisely defined , one could imagine that a magnetized neutron star moves through an external magnetic field , generating an `` external '' electric field in the local rest frame .
this would be an astrophysical example of the original hidden momentum setup , and our heuristic would predict a kick if the neutron star underwent collapse .
however , even assuming the strongest known magnetar moving through the field of another such object at an orbital separation of ten times the radius , the total hidden momentum corresponds to a kick of only about a millimeter per second .
furthermore , plasma dynamics would likely be important for this system .
having a large electromagnetic hidden momentum kick requires compact objects with strong intrinsic electric and magnetic fields .
one possible avenue for further exploration would be the addition of electric charge to the spinning black holes already considered , which generates an effective magnetic dipole moment as well . in this case
there should be significant hidden momentum of both the electromagnetic and gravitational variety .
99 w. shockley and r.p .
james , _ phys
lett . _ * 18 * 876 ( 1967 )
. v. hnizdo _ am .
j. phys . _ * 65 * 515 - 18 ( 1997 ) d. babson , s. reynolds , r. bjorkquist and d. griffiths , _ am
. j. phys .
_ * 77 * 826 ( 2009 ) s. gralla , a. harte , and r. wald , _ phys
d _ * 81 * 104012 ( 2010 ) l. costa , c. herdeiro , j. natario , and m. zilhao , _ phys .
* 85 * 024001 ( 2012 ) see a number of worked examples by kirk mcdonald , available on the arxiv and at http://www.physics.princeton.edu/.17exmcdonald / examples/. d. griffiths and v. hnizdo , arxiv:1303.0732 ( 2013 ) d. j. griffiths , _ introduction to electrodynamics _
( prentice - hall , new jersey , 1999 ) v. de la cruz , j. chase and w. israel , _ phys . rev .
* 24 * 423 ( 1970 ) m. campanelli , c. lousto , y. zlochower , and d. merritt , _ ap . j. _ * l5 * , 659 ( 2007 ) m. campanelli , c. lousto , y. zlochower , and d. merritt , _ phys . rev .
_ * 98 * , 231102 ( 2007 ) j. a. gonzalez , u. sperhake , b. brgmann , m. hannam and s. husa , _ phys
lett . _ * 98 * , 091101 ( 2007 ) f. pretorius , in relativistic objects in compact binaries : from birth to coalescence , edited by colpi et al . ,
springer verlag ( 2009 ) , arxiv:0710.1338 [ gr - qc ] w. israel , _
nuovo cim .
b _ * 44 * 1 ( 1966 ) e. poisson , _ a relativist s toolkit _ ( university press , cambridge , 2004 ) c. cunningham , r. price , and v. moncrief , _ ap
. j. _ * 224 * 643 - 667 ( 1978 ) c. cunningham , r. price , and v. moncrief , _ ap . j. _ * 230 * 870 - 892 ( 1979 ) r. haas , _ phys
d _ * 75 * , 124011 ( 2007 ) j. d. jackson , _ classical electrodynamics _
( john wiley & sons , new york , 1999 ) j. thornburg , _ grg _ * 43 * 1211 - 1251 ( 2011 ) r. gomez and j. winicour , _ j. math .
physics _ * 33*(4 ) 1445 ( 1992 ) u. sperhake , e. berti , v. cardoso , f. pretorius and n. yunes , _ phys .
* 83 * 024037 ( 2011 ) u. sperhake , _ private comment _ | a stationary magnetic dipole immersed in an electric field carries `` hidden '' mechanical momentum .
however , the fate of this momentum if the fields are turned off is unclear .
we consider a charge - and - dipole hidden momentum configuration , and turn off the fields by collapsing a null shell onto the system , forming a black hole . in numerical calculations
we find that the black hole receives a kick corresponding to @xmath0 of the initial stored momentum .
when extrapolated to apply to purely gravitational phenomena , this efficiency suggests a role for the hidden momentum kick mechanism in generating the binary black hole `` superkicks '' observed in numerical simulations of einstein s equation . |
let @xmath9 be a finite undirected graph . a vertex @xmath10 _ dominates _ itself and its neighbors .
a vertex subset @xmath11 is an _ efficient dominating set _ ( _ e.d.s . _ for short ) of @xmath1 if every vertex of @xmath1 is dominated by exactly one vertex in @xmath0 ; for any e.d.s .
@xmath0 of @xmath1 , @xmath12| = 1 $ ] for every @xmath13 ( where @xmath14 $ ] denotes the closed neighborhood of @xmath15 ) .
note that not every graph has an e.d.s . ; the efficient dominating set ( ed ) problem asks for the existence of an e.d.s . in a given graph @xmath1 .
the exact cover problem ( x3c [ sp2 ] in @xcite ) asks for a subset @xmath16 of a set family @xmath17 over a ground set , say @xmath18 , containing every vertex in @xmath18 exactly once . as shown by karp @xcite ,
the problem is @xmath2-complete even for set families containing only @xmath19-element subsets of @xmath18 ( see @xcite ) .
clearly , ed is the exact cover problem for the closed neighborhood hypergraph of @xmath1 .
the notion of efficient domination was introduced by biggs @xcite under the name _
perfect code_. the ed problem is motivated by various applications , including coding theory and resource allocation in parallel computer networks ; see e.g. @xcite . in @xcite
, it was shown that the ed problem is @xmath2-complete .
moreover , ed is @xmath2-complete for @xmath3-free chordal unipolar graphs @xcite . in this paper
, we will also consider the following weighted version of the ed problem : the relationship between we d and ed is analyzed in @xcite .
for a set @xmath17 of graphs , a graph @xmath1 is called _
@xmath17-free _ if @xmath1 contains no induced subgraph isomorphic to a member of @xmath17 . in particular
, we say that @xmath1 is _ @xmath5-free _ if @xmath1 is @xmath20-free .
let @xmath21 denote the disjoint union of graphs @xmath22 and @xmath23 , and for @xmath24 , let @xmath25 denote the disjoint union of @xmath26 copies of @xmath5 . for @xmath27 , let @xmath28 denote the chordless path with @xmath29 vertices , and
let @xmath30 denote the complete graph with @xmath29 vertices ( clearly , @xmath31 ) .
for @xmath32 , let @xmath33 denote the chordless cycle with @xmath29 vertices . for indices
@xmath34 , let @xmath35 denote the graph with vertices @xmath36 , @xmath37 , @xmath38 such that the subgraph induced by @xmath36 forms a @xmath39 @xmath40 , the subgraph induced by @xmath41 forms a @xmath42 @xmath43 , and the subgraph induced by @xmath44 forms a @xmath45 @xmath46 , and there are no other edges in @xmath35 .
thus , claw is @xmath47 , chair is @xmath48 , and @xmath49 is isomorphic to e.g. @xmath50 .
@xmath5 is a _ linear forest _ if every component of @xmath5 is a chordless path , i.e. , @xmath5 is claw - free and cycle - free .
@xmath5 is a _ co - chair _ if it is the complement graph of a chair .
@xmath5 is a @xmath6 if @xmath5 has five vertices such that four of them induce a @xmath51 and the fifth is adjacent to exactly one of the @xmath51-vertices .
@xmath5 is a co-@xmath6 if @xmath5 is the complement graph of a @xmath6 .
@xmath5 is a _ bull _ if @xmath5 has five vertices such that four of them induce a @xmath52 and the fifth is adjacent to exactly the two mid - points of the @xmath52 .
@xmath5 is a _ net _ if @xmath5 has six vertices such that five of them induce a bull and the sixth is adjacent to exactly the vertex of the bull with degree 2 .
@xmath5 is a _ gem _ if @xmath5 has five vertices such that four of them induce a @xmath52 and the fifth is adjacent to all of the @xmath52 vertices .
@xmath5 is a _ co - gem _ if @xmath5 is the complement graph of a gem . for a vertex @xmath13
, @xmath53 denotes its ( _ open _ ) _ neighborhood _ , and @xmath14=\{v\ } \cup n(v)$ ] denotes its _ closed neighborhood_. a vertex @xmath10 _ sees _ the vertices in @xmath54 and _ misses _ all the others .
the _ non - neighborhood _ of a vertex @xmath10 is @xmath55 $ ] . for @xmath56 , @xmath57 and @xmath58 .
we say that for a vertex set @xmath59 , a vertex @xmath60 has a join ( resp .
, co - join ) to @xmath61 if @xmath62 ( resp . , @xmath63 ) .
join ( resp .
, co - join ) of @xmath10 to @xmath61 is denoted by @xmath64 ( resp . , @xmath65 ) .
correspondingly , for vertex sets @xmath66 with @xmath67 , @xmath68 denotes @xmath69 for all @xmath70 and @xmath71 denotes @xmath72 for all @xmath70 . a vertex @xmath73 _
contacts @xmath74 _ if @xmath15 has a neighbor in @xmath74 . for vertex sets
@xmath75 with @xmath76 , @xmath74 _ contacts @xmath77 _ if there is a vertex in @xmath74 contacting @xmath77 .
if @xmath78 but @xmath10 has neither a join nor a co - join to @xmath61 , then we say that @xmath10 _ distinguishes _ @xmath61 .
a set @xmath5 of at least two vertices of a graph @xmath1 is called _ homogeneous _
if @xmath79 and every vertex outside @xmath5 is either adjacent to all vertices in @xmath5 , or to no vertex in @xmath5 .
obviously , @xmath5 is homogeneous in @xmath1 if and only if @xmath5 is homogeneous in the complement graph @xmath80 .
a graph is _ prime _ if it contains no homogeneous set .
a graph @xmath1 is _ chordal _ if it is @xmath33-free for any @xmath32 .
@xmath9 is _ unipolar _ if @xmath18 can be partitioned into a clique and the disjoint union of cliques , i.e. , there is a partition @xmath81 such that @xmath82 $ ] is a complete subgraph and @xmath83 $ ] is @xmath84-free .
@xmath1 is a _ split graph _ if @xmath1 and its complement graph are chordal .
equivalently , @xmath1 can be partitioned into a clique and an independent set .
it is well known that @xmath1 is a split graph if and only if it is ( @xmath85)-free @xcite .
it is well known that ed is @xmath2-complete for claw - free graphs ( even for ( @xmath86)-free perfect graphs @xcite ) as well as for bipartite graphs ( and thus for triangle - free graphs ) @xcite and for chordal graphs @xcite .
thus , for the complexity of ed on @xmath5-free graphs , the most interesting cases are when @xmath5 is a linear forest . since ( w)ed is @xmath2-complete for @xmath3-free graphs and polynomial for @xmath87-free graphs @xcite , @xmath4-free graphs were the only open case finally solved in @xcite by a direct polynomial time approach ( and in @xcite by an indirect one ) .
it is well known that for a graph class with bounded clique - width , ed can be solved in polynomial time @xcite .
thus we only consider ed on @xmath5-free chordal graphs for which the clique - width is unbounded .
for example , the clique - width of @xmath5-free chordal graphs is unbounded for claw - free chordal graphs while it is bounded if @xmath88bull , gem , co - gem , co - chair@xmath89 ; there are only two open cases for the clique - width of @xmath5-free chordal graphs @xcite .
for graph @xmath9 , the square @xmath90 has the same vertex set @xmath18 , and two vertices @xmath91 , @xmath92 , are adjacent in @xmath90 if and only if @xmath93 .
the we d problem on @xmath1 can be reduced to maximum weight independent set on @xmath90 ( see @xcite ) . while the complexity of ed for @xmath3-free chordal graphs is @xmath2-complete ( as mentioned above ) , it was shown in @xcite that we d is solvable in polynomial time for @xmath4-free chordal graphs , since for @xmath4-free chordal graphs @xmath1 with e.d.s . , @xmath90 is chordal . for @xmath5-free chordal graphs ,
however , there are still many open cases . motivated by the @xmath90 approach in @xcite , and the result of milani @xcite showing that for ( @xmath7,net)-free graphs @xmath1 , its square @xmath90 is claw - free , we show in the next section that we d is solvable in polynomial time for @xmath5-free chordal graphs when @xmath94,net@xmath89 .
[ eds122frchordalg2 ] if @xmath1 is an @xmath7-free chordal graph with e.d.s . , then @xmath90 is chordal
. * proof .
* let @xmath9 be an @xmath7-free chordal graph and assume that @xmath1 contains an e.d.s .
@xmath0 . * case 1 . *
assume first that @xmath90 contains @xmath51 , say with vertices @xmath95 such that @xmath96 and @xmath97 , @xmath98 ( index arithmetic modulo 4 ) .
since @xmath1 is chordal , we have @xmath99 for each @xmath98 ; let @xmath100 denote a common neighbor of @xmath101 ( an _ auxiliary vertex _ ) . clearly , @xmath102 and @xmath103 for each @xmath98 . since @xmath1 is chordal , @xmath104 either induce a diamond or @xmath105 in @xmath1 .
assume first that @xmath104 induce a diamond in @xmath1 , say with @xmath106 .
* case 1.1 * @xmath107 . if @xmath108 then by the e.d.s . property , @xmath109 .
let @xmath110 with @xmath111 and @xmath112 with @xmath113 . clearly , since @xmath1 is chordal , @xmath114 , and by the e.d.s .
property , @xmath115 , and @xmath116 , but then @xmath117 induce an @xmath7 in @xmath1 .
thus , @xmath118 and correspondingly , @xmath119 .
if @xmath120 then by the e.d.s . property , @xmath121
let @xmath122 with @xmath123 and @xmath113 . clearly , since @xmath1 is chordal , @xmath124 .
if neither @xmath125 nor @xmath126 is adjacent to @xmath127 then let @xmath128 be a @xmath0-neighbor of @xmath127 but then @xmath129 induce an @xmath7 in @xmath1 .
thus assume that @xmath130 but then @xmath131 induce an @xmath7 in @xmath1 . * case 1.2 .
* @xmath132 and @xmath133 . without loss of generality ,
let @xmath134 ; then by the e.d.s .
property , @xmath135 .
let @xmath136 with @xmath137 . clearly , by the e.d.s .
property , since @xmath138 and since @xmath1 is chordal , we have @xmath139 , @xmath140 , and @xmath141 , but now @xmath142 induce an @xmath7 in @xmath1 .
thus , also @xmath143 .
* case 1.3 . * @xmath144 . for @xmath98 ,
let @xmath145 be the @xmath0-neighbor of @xmath146 ; @xmath147 , @xmath148 . clearly , since @xmath149 and since @xmath1 is chordal , @xmath150 are pairwise distinct .
if neither @xmath125 nor @xmath151 is adjacent to @xmath152 then for a @xmath0-neighbor @xmath128 of @xmath152 , @xmath153 induce an @xmath7 in @xmath1 .
thus , without loss of generality , assume that @xmath154 , but now @xmath155 induce an @xmath7 in @xmath1 . in a very similar way
, we can show that we can exclude a @xmath51 in @xmath90 when @xmath104 induce a @xmath105 in @xmath1 .
* now assume that @xmath90 contains @xmath156 , @xmath157 , say with vertices @xmath158 such that @xmath96 and @xmath159 , @xmath160 , @xmath161 ( index arithmetic modulo @xmath26 ) .
since @xmath1 is chordal , we have @xmath99 for each @xmath162 ( otherwise , if , e.g. , @xmath163 , then @xmath164 and @xmath165 , for the auxiliary vertices , @xmath166 , and now we get a chordless cycle in @xmath1 ) ; let @xmath100 denote a common neighbor of @xmath101 .
again , the auxiliary vertices are pairwise distinct .
since @xmath1 is chordal , we have @xmath167 for each @xmath162 .
we first claim that @xmath168 induce a clique @xmath169 in @xmath1 : without loss of generality , suppose to the contrary that there is a non - edge @xmath170 for some @xmath171 ; we can first consider a @xmath172 , i.e. , @xmath173 .
if @xmath174 and @xmath175 then @xmath176 induce an @xmath7 in @xmath1 .
thus @xmath174 implies @xmath177 , and by symmetry , @xmath177 implies @xmath106 and so on .
finally , @xmath168 induce a chordless @xmath172 in @xmath1 which is a contradiction .
the same arguments can be used for @xmath178 .
thus , @xmath168 induce a clique @xmath169 in @xmath1 .
now we consider the following cases : * case 2.1 .
* @xmath179 . without loss of generality ,
let @xmath108 .
then clearly , by the e.d.s . property , @xmath180 .
let @xmath110 with @xmath111 and @xmath181 with @xmath182 . clearly , since @xmath183 , @xmath184 , by the e.d.s .
property , @xmath185 , @xmath186 , and since @xmath187 does not induce a @xmath51 in @xmath1 , we have @xmath188 , but now , @xmath189 induce an @xmath7 in @xmath1 .
* case 2.2 .
* @xmath190 and @xmath191 . without loss of generality ,
let @xmath134 . then clearly , by the e.d.s .
property , @xmath192 .
let @xmath136 with @xmath137 and @xmath181 with @xmath182 .
clearly , as above , @xmath193 and @xmath194 , @xmath139 , and since @xmath195 does not induce a @xmath51 in @xmath1 , we have @xmath196 but now , @xmath197 induce an @xmath7 in @xmath1 .
* case 2.3 .
* @xmath198 . without loss of generality ,
let @xmath199 with @xmath123 and @xmath136 with @xmath137 .
then clearly , @xmath200 ( otherwise , @xmath125 could be considered as an auxiliary vertex as in case 2.1 ) . if neither @xmath125 nor @xmath151 see @xmath152 then there is a @xmath0-vertex @xmath201 seeing @xmath152 but then @xmath202 induce an @xmath7 in @xmath1 . thus , without loss of generality , let @xmath154 .
clearly , as above , @xmath194 , @xmath139 , @xmath196 , and @xmath203 but now , @xmath197 induce an @xmath7 in @xmath1 . thus , lemma [ eds122frchordalg2 ] is shown .
by @xcite , and since mwis is solvable in linear time for chordal graphs @xcite , we obtain : [ weds122frchordalpol ] we d is solvable in time @xmath204 for @xmath7-free chordal graphs . actually , the time bound is even better as described in a more detailed way in @xcite .
similarly , we can show : [ ednetfrchordalg2 ] if @xmath1 is a net - free chordal graph with e.d.s . , then @xmath90 is chordal
. * proof .
* let @xmath9 be a net - free chordal graph and assume that @xmath1 contains an e.d.s .
@xmath0 .
* assume first that @xmath90 contains @xmath51 , say with vertices @xmath95 such that @xmath96 and @xmath97 , @xmath98 ( index arithmetic modulo 4 ) .
since @xmath1 is chordal , we have @xmath99 for each @xmath98 ; let @xmath100 denote a common neighbor of @xmath101 ( an _ auxiliary vertex _ ) .
clearly , @xmath102 and @xmath103 for each @xmath98 . since @xmath1 is chordal , @xmath104 either induce a diamond or @xmath105 in @xmath1 .
assume first that @xmath104 induce a diamond in @xmath1 , say with @xmath106 .
* case 1.1 * @xmath107 . if @xmath108 then by the e.d.s . property , @xmath109 .
let @xmath128 with @xmath205 . clearly , by the e.d.s .
property , @xmath206 , @xmath207 , @xmath208 , and since @xmath1 is chordal , @xmath209 but then @xmath210 induce a net in @xmath1 .
thus , @xmath118 and correspondingly , @xmath119 . if @xmath211 then @xmath212 . let @xmath128 with @xmath213 . clearly , by the e.d.s .
property , @xmath214 , @xmath207 , and since @xmath1 is chordal , @xmath209 .
let @xmath215 be the @xmath0-neighbor of @xmath216 .
now , if @xmath217 then @xmath218 induce a net in @xmath1 , and if @xmath219 then @xmath220 induce a net in @xmath1 .
thus , @xmath221 and correspondingly , @xmath222 .
this implies @xmath132 .
* case 1.2 . * @xmath133 and @xmath132 .
if without loss of generality , @xmath134 then @xmath223 .
let @xmath128 with @xmath205 . clearly , since case 1.1 is excluded ( and if @xmath224 then @xmath201 would be an auxiliary vertex ) , we have @xmath209 ; let @xmath215 be the @xmath0-neighbor of @xmath216 . if @xmath225 then @xmath226 induce a net in @xmath1 .
now assume @xmath227 ; then by the e.d.s .
property , @xmath228 .
clearly , @xmath229 and @xmath230 but now , if @xmath231 , then @xmath232 induce a net in @xmath1 , and if @xmath233 , then @xmath234 induce a net in @xmath1 .
thus , @xmath143 .
* case 1.3 .
* @xmath235 .
let @xmath145 be the @xmath0-neighbor of @xmath146 ; @xmath147 .
clearly , @xmath150 are pairwise distinct . if @xmath236 and @xmath237 then @xmath238 induce a net in @xmath1 , and correspondingly by symmetry , for @xmath239 , @xmath240 .
thus , we can assume that for each @xmath29 , @xmath241 sees at least one of @xmath242 .
if @xmath154 and @xmath130 then clearly , @xmath139 and @xmath243 and thus , by the above , we can assume that @xmath244 and @xmath116 but now , @xmath245 induce a net in @xmath1 .
thus , assume that @xmath125 is adjacent to exactly one of @xmath246 , say @xmath154 ( which implies @xmath139 ) and @xmath237 . by symmetry , this holds for @xmath247 as well , i.e. , @xmath248 , @xmath249 , and @xmath250 .
then @xmath251 induce a net in @xmath1 . in a very similar way
, we can show that we can exclude a @xmath51 in @xmath90 when @xmath104 induce a @xmath105 in @xmath1 .
* now assume that @xmath90 contains @xmath156 , @xmath157 , say with vertices @xmath158 such that @xmath96 and @xmath159 , @xmath160 , @xmath161 ( index arithmetic modulo @xmath26 ) .
since @xmath1 is chordal , we have @xmath99 for each @xmath162 ( otherwise , if , e.g. , @xmath163 , then @xmath164 and @xmath165 , for the auxiliary vertices , @xmath166 , and now we get a chordless cycle in @xmath1 ) ; let @xmath100 denote a common neighbor of @xmath101 .
again , the auxiliary vertices are pairwise distinct .
since @xmath1 is chordal , we have @xmath167 for each @xmath162 .
since @xmath157 , there is an edge @xmath167 having a common neighbor @xmath252 , @xmath253 , say without loss of generality , @xmath254 induce a @xmath255 in @xmath1 . then @xmath256 induce a net in @xmath1 ( note that for @xmath157 , we do not need the existence of an e.d.s . in @xmath1 ) .
similarly , we get a net for any @xmath156 in @xmath90 , @xmath178 .
thus , lemma [ ednetfrchordalg2 ] is shown . by @xcite
, we obtain : [ wednetfrchordalpol ] we d is solvable in time @xmath204 for net - free chordal graphs .
lemma [ ednetfrchordalg2 ] generalizes the corresponding result for at - free chordal graphs ( i.e. , interval graphs ) .
as mentioned in the introduction , the reduction from x3c to efficient domination shows that ed is @xmath2-complete . for making this manuscript
self - contained , we describe the reduction here : let @xmath257 with @xmath258 and @xmath259 be a hypergraph with @xmath260 for all @xmath261 .
let @xmath262 be the following reduction graph : @xmath263 such that @xmath264 , @xmath265 and @xmath266 are pairwise disjoint .
the edge set of @xmath262 consists of all edges @xmath267 whenever @xmath268 .
moreover @xmath18 is a clique in @xmath262 , and every @xmath269 is only adjacent to @xmath100 . clearly , @xmath257 has an exact cover if and only if @xmath262 has an e.d.s .
@xmath0 : for an exact cover @xmath270 of @xmath5 , every @xmath271 corresponds to vertex @xmath272 , and every @xmath273 corresponds to vertex @xmath274 .
conversely , if @xmath0 is an e.d.s . in @xmath262 we can assume without loss of generality that @xmath275 , and now , @xmath276 corresponds to an exact cover of @xmath5 .
clearly , @xmath262 is chordal and unipolar ; actually , it corresponds to a slight generalization of split graphs ; we call it _
split - matching - extended graphs_. moreover , the reduction shows that @xmath262 is not only @xmath3-free but also @xmath5-free for various other graphs @xmath5 such as @xmath277 , @xmath278 , butterfly , extended butterfly , extended co-@xmath6 , extended chair , and double - gem as shown in figure [ butterfly ] . in a forthcoming paper , we show that split - matching - extended graphs can be characterized by the forbidden induced subgraphs mentioned above .
thus , we have : [ butterflynpc ] ed is @xmath2-complete for @xmath279 , @xmath277 , @xmath278 , butterfly , extended butterfly , extended co-@xmath6 , extended chair , double - gem@xmath280-free chordal graphs on the other hand , ed is polynomial for various smaller forbidden subgraphs such as gem - free chordal graphs ; as already mentioned , the clique - width of gem - free chordal graphs is bounded . in @xcite , for all but two stubborn cases , the clique - width of @xmath5-free chordal graphs is classified .
for example , the clique - width of co-@xmath6-free chordal graphs , of @xmath281-free chordal graphs , and of claw - free chordal graphs is unbounded .
[ copk3p2fr ] for @xmath281-free chordal graphs and for co-@xmath6-free chordal graphs , we d is solvable in polynomial time
. * proof . *
first let @xmath9 be a connected @xmath281-free chordal graph , and assume that @xmath1 contains a @xmath255 , say @xmath282 ( otherwise @xmath1 is a tree since a @xmath255-free chordal graph is a forest ) .
clearly , the non - neighborhood of @xmath282 is an independent vertex set , and according to the approach described in @xcite , it is clear that there are at most three vertices in an e.d.s .
@xmath0 dominating @xmath282 ; then the remaining non - dominated vertices in the non - neighborhood of @xmath282 belong to @xmath0 .
another direct approach works as follows : if @xmath1 has an e.d.s .
@xmath0 such that for a @xmath255 @xmath5 , one of its vertices , say @xmath201 , is in @xmath0 then for distance levels @xmath283 of @xmath201 , @xmath284 is independent and thus , @xmath285 ( recall that @xmath286 ) , and it has to be checked whether @xmath287 is an e.d.s . of @xmath1 .
now assume that no vertex of any @xmath255 in @xmath1 is in @xmath0 .
let again @xmath5 be a @xmath255 in @xmath1 , say with vertices @xmath288 .
then @xmath289 and @xmath290 is independent .
let @xmath291 be the @xmath0-neighbors of @xmath288 , respectively , and let @xmath292 .
clearly , @xmath293 .
thus , for each possible triple of @xmath0-vertices @xmath294 , one has to check whether @xmath295 is an e.d.s . of @xmath1 .
now let @xmath9 be a prime co-@xmath6-free chordal graph ( by @xcite , we can reduce the we d problem to prime graphs ) .
as above , we can assume that @xmath1 contains a @xmath255 @xmath5 , say with vertices @xmath288 .
if @xmath296 , say @xmath297 , then let @xmath298 be a vertex distinguishing an edge , say @xmath299 , in @xmath300 .
now , since @xmath301 , there is a @xmath0-neighbor @xmath302 of @xmath15 but now @xmath303 induce a co-@xmath6 in @xmath1 which is a contradiction .
thus assume that for every @xmath255 @xmath5 in @xmath1 , @xmath304 .
this implies that any @xmath0-vertex @xmath201 has at most one neighbor in any @xmath255 @xmath5 in @xmath1 .
thus assume that each vertex @xmath288 in @xmath5 has its personal @xmath0-neighbor , say @xmath291 .
if one of the @xmath0-neighbors of @xmath5 , say , @xmath305 , has another neighbor @xmath15 then @xmath306 since otherwise , the @xmath255 induced by @xmath307 would contain a @xmath0-vertex . clearly , @xmath308 and @xmath309 since @xmath310 and @xmath311 and @xmath1 is chordal .
thus , @xmath312 would induce a co-@xmath6 .
thus , assume that @xmath294 , respectively , have only one neighbor .
now let @xmath313 and @xmath314 .
we claim that @xmath315 is independent : if there is an edge @xmath316 in @xmath315 then , since @xmath1 is prime , there is a vertex @xmath317 distinguishing an edge , say @xmath318 in the connected component of @xmath315 containing @xmath316 but then @xmath1 contains a co-@xmath6 with vertices from @xmath319 ( depending on the neighborhood of @xmath320 in @xmath321 ) .
thus , @xmath315 is independent and clearly , @xmath285 , and now for the algorithmic approach , we have to check for a @xmath255 in @xmath1 whether the corresponding set @xmath322 is an e.d.s . of @xmath1 . [ copk3p2frcorollary ] for every proper induced subgraph @xmath323 of any graph @xmath324 @xmath3 , @xmath277 , @xmath278 , butterfly , extended butterfly , extended co-@xmath6 ,
extended chair , double - gem @xmath89 , we d is solvable in polynomial time for @xmath323-free chordal graphs
. * proof .
* by @xcite , the clique - width of co - chair - free chordal graphs is bounded , and by @xcite , the clique - width of gem - free chordal graphs is bounded . by lemma [ eds122frchordalg2 ] , we d is solvable in polynomial time for chair - free chordal graphs . by lemma [ copk3p2fr ] , we d is solvable in polynomial time for @xmath281-free chordal graphs , and for co-@xmath6-free chordal graphs . in all other cases
, we can use the fact that if we d is solvable in polynomial time for @xmath5-free graphs then it is solvable in polynomial time for @xmath325-free graphs ( see @xcite ) and the fact that we d is solvable in polynomial time for @xmath326-free graphs ( and thus also for @xmath327-free graphs ) .
in this section , we generalize lemma [ eds122frchordalg2 ] and the polynomial - time solvability of we d for @xmath7-free chordal graphs as well as for @xmath4-free chordal graphs by a direct approach for @xmath8-free chordal graphs .
let @xmath9 be @xmath8-free chordal .
for any vertex @xmath13 , let * @xmath328 \subset n[u]\}$ ] , and * @xmath329 \subset n[v]\}$ ] . clearly such sets @xmath330 , @xmath331 may be empty and , for any @xmath91 , @xmath332 $ ] if and only if @xmath333 $ ] . then for any @xmath91 , let us say that @xmath334 if @xmath335 or @xmath336 , i.e. , @xmath337 \subset n[y]$ ] . clearly , @xmath338 is a partial order ( in particular , anti - symmetric and transitive ) .
a vertex @xmath13 is _ maximal _ if there is no @xmath339 , @xmath340 , with @xmath341 , i.e. , @xmath342 .
clearly , @xmath343 has such a maximal element .
[ wedmaxvdirects123frchordalgr ] let @xmath13 be a maximal element of @xmath343 . then a minimum ( finite ) weight e.d.s
containing @xmath10 in the connected component of @xmath344 $ ] with @xmath10 ( if it exists ) can be computed in polynomial time .
* let us assume without loss of generality that @xmath1 is connected . as usual
, let @xmath345 ( for some natural @xmath346 ) denote the distance levels of @xmath10 in @xmath1 .
then @xmath347 is a partition of @xmath18 .
clearly , @xmath348 .
since @xmath1 is chordal , we have : [ xnicliquen(x)ni-1 ] for every @xmath349 and every vertex @xmath350 , @xmath351 is a clique , and in particular , @xmath15 contacts exactly one component of @xmath352 $ ] . [ u1inn1nonadjz1 ] for any vertex @xmath353 , there is a non - neighbor @xmath354 of @xmath355 . for any vertex @xmath356 , with neighbor @xmath353 , there is a vertex @xmath354 nonadjacent to both @xmath357 . for any vertex @xmath358 , for @xmath359 , there is an induced path of at least three vertices from @xmath360 to @xmath10 .
the first statement holds since @xmath10 is a maximal element of @xmath343 .
the second statement holds by the first one and since @xmath1 is chordal .
the third statement trivially holds by construction .
assume that @xmath0 is a ( possible ) e.d.s . of finite weight of @xmath1 containing @xmath10 .
for any fixed @xmath29 , @xmath349 , let * @xmath362 has a neighbor in @xmath363 , and let * @xmath364 ( for some natural @xmath365 ) be the family of components of @xmath366 $ ] contacting @xmath61 . [ fact1 ] for any @xmath70 , @xmath15 contacts exactly one component of @xmath366 $ ] . _ proof .
_ suppose to the contrary that @xmath15 contacts two components of @xmath366 $ ] , say @xmath367 , and let us assume without loss of generality that the neighbor of @xmath15 in @xmath368 , say @xmath201 , belongs to @xmath5 .
then let @xmath369 be any neighbor of @xmath15 in @xmath323 : then by the e.d.s .
property @xmath369 has a neighbor in @xmath0 , say @xmath370 , with @xmath371 . by claim [ xnicliquen(x)ni-1 ] , @xmath372 ,
i.e. , @xmath373
. then by claim [ u1inn1nonadjz1 ] , @xmath374 , and three further vertices of @xmath1 ( with respect to @xmath15 ) induce an @xmath8 , which is a contradiction .
@xmath361 [ p3 ] 1 .
@xmath61 admits a partition @xmath375 such that @xmath376 contacts @xmath377 and does not contact @xmath378 ( for @xmath379 ) , @xmath380 .
2 . @xmath381 , for @xmath382 , say @xmath383 , and @xmath384 dominates @xmath385 .
_ ( i ) follows directly by claim [ fact1 ] and by definition of @xmath61 and @xmath386 .
+ ( ii ) : first we claim that @xmath381 ( clearly , @xmath387 ) : suppose to the contrary that there are @xmath388 , @xmath389 .
since @xmath1 is connected and by definition of @xmath61 , there are @xmath70 with @xmath390 and @xmath391 with @xmath392 . by the e.d.s .
property , the shortest path , say @xmath6 in @xmath376 from @xmath201 to @xmath370 has at least two internal vertices , i.e. , there exist @xmath393 with @xmath394 and @xmath395 . since @xmath1 is @xmath8-free , by claim [ u1inn1nonadjz1 ] and by the e.d.s .
property , @xmath15 is nonadjacent to all vertices of @xmath396 , while @xmath369 is nonadjacent to all vertices of @xmath397 which contradicts the fact that @xmath1 is chordal .
let @xmath383 .
next we claim that @xmath384 dominates @xmath377 : this follows by definition of @xmath61 , by statement ( i ) , and by the e.d.s .
property .
finally we claim that @xmath384 dominates @xmath376 : suppose to the contrary that @xmath384 does not dominate @xmath376 , i.e. , there is a vertex @xmath398 with @xmath399 .
then there is @xmath128 , @xmath400 , with @xmath401 .
let @xmath402 be a shortest path in @xmath376 between @xmath384 and @xmath403 , and let @xmath70 be adjacent to @xmath384 ( by the above , @xmath384 dominates @xmath377 ) . clearly , by the e.d.s .
property , @xmath404 .
_ case _ 1 : @xmath405 . then by claim [ xnicliquen(x)ni-1 ] , @xmath406 ,
i.e. , @xmath407 ; then by claim [ u1inn1nonadjz1 ] , @xmath408 , and three further vertices of @xmath1 ( with respect to @xmath15 ) induce a @xmath8 , which is a contradiction
. _ case _ 2 : @xmath409 .
then the only possible sub - cases are : _ case _ 2.1 : @xmath410 .
then , by considering the ( not necessarily induced ) path formed by @xmath411,@xmath412 , one gets a contradiction to the fact that @xmath1 is chordal .
_ case _ 2.2 : @xmath407 .
then let @xmath413 be a neighbor of @xmath403 in @xmath414 ; clearly , by the e.d.s .
property , @xmath415 .
note that @xmath413 is nonadjacent to @xmath384 ( else one would be again in case 1 in which @xmath15 is adjacent to @xmath403 ) , and @xmath413 is adjacent to @xmath15 ( else by an argument similar to that of the previous paragraph one would get a contradiction since @xmath1 is chordal ) .
then there is @xmath215 adjacent to @xmath413 .
clearly , @xmath416 by the above .
furthermore @xmath371 : otherwise @xmath417 , and then by considering the path in @xmath418 between @xmath201 and @xmath384 , we get a contradiction by an argument similar to the one above for showing that @xmath381 .
then let us consider the following sub - cases : _ case _ 2.2.1 : @xmath419 .
then by claim [ u1inn1nonadjz1 ] , @xmath420 , and two further vertices of @xmath1 ( with respect to @xmath370 ) induce an @xmath8 , which is a contradiction .
_ case _ 2.2.2 : @xmath421 ( which implies @xmath422 ) .
then , since @xmath1 is chordal , @xmath413 and @xmath370 have a common neighbor in @xmath423 , say @xmath320 , and then @xmath320 is adjacent to @xmath15 ( since otherwise , by claim [ u1inn1nonadjz1 ] , @xmath424 , and two further vertices of @xmath1 ( with respect to @xmath320 ) induce an @xmath8 ) . now , since @xmath425 , the vertices @xmath426 and three further vertices of @xmath1 ( according to claim [ u1inn1nonadjz1 ] with respect to @xmath320 ) induce an @xmath8 ) , which is a contradiction .
_ case _ 2.2.3 : @xmath427
. then by claim [ u1inn1nonadjz1 ] , @xmath428 , and three further vertices of @xmath1 ( with respect to @xmath413 ) induce an @xmath8 , which is a contradiction .
thus , claim [ p3 ] is shown .
@xmath361 [ p4 ] for every component @xmath282 of @xmath429 $ ] , @xmath359 , we have 1 .
@xmath430 , and 2 .
if @xmath431 , say @xmath432 then @xmath201 dominates @xmath282 . _ proof . _
( i ) : the first statement can be proved similarly to the first paragraph of the proof of claim [ p3 ] ( ii ) .
( ii ) : the second statement follows by claim [ p3 ] ( ii ) since @xmath433 ( and thus @xmath282 ) contacts a set of vertices of @xmath423 which consequently have a neighbor in @xmath434 . @xmath361 now let us consider the problem of checking whether such an e.d.s .
@xmath0 does exist . according to claim [ xnicliquen(x)ni-1 ] , graph @xmath1
can be viewed as a tree @xmath435 rooted at @xmath436 , whose nodes are the components of @xmath429 $ ] for @xmath437 ( for @xmath438 ) , such that two nodes are adjacent if and only if the corresponding components contact each other . then for any component @xmath282 of @xmath429 $ ] , @xmath439 , let us denote as @xmath440 $ ] the subgraph of @xmath1 corresponding to the subtree of @xmath435 rooted at @xmath282 . in particular ,
@xmath441 = g$ ]
. according to claim [ p4 ] , let us a say that a vertex @xmath201 of @xmath1 of finite weight , belonging to a component say @xmath282 of @xmath429 $ ] , @xmath439 , is a _ @xmath0-candidate _ ( or equivalently let us say that @xmath282 admits a _ @xmath0-candidate _
@xmath201 ) if 1 .
@xmath201 dominates @xmath282 , and 2 .
there is an e.d.s . in
@xmath440 $ ] containing @xmath201 .
[ p5 ] @xmath0 does exist if and only if @xmath10 is a @xmath0-candidate .
_ it directly follows by the above .
@xmath361 [ p6 ]
let @xmath282 be a component of @xmath429 $ ] , @xmath442 , and let @xmath201 be a vertex of finite weight of @xmath282 .
then let us write : @xmath443 for @xmath444 . then let @xmath445 is nonadjacent to @xmath446 , @xmath447 is adjacent to @xmath446 .
then let @xmath386 be the family of components of @xmath448 $ ] contacting @xmath61 , and let @xmath449 be the family of components of @xmath448 $ ] not contacting @xmath61 .
finally let @xmath450 be the family of components of @xmath451 $ ] contacting components in @xmath449 .
then the following statements hold : * if @xmath452 and @xmath453 then @xmath201 is a @xmath0-candidate if and only if @xmath201 dominates @xmath282 . * if @xmath454 and @xmath453 then @xmath201 is a @xmath0-candidate if and only if @xmath201 dominates @xmath282 and each component of @xmath386 admits a @xmath0-candidate . * if @xmath452 and @xmath455 then @xmath201 is a @xmath0-candidate if and only if @xmath201 dominates @xmath282 and each component of @xmath450 admits a @xmath0-candidate . *
if @xmath454 and @xmath455 then @xmath201 is a @xmath0-candidate if and only if @xmath201 dominates @xmath282 and each component of @xmath386 admits a @xmath0-candidate and each component of @xmath450 admits a @xmath0-candidate . _ proof .
_ it directly follows by definition of @xmath0-candidate and by the e.d.s .
@xmath361 then by claims [ p5 ] and [ p6 ] , one can check if e.d.s .
@xmath0 with @xmath456 does exist by the following procedure which can be executed in polynomial time : + * procedure @xmath10-maximal - wed * _ input _ : a maximal element @xmath10 of ( @xmath457 ) .
+ _ output _ : an e.d.s .
@xmath0 of @xmath1 containing @xmath10 ( if it exists ) . +
* begin * * let @xmath458 ( for some natural @xmath346 ) , with @xmath459 , be as usual the distance levels of @xmath10 in @xmath1 .
* * for * @xmath460 * do * * * begin * * * for each component @xmath282 of @xmath429 $ ] , check whether @xmath282 admits a @xmath0-candidate * * end * * * if * @xmath10 is a @xmath0-candidate * then * return `` @xmath0 does exist '' ; * * else * return `` @xmath0 does not exist '' * end * let us observe that , if @xmath0 does exist , then : * the vertices of an e.d.s .
@xmath0 containing @xmath10 can be easily obtained according to claim [ p6 ] ( by iteratively choosing , for each component @xmath282 containing a @xmath0-candidate , @xmath461 @xmath0-candidate in @xmath282 ) , and * the vertices of a minimum weight e.d.s .
@xmath0 containing @xmath10 can be easily obtained according to claim [ p6 ] ( by iteratively choosing , for each component @xmath282 containing a @xmath0-candidate , a @xmath0-candidate in @xmath282 of minimum weight ) .
this completes the proof of lemma [ wedmaxvdirects123frchordalgr ] .
[ weddirects123frchordalgr ] for @xmath8-free chordal graphs , we d is solvable in polynomial time .
* let us observe that , if all vertices @xmath1 are maximal elements of @xmath343 , then by lemma [ wedmaxvdirects123frchordalgr ] , the we d problem can be solved for @xmath1 by computing a minimum finite weight e.d.s .
@xmath0 containing @xmath10 in the connected component of @xmath344 $ ] with @xmath10 ( if @xmath0 exists ) , for all @xmath13 .
then let us focus on those vertices @xmath15 which are not maximal elements , i.e. , there is a vertex @xmath403 with @xmath337 \subset n[y]$ ] ( which means @xmath336 ) .
thus , there is a maximal vertex @xmath10 such that @xmath462 .
in particular removing such maximal vertices @xmath10 leads to new maximal vertices in the reduced graph . recall that for any graph @xmath9 and any e.d.s .
@xmath0 of @xmath1 , @xmath463| = 1 $ ] for every @xmath464 .
[ maxreduction ] let @xmath13 be a maximal element of @xmath343 , with @xmath465 , and let @xmath462 . then , if @xmath1 has an e.d.s .
, say @xmath0 , then @xmath466 .
in particular , one can define a reduced weighted graph @xmath467 from @xmath1 as follows : * for each vertex @xmath462 , assign weight @xmath468 to all vertices in @xmath469 , and * remove @xmath10 , i.e. , @xmath470 ( and reduce @xmath467 to its prime connected components ) .
in particular , the problem of checking if @xmath1 has an ( minimum weight ) e.d.s .
not containing @xmath10 can be reduced to that of checking if @xmath467 has a finite ( minimum weight ) e.d.s .
_ the reduction is correct by the e.d.s .
property and by definition of @xmath331 .
moreover , by the e.d.s .
property , by definition of @xmath331 and by construction of @xmath467 , every ( possible ) e.d.s . of finite weight of @xmath467 contains exactly one vertex which is a neighbor of @xmath10 in @xmath1 since @xmath463| = 1 $ ] for a vertex @xmath462 . + according to the fact that the above holds in a hereditary way for any subgraph of @xmath1 , and to the fact that we d for any graph @xmath5 can be reduced to the same problem for the connected components of @xmath5 , let us introduce a possible algorithm to solve we d for @xmath1 in polynomial time . + * algorithm wed-@xmath8-free - chordal - graphs * _ input _ : graph @xmath471 .
+ _ output _ : a minimum ( finite ) weight e.d.s . of @xmath1
( if it exists ) .
+ * begin * * set @xmath472 ; * * while * @xmath473 * do * * * begin * * * take any maximal element of @xmath338 , say @xmath13 , and set @xmath474 ; * * compute a minimum ( finite ) weight e.d.s .
containing @xmath10 in the connected component of @xmath344 $ ] with @xmath10 ( if it exists ) @xmath475according to lemma [ wedmaxvdirects123frchordalgr ] and procedure @xmath10-maximal - wed@xmath89 ; * * * if * @xmath465 * then * @xmath475according to claim [ maxreduction]@xmath89 * * * begin * * * * * for each * vertex @xmath462 , assign weight @xmath468 to all vertices in @xmath469 ; * * * remove @xmath10 from @xmath18 , i.e. , set @xmath476 * * * end * * * end * * if * there exist some e.d.s . of finite weight of @xmath1 ( in particular , for each resulting set of e.d.s .
candidates , check whether this is an e.d.s . of @xmath1 ) * then * choose one of minimum weight , and return it * else * return `` @xmath1 has no e.d.s . ''
+ * end * the correctness and the polynomial time bound of the algorithm is a consequence of the arguments above and in particular of lemma [ wedmaxvdirects123frchordalgr ] and claim [ maxreduction ] .
this completes the proof of theorem [ weddirects123frchordalgr ] .
the results described in lemmas [ eds122frchordalg2 ] , [ ednetfrchordalg2 ] , and [ copk3p2fr ] are still far away from a dichotomy for the complexity of ed on @xmath5-free chordal graphs . for chordal graphs @xmath5 with four vertices ,
all cases are solvable in polynomial time as described in lemma [ wedopenhfrchordalgr ] .
* for every chordal graph @xmath5 with exactly four vertices , we d is solvable in polynomial time for @xmath5-free chordal graphs .
* for every chordal graph @xmath5 with exactly five vertices , the five cases described in figure [ k14dart ] are the only cases for which the complexity of we d is open for @xmath5-free chordal graphs
. * proof . * ( i ) : it is well known that for @xmath480 , the clique - width is bounded for @xmath5-free chordal graphs and thus , we d is solvable in polynomial time . by lemma [ eds122frchordalg2 ] , or more generally , by theorem [ weddirects123frchordalgr ] , we d is solvable in polynomial time for claw - free chordal graphs .
clearly , if @xmath482 $ ] is a @xmath327 then @xmath5 is a butterfly and thus , we d is @xmath2-complete . if @xmath482 $ ] is a @xmath105 , or paw , or @xmath52 , or @xmath483 , then the clique - width is bounded @xcite ; in particular , if @xmath482 $ ] is a paw or @xmath483 then @xmath5 is a subgraph of @xmath484 , and according to theorem 1 of @xcite , the clique - width is bounded .
the other cases correspond to graphs @xmath485 of figure [ k14dart ] ( by theorem 1 of @xcite , their clique - width is unbounded ) . in particular , for the same reason
, we d is solvable in polynomial time whenever @xmath5 is not connected ( since in that case , at least one connected component of @xmath5 has at most two vertices ) .
thus , from now on , we can assume that @xmath5 is connected .
if @xmath10 has has exactly one non - neighbor in diamond , namely one of degree 2 , or exactly one non - neighbor in paw , namely one of degree 1 , then @xmath5 is a subgraph of @xmath484 .
moreover , if @xmath10 has has exactly one non - neighbor in paw , namely one of degree 2 , then @xmath5 is a gem , and if @xmath10 has exactly one non - neighbor in @xmath52 , namely one of degree 1 , then @xmath5 is a co - chair . if @xmath10 has exactly one non - neighbor in @xmath492 , namely one of degree 1 , then @xmath5 is a bull . in all these cases ,
the clique - width is bounded according to theorem 1 of @xcite .
if @xmath10 has exactly two neighbors in @xmath105 or if @xmath10 is adjacent to degree 2 and degree 3 vertices in diamond or if @xmath10 is adjacent to the two degree 2 vertices in paw or if @xmath10 is adjacent to the two degree 2 vertices ( midpoints ) in @xmath52 , then by theorem 1 of @xcite , the clique - width is bounded . if @xmath10 is adjacent to the two vertices of degree 3 of diamond then @xmath495 . if @xmath10 is adjacent to degree 2 and degree 3 vertices in paw then @xmath491 . if @xmath10 is adjacent to degree 1 and degree 3 vertices in claw then @xmath496 . in all other cases ,
@xmath5 is @xmath326 , chair or co - p , and thus , we d is solvable in polynomial time ( by lemma [ copk3p2fr ] for co - p - free chordal graphs , and by theorem [ weddirects123frchordalgr ] , for @xmath326-free chordal graphs , and for chair - free chordal graphs ) . if @xmath10 is adjacent to a degree 3 vertex of diamond then @xmath491 . if @xmath10 is adjacent to a degree 3 vertex of paw then @xmath496 . if @xmath10 is adjacent to a degree 3 vertex of claw then @xmath500 .
in general , one can restrict @xmath5 by various conditions such as diameter ( if the diameter of @xmath5 is at least 6 then @xmath5 contains @xmath3 ) and size of connected components ( if @xmath5 has at least two connected components of size at least 3 then @xmath5 contains @xmath3 , @xmath277 , or @xmath278 ) .
it would be nice to classify the open cases in a more detailed way .
a. brandstdt , e. eschen , and e. friese , efficient domination for some subclasses of @xmath4-free graphs in polynomial time , corr arxiv:1503.00091 , 2015 ; extended abstract in : proceedings of wg 2015 , e.w .
mayr , ed .
, lncs 9224 , pp .
78 - 89 , 2015 .
a. brandstdt , p. fiur , a. leitert , and m. milani , polynomial - time algorithms for weighted efficient domination problems in at - free graphs and dually chordal graphs , _
information processing letters _ 115 ( 2015 ) 256 - 262 .
a. brandstdt , a. leitert , and d. rautenbach , efficient dominating and edge dominating sets for graphs and hypergraphs , extended abstract in : conference proceedings of isaac 2012 , lncs 7676 , 2012 , 267 - 277 .
a. brandstdt , m. milani , and r. nevries , new polynomial cases of the weighted efficient domination problem , extended abstract in : conference proceedings of mfcs 2013 , lncs 8087 , 195 - 206 .
full version : corr arxiv:1304.6255 , 2013 .
a. brandstdt and r. mosca , weighted efficient domination for @xmath326-free and @xmath4-free graphs , extended abstract in : proceedings of wg 2016 , p. heggernes , ed .
, lncs 9941 , pp .
38 - 49 , 2016 .
full version : _ siam j. discrete math . _ 30 , 4 ( 2016 ) 2288 - 2303 .
liang , c.l .
lu , and c.y .
tang , efficient domination on permutation graphs and trapezoid graphs , in : _
proceedings cocoon97 _ , t. jiang and d.t .
lee , eds .
, _ lecture notes in computer science _ vol . 1276 ( 1997 ) 232 - 241 . | a vertex set @xmath0 in a finite undirected graph @xmath1 is an _ efficient dominating set _ ( _ e.d.s . _ for short ) of @xmath1 if every vertex of @xmath1 is dominated by exactly one vertex of @xmath0 .
the _ efficient domination _
( ed ) problem , which asks for the existence of an e.d.s . in @xmath1 , is known to be @xmath2-complete even for very restricted graph classes such as for @xmath3-free chordal graphs while it is solvable in polynomial time for @xmath4-free chordal graphs ( and even for @xmath4-free graphs ) .
a standard reduction from the @xmath2-complete exact cover problem shows that ed is @xmath2-complete for a very special subclass of chordal graphs generalizing split graphs ; we call this graph class _ split - matching - extended graphs _ ( which can be characterized by forbidden induced subgraphs ) .
the reduction implies that ed is @xmath2-complete e.g. for double - gem - free chordal graphs while it is polynomial for gem - free chordal graphs ( by various reasons such as bounded clique - width , distance - hereditary graphs , chordal square etc . )
moreover , we investigate the complexity of ( weighted ) ed for some other cases of @xmath5-free chordal graphs for examples of @xmath5 such as co-@xmath6 , net , @xmath7 , and @xmath8 .
* keywords * : weighted efficient domination ; @xmath5-free chordal graphs ; split - matching - extended graphs ; @xmath2-completeness ; @xmath7-free chordal graphs ; @xmath8-free chordal graphs ; net - free chordal graphs ; polynomial time algorithm ; clique - width . |
a cloud monitoring system , which watches the sky to detect clouds above an observatory , is a powerful apparatus for ground - based telescopes if we want to check the sky easily and to execute some remote or automated observations .
the telescopes and their instruments are sure to be safe if we can monitor clouds on time and can close the dome slit before cloud coverage becomes heavy and rain drops come .
other types of weather sensors , such as rain or humidity sensors , are sometimes late to alert us to close the dome . with a cloud monitor , we also can always be sure whether the data acquired by the telescope has been affected by clouds or not .
one of the smartest methods of seeing clouds from the ground is to use some thermal infrared wavebands in which clouds themselves emit thermal radiation or reflect radiation originating from ground or sea .
a ccd camera with a fish - eye lens is cheap , but the appearance of clouds in optical is deceptive because their brightness depends strongly on the intensity of the moon and city lights that illuminate them . using a similar system ,
shamir & nemiroff ( 2005 ) developed an algorithm to make a whole - sky opacity map by means of measuring the extinctions for many stars .
however , it does not give us a direct view of the cloud distribution in the sky .
an uncooled thermal imager with panoramic optics suits a robotic telescope because it requires little maintenance .
the sloan digital sky survey ( sdss ) project has developed a scanning system using a single channel photometer cooled by liquid nitrogen ( hull et al .
1994 ; hogg et al .
it has enough sensitivity and field of view , but their scanning mechanism is not easy to construct and its cooling system needs frequent hands .
recently , thermal infrared imagers without cooling parts sensitive enough to detect thin clouds have become available , and by combining these with some panoramic optics , now we can easily see the whole sky in thermal infrared .
this idea was presented by mallama & deganan ( 2002 ) , and a similar one was developed by the apache point observatory .
however , it requires stability and reliability for these systems to be put into practical use for a robotic telescope where no operator or engineer is onsite .
we developed an infrared cloud monitor system that is successful for this use .
owing to this , we have achieved unmanned automated observation at the magnum observatory at haleakala since 2001 ( kobayashi et al . 2003 ; 2004 ) .
the magnum ( ulticolor ctive alactic clei onitoring ) project newly built a 2-m optical - infrared telescope at the university of hawaii s haleakala observatory site on the hawaiian island of maui , and has been monitoring many active galaxies and quasars in optical and near - infrared wavebands for more than several years ( yoshii et al .
we study their structure and physical environment , and finally , determine the cosmological parameters by an entirely original method ( kobayashi et al .
1998 ; yoshii 2002 ) .
the other distinct challenge of this project is to achieve unmanned automated observation .
we aimed at and have achieved months - long automated observation without anyone being required at the observatory . in order to fulfill our purpose
, our whole - sky infrared cloud monitor has many salient characteristics for automated operation .
it can see almost whole - sky at once in a thermal infrared waveband with sensitivity high enough to detect thin clouds .
it is weather - proof , and has been outdoors under hard weather conditions on the top of a high - altitude mountain , about 3,000 m high , for many years .
furthermore , the raw whole - sky images are immediately reduced to apparent emissivity maps of the cloud which will be classified into several ranks of observational condition .
these maps and ranks can be referred to by our automated observing system and the remote watchers in japan .
they are also referred to by many facilities other than the magnum telescope at haleakala .
following our success , systems using copies of our design have been used at several observatories and sites such as mauna kea in hawaii ( takato et al .
2002 ) , atacama in chile , and okayama and kiso in japan .
these attract site studies for new observatories where it is difficult for people to remain for any length of time . in this paper , we describe our infrared cloud monitor system , breaking down into its design , its hardware contents , its data analysis software , and its performance and statistical data . the main instrument design and its components including the thermal imager , reflecting optics , and data acquisition system are described in @xmath2 [ system ] . in @xmath2 [ software ] , the analysis software that detects clouds and evaluates the whole - sky condition is described .
the performance in some respects in operating the magnum observatory is presented and discussed in @xmath2 [ performance ] .
finally , the weather trends seen in our accumulated whole - sky condition data over four years are discussed in @xmath2 [ weather_trends ] . +
figure [ sysicm ] shows a schematic diagram of the magnum infrared cloud monitor .
the outdoor hardware of the system is to the left of the figure .
there are two aspherical convex mirrors with cassegrain - like alignment .
a blackbody reference plate for calibration is installed where the camera can see it near the edge of its field of view .
all electrical devices , thermal imager , signal converter for the output signal of the imager , shutter , and thermometer circuit for the blackbody reference plate , are attached under the primary mirror and surrounded by an aluminum - pipe housing whose ceiling is the primary mirror . on the central hole of the primary mirror
is installed a diamond window which is transparent to thermal infrared and prevents water from dripping into the housing .
photographs of the outdoor hardware are shown in figures [ magicm ] and [ weath1 ] .
it is about 80 cm in height and about 35 cm in diameter .
the data observed by the outdoor hardware are acquired by a linux pc , illustrated on the lower right of figure [ sysicm ] .
the pc controls the shutter with a digital i / o board , and also triggers data acquisition of images and temperatures .
the controlling and data - acquiring software , as well as analysis software , are always working and getting sets of the data including whole - sky raw images every one or two minutes throughout the night .
the output data are directed to various types of software through a lan in the magnum observatory that includes the main manager of the observatory , the real - time scheduler or selector of target astronomical objects , and the information - collecting server for image headers of astronomical observations .
the specifications of our infrared cloud monitor are listed in table [ speicm ] .
owing to the uncooled thermal imager and reflecting optics that widen the camera field of view , we can obtain almost whole - sky images in thermal infrared .
this hardware alignment , a cassegrain - like mirror system above an aluminum - pipe housing containing the thermal imager and electronic parts , has a significant advantage in that the outdoor system becomes compact and is waterproof .
the only movable component in the system is a shutter , the pronter magnetic e/100 " , which works both as a shield against the sunlight during daytime and as a flat - fielding plate for image reduction .
the main hardware components of our infrared cloud monitor are the thermal imager , the reflecting optics , and the data acquisition system , the details of which are described individually in the next subsections .
+ the uncooled infrared imager that is sensitive in the @xmath3 waveband is one of the key components of our cloud monitor . in order to see high - altitude clouds with good visibility , or to detect the thermal emission of clouds standing out against the dark background of cold space , we need to select a waveband
in which the atmosphere is highly transparent .
there are two wavebands satisfying this requirement : one is the @xmath4 @xmath5 m band , and the other is the @xmath6 @xmath5 m band .
so far as we use only one waveband and can not measure cloud temperature , the @xmath6 @xmath5 m band is better to estimate cloud emissivity .
this is because this wavelength is around the flat - top part of the blackbody radiation of the clouds and ground , and flux from the clouds in this band is less dependent on temperatures than in the @xmath4 @xmath5 m band .
we use an amber sentinel camera commercially produced by the amber a. raytheon co. in 1997 .
this imager has an uncooled bolometer array of 320@xmath7240 pixels for its detector , and is sensitive enough to detect thin cirrus clouds in thermal infrared .
it outputs both an analog signal of ntsc standard and digital signals of 12-bit depth with a frame rate of 30 hz .
the specifications of the imager are tabulated in table [ sentinel ] .
the camera has an automatically offset flat - fielding function .
it compensates for pix - to - pix scatter of bias and dark current signals that depend heavily on the temperature of the imager .
this calibration is necessary for a quick look at raw images and is recommended every several minutes ; calibration can be triggered by pc through an rs232c interface .
however , we do this calibration less frequently , and take integrated images of a closed shutter plate for more precise compensation ( see [ reduction ] ) .
several cautions are in order here .
some imagers could not take clear images in a cold environment , since their detector outputs are reduced outside the operation range .
generally , outputs of uncooled bolometers significantly change due to a large change of thermal background , because they are surrounded by internal parts of ambient temperature .
imagers being optimized for use at room temperature have relatively narrow operation ranges , and sometimes have the problem under low temperature environment .
our imager displayed this defect in winter in japan or at the haleakala site , so we asked the company to tune the electric circuit of the imager in colder environments .
moreover , we resistively heated the imager up to about room temperature at night to increase the thermal background signals from the imager itself . incidentally , the imager cost as much as $ 40,000 when we developed our cloud monitor .
recently , products with various specifications have been made available by many manufacturers at lower costs .
+ the cassegrain - like reflecting optics is another key component of our infrared cloud monitor .
the camera s field of view is not large enough to cover the whole sky , and must be expanded by other optics .
germanium crystals are generally used for the lenses of these thermal imagers because it is well transparent in thermal infrared .
fish - eye lenses made of this material , however , are not commercially available , and are also hard to develop or fabricate without a great deal of expenditure .
thus , it is reasonable to use some kind of convex mirror system .
for the fundamental shape of the mirrors , we based ours on the particular aspherical ones introduced by chahl & srinivasan ( 1997 ) .
when we look into these types of mirrors along with their optical axis , the appearance of the reflected field is not radially deformed ; they preserve a linear relationship between the apparent angle from the image center and the real radial angle from the field center .
a simple spherical mirror produces a radially compressed image toward the image edge ; the more distant from the image center we see through the mirror , the more radially compressed the objects appear .
now , we align the camera with the surface of a mirror using polar coordinates as shown in figure [ opt_design1 ] , where the lens node of the imager is located at the origin , and the mirror surface should be adjusted by the revolution of the function @xmath8 around @xmath9 axis . here
, @xmath10 is an angle of line - of - sight with the @xmath9 axis in the camera field , and @xmath11 is an angle of line - of - sight with the @xmath9 axis in the negative direction in the real field that is seen through both the camera and the mirror . according to chahl & srinivasan
, @xmath11 would proportionally correspond to @xmath10 if @xmath8 is given by @xmath12^{-1/\kappa } } \;\;\;,\ ] ] where @xmath13 is the distance between the mirror and @xmath14 along the @xmath9 axis , @xmath15 $ ] is the initial angle of the mirror , in other words , a half of the vertex angle , and @xmath16 relates to the proportionality constant @xmath17 between @xmath10 and @xmath11 , which means field widening power , as @xmath18 if we place the mirror with the convex side upwards and direct the camera to look down on the mirror vertically , we can see the sky at zenith angle between @xmath19 and @xmath20 in degrees : @xmath21 \right)\;\;\;. \end{aligned}\ ] ] a circular image of the whole - sky is obtained by setting @xmath22 as half the shorter angle of the rectangular field - of - view of the imager .
chahl & srinivasan ( 1997 ) also suggested a two - mirror system of cassegrain - like alignment , in which the aspherical shape introduced above is used as a primary mirror and a cone - like shape is used as secondary , as shown in figure [ opt_design2 ] .
here , the secondary with half a vertex angle of @xmath23 is placed with its surface facing the imager , which is equivalent to the configuration in figure [ opt_design1 ] with @xmath24 .
the section of its surface is triangular , equivalent to @xmath25 , which means the surface has no field widening power .
the section of primary mirror surface should be drawn similarly to figure [ opt_design1 ] , but in an @xmath26 coordinate system , in which @xmath27 is symmetrical with @xmath14 about the section of the secondary surface .
if @xmath23 , @xmath28 , and @xmath29 are properly optimized , the light coming from the zenith ( @xmath30 ) can reach the imager , avoiding the secondary mirror by way of @xmath28 and the secondary vertex .
then , the entire sky including the zenith can be seen .
however , though the idea is very attractive , we found a serious astigmatism aberration in these optics which is mostly derived from the cone - shape of the secondary .
when seen from above along with the optical axis , there is no curvature along with sagittal directions on the completely cone - shaped surface , while its curvature exists tangentially .
this aberration becomes extremely large if we use an imager with a large lens aperture .
a combination of our 71 mm - aperture lens imager and 240 mm - diameter primary mirror , being restricted by our manufacturing capacity , results in a point - spread - function ( psf ) size over one millimeter on the detector , which corresponds to 10 degrees in the sky .
we therefore did not completely follow the original two - mirror system scheme , and improved it to upgrade image quality at the cost of view near zenith and some amount of sensitivity . in detail
, we significantly flattened the vertex angle of the cone - shaped secondary mirror to reduce its tangential curvature .
next , we introduced an aspherical shape similar to that of the secondary mirror in figure [ opt_design1 ] , so that it also has a field - widening power to some extent , similar to a primary mirror . in fact
, @xmath27 projected by the section of this type of secondary surface does not strictly converge at one point , but its effect on the field deformation was found to be ignored .
moreover , we adopted a smaller primary mirror hole to reduce the shadow area in the image center , and also to extend the focal depth . because the primary hole squeezes the imager s lens aperture , the decline in sensitivity is compensated by frame integration . as a result of these improvements
, we adopted the parameters listed for the improved " model in table [ sys_mirror ] .
the parameters determined following the original idea of figure [ opt_design2 ] are also listed as a reference for the original " model .
our improved " model gives a shadow circle with a radius of 11 degrees at the zenith .
however , the area is not critical for us because the magnum telescope rarely observes objects around the zenith because of the sparse distribution of celestial coordinates of our targets , and because of some operational restrictions of our telescope .
most remarkable is that the image resolution was dramatically increased in our improved " model .
figure [ spt ] shows the spot - diagrams of ray - traced images for both original " and improved " models .
the specifications of the two models are also tabulated in table [ spe_mirror ] .
the improved " model decreases the size of the point - spread - functions ( psfs ) by an order of magnitude or more .
there remain some astigmatism and field curvature in our improved " model , but the psf size is within a few pixels over almost the entire field of view .
+ the surfaces of the mirrors were shaped by diamond turning on brass that is easily worked and goes well with gold coating .
using an ultra - precise , computerized , numerically controlled ( cnc ) turning machine , we obtained a surface roughness of about 20 nm in rms , which is fine enough as a mirror at a wavelength of @xmath31 m .
the surface was plated with solid gold , containing 5 % cobalt , with a thickness of 2 @xmath32 m .
finally , a physical vapor sapphire was deposited on the surface with a thickness of 0.2 @xmath32 m for protection .
the surface reflectivity is about 95 % and no serious degradation has been seen in an outdoor operation of five years .
for the first attempt to process the mirror surface , we tried an aluminum - based alloy plated by electroless nickel , and coated gold on the surface .
but we found that the gold coat degraded in a few months exposure to air .
the pinholes in the nickel plating might make water erode the aluminum base rapidly . on the central hole of the primary mirror
was placed a chemical vapor - deposited ( cvd ) diamond plate with a thickness of 0.2 mm and transparency of about 80 % at @xmath33 with no coat .
a germanium plate processed with anti - reflection coating on both top and bottom and protective coating on the top could also work in some environments , and has been used in similar systems at several sites such as mauna kea in hawaii and atacama in chile .
however , during our test operation in tokyo , japan , the upper surface of the germanium window degraded in a few months .
we guess that rain in tokyo is very acidic , which might enhance degradation of the protective coating . + the thermal imager , the amber sentinel camera , outputs images with analog signals of ntsc standard as well as digital signals of 12-bit parallel channels at the rate of 30-frames per second .
we use digital output because an analog signal of 8-bit depth data loses the lower 4 bits of original signal that is much larger than the noise signal of one or two analog - to - digital units ( adus ) , and is difficult to restore by frame integration afterwards . for each data bit , together with synchronizing clocks for frame acquisition ,
the digital signals are single - ended transistor - transistor - logic ( ttl ) standard .
we convert these signals to differential signals of rs-422 standard level by a hand - made digital electrical circuit with ic in order to transmit the signals to a pc about 15 m away from the outdoor system .
this is because the ttl signals are too delicate in a noisy environment to send more than several tens of centimeters at a high data rate .
the signals are acquired by a linux pc using a digital frame - grabber board , pc - dig produced by coreco inc .
it can grab digital data of 12 parallel channels at the rate of 4.6 mega - bytes per second for our imager , and its driver for linux os is supported by the company .
we integrate the images for 5 seconds , corresponding to 150 frames , to increase the signal - to - noise ratio .
integration for more than 10 seconds is not favorable because the whole - sky images of clouds are often blurred by migrations of clouds . for each acquisition of the whole - sky image
, we take shutter images immediately before and after it . in addition
, the temperature of the blackbody reference plate is measured at the same time .
the sets of data are acquired about every one or two minutes while the elevation of the sun is below 15 degrees .
we cease the operation almost entirely during the daytime , in case direct rays of sunlight degrade the imager detector .
each whole - sky image is processed immediately after acquisition with the shutter images , the temperature of the blackbody reference , and some calibration data measured in advance .
figure [ kumored ] illustrates how we reduce a raw whole - sky image into the apparent emissivity map of the clouds .
first , offset flat fielding for the pix - to - pix pattern is done by subtracting the average of the two shutter images obtained before and after the incident exposure of the whole - sky image .
the automatic offset flat fielding by the camera is convenient for snapshot images , but not complete for our frame - integrated whole - sky images .
we can do similar calibration with a higher signal - to - noise ratio using the frame - integrated shutter images .
next , background signals from the optics and the interior of the imager are subtracted .
there are two types of components of background signal : flat - offset components and spatial - pattern components .
they differ from each other with varying internal and environmental temperatures , and have to be subtracted separately .
the flat - offset background component includes bias and dark current signals of the detector along with thermal radiation from the interior of the imager and the optics , which should be compensated for each whole - sky image .
subtraction of the shutter image from the whole - sky image does not work well because the surface brightness of the shutter and the thermal background from the optics vary independently .
the flat - offset value of the background signal @xmath34 on the whole - sky image is calculated from the temperature @xmath35 of the blackbody reference plate and its signal @xmath36 on the incident whole - sky image as @xmath37 where @xmath38 is the planck function for temperature @xmath35 , and @xmath39 is the signal - to - surface brightness ratio measured beforehand .
the units of @xmath34 and @xmath36 are adu .
we subtract the single value of @xmath34 from those of all pixels in the incident whole - sky image .
there remains a spatial - pattern background component , which mainly originates in the baffle of the reflecting optics and the atmosphere .
the pattern of this type of background is radially symmetrical and almost stable on clear nights .
we therefore prepare a template whole - sky image for a clear night beforehand .
the template image should be acquired when the sky is certain to be clear and reduced up to compensating the flat - offset background .
we subtract the template image from all whole - sky images .
now that the background signals are subtracted , we calibrate the signals in the whole - sky image into a surface brightness value using the signal - to - surface brightness ratio @xmath39 .
one can determine the value of @xmath39 in laboratory by exposing the blackbody targets of different temperatures , or can determine it at the observatory site by exposing both clear sky and a black object of ambient temperature at once .
note that the value of @xmath39 is dependent on zenith angle , mainly because of vignetting on the aperture of the camera lens .
we should therefore measure @xmath39 for several zenith angles and using a function fitted to them .
figure [ gain ] shows the data of @xmath39 measured for our system at the magnum observatory site .
finally , the surface brightness @xmath40 in the image is converted to the apparent emissivity @xmath41 of the clouds at a 10 @xmath32 m waveband , which is related to @xmath40 as @xmath42 where @xmath43 is the plank function for the cloud temperature @xmath44 . according to the average annual air temperature of 296 k at sea level in maui island and lapse rate of @xmath45 @xmath46 for standard atmosphere ( coesa , u.s . standard atmosphere 1976 ) , the expected ambient temperature at 10,000-m altitude above haleakala observatory should be about 240 k. we therefore calculate @xmath41 assuming ( hereafter fixing ) the temperature of @xmath47 k as representative of high - altitude clouds or cirrus .
note that @xmath41 includes reflection efficiency of a cloud as well as absorption efficiency , and in @xmath40 , there is a significant amount of reflected emission by the cloud which originates in the surface of the ground or sea .
this means we can not convert the apparent emissivity @xmath41 simply to optical depth , which relates to actual absorption efficiency .
however , particularly for high - altitude clouds , it is reported that a large amount of emission still originates thermally in clouds themselves ( platt & stephens 1980 ) .
figure [ cloudimage ] shows the whole - sky cloud emissivity maps obtained and processed under various sky conditions .
it is very convenient for a remote watcher from japan as well as at the haleakala site to see the whole - sky cloud emissivity maps on the internet .
however , our main objective in operating the magnum observatory is automated observation in real - time consideration of weather conditions .
we therefore developed software to detect clouds from the whole - sky cloud emissivity maps and evaluate observational conditions from them ( @xmath2 [ acd ] and @xmath2 [ whole - sky ] ) .
+ to determine whether clouds exist or not in a certain part of the sky , it is important to measure both the average and the standard deviation value of the emissivities in a small area in about that direction , rather than to refer to just one pixel value . here , two elements limit sensitivity : one is variation in zero emissivity level caused by residual thermal background signal , and the other is pix - to - pix noise .
the empirical value of the former for our system is about @xmath48 , which is considerably large compared to those of thin clouds .
this mainly comes from the residual pattern of the background radiation that is difficult to subtract completely from a single template whole - sky image . also , humidity has some correlation with residual background .
the latter limit for the sensitivity is about @xmath49 as a noise equivalent signal of the image , which is much less than the former .
thin clouds are easy to detect by their spatial fluctuations of emissivity rather than by emissivity values themselves .
we therefore divide a whole - sky cloud emissivity map into 90 sub - areas , each being 10 degrees in elevation and 20 degrees in azimuth . for each sub - area
, we categorize the cloud condition into several levels using the average emissivity @xmath50 and the rms emissivity @xmath51 calculated for the area .
figure [ pointsky ] shows the @xmath52 versus @xmath41 diagram on which each sub - area can be evaluated . a sub - area
is evaluated as clear " only when both @xmath53 and @xmath54 , otherwise regarded as being covered by some clouds . except for
clear " , the sub - area is evaluated into thin " , thick " , or rain " when @xmath55 , @xmath56 , and @xmath57 , respectively .
the condition of rain " means that the surface brightness is larger than that for blackbody of 300 k , and the mirror system is possibly wet due to rainfall or moisture , though the direct detection of rainfall should be done by rain sensors .
each sub - area is given a level of 0 for clear " , 1 for thin " , 2 for thick " , or 3 for rain " in order to calculate whole - sky cloud condition from statistics over all sub - areas ( see next subsection ) .
the main cause preventing detection of even thinner clouds is the residual pattern of the thermal background on the whole - sky emissivity maps , which increases the rms value of emissivities even in a small sub - area .
this could be improved if a relation between the radial pattern and the temperature of the reflecting optics is contributed or the temperature of the reflecting optics is regulated .
more fundamentally , we are soon going to improve the design of the reflecting optics so that there would be no vignetting objects in the optical pass .
+ to determine whether the sky allows observation or not and what type of observation is best to execute , we evaluate the whole - sky cloud condition using statistics over sub - area values calculated and labeled in @xmath2 [ acd ] .
we classify the whole - sky cloud condition into five types : clear " , thinorpartial " , medium " , cloudy " , and rainy " .
first , for safety , we strictly exclude conditions when there are many sub - areas of rain " .
figure [ allsky_rain ] shows classifications of whole - sky cloud condition on the average over sub - area values vs. rain " sub - area coverage plane .
in so far as the rain " sub - area coverage is larger than 10 % , the whole - sky condition is evaluated as medium " , cloudy " , or
rainy " , and we do not start any type of observation .
next , when the rain " sub - area coverage is less than 10 % , we classify whole - sky cloud conditions into five types , as in figure [ allsky_cloud ] . here , cloud coverage includes both thin " and thick " sub - areas .
the whole - sky cloud condition is evaluated as clear " only at the origin of figure [ allsky_cloud ] , which means all sub - areas are clear " . except for
clear " , whole - sky cloud condition is classified as thinorpartial " , medium " , or cloudy " , according to the cloud coverage and mean sub - area level .
the classifications above have mainly been working successfully , though they are empirical and somewhat inelegant .
our cloud monitor was located at the haleakala site and began to give whole - sky cloud emissivity maps when magnum observatory started its telescope operation in august 2000 . then , an automated monitoring observation of active galaxies with the cloud monitor was put into practical use in early 2001 . after refinements of several months , we achieved fully automated astronomical observation for an entire night .
now , we have continuous unmanned observation , except for maintenance every several months ( kobayashi et al . 2003 ; 2004 ) according to the whole - sky cloud condition evaluated by the cloud monitor , our automated observing system decides whether observation is possible or not .
when the whole - sky cloud condition is either clear " or thinorpartial " , the observing system opens the dome slit and commands observation . when the whole - sky cloud condition is cloudy " or rainy " , the observing system closes the dome slit , and carries out no observation . when the whole - sky cloud condition is medium " , the observing system maintains the ongoing operation .
moreover , according to the whole - sky cloud condition , we also determine what type of observation should be executed . if the condition is clear " , which can be regarded as a photometric sky , all types of observations are possible .
if the condition is thinorpartial " , certain observations that are delicate under cloud extinction are restricted ; observations such as standard star calibration , relative photometry between several separate fields , or imaging of faint objects are allowed only in clear " conditions . instead , differential photometry between the bright objects in the same field of view is permitted in thinorpartial " conditions because it is barely affected by extinction fluctuations .
the whole - sky cloud condition and status of the sub - area at which the telescope is pointing are recorded in the fits header of observed images .
the whole - sky emissivity maps are also archived so that we can check the quality of the observed astronomical data when we analyze them .
the cloud monitor has mainly been working stably until now , except for several months of trouble with the frame grabber board .
regular maintenance includes wiping the dust on the mirror and shipping whole - sky images to japan when we visit the site every several months . in the following two sections ( @xmath2 [ defence_rain ] and @xmath2 [ det_phot ] ) , we present and discuss performance of our cloud monitor , comparing it with some other weather sensors and photometric data .
+ the most primitive function required of our cloud monitor is to determine whether the sky allows observation or not . when the sky becomes cloudy
, the cloud monitor should close the dome slit before rain falls .
it also should stop meaningless and risky opening of the dome when the sky is still covered by thick clouds .
table [ sky_rain ] shows the frequency distribution of various weather conditions over four years from two different weather - sensing systems including the whole - sky cloud monitor and the rain sensor .
note that the rain sensor directly senses rain drops by means of changes in the resistivity of the electrical circuit , while the cloud monitor only inspects surface brightness of the sky image .
the percentages of rain - sensor output in the night are 85.8 % for dry " and 14.2 % for rain " .
this table indicates that a combination of cloudy " and rainy " comprises 96 % of rain " , i.e. , rain drops can be avoided by this high probability from such a combined cloud condition .
the remaining 4 % probability corresponds to a situation in which the rain sensor catches rain drops while the output from the cloud monitor is clear " , thinorpartial " or medium " , and would decrease further if acquisition of whole - sky data were carried out more frequently , because the approach of moisture is sometimes very rapid .
the humidity sensor , however , usually helps to catch the moisture on its way up to the observatory .
+ the next important function required of the cloud monitor is to determine whether the night is photometric or not .
the flux calibration of active galaxies using reference stars or standard stars in different telescope directions often fails if we are uncertain whether the sky is entirely clear .
reliability of the whole - sky condition clear " can be estimated from statistics of accumulated standard star flux data , because these have been observed quickly while the whole - sky is clear " .
table [ sig_zerop ] presents several statistical values of our standard star observations over two years while the instrumental throughput was relatively stable .
columns ( 1 ) and ( 2 ) are the waveband and effective wavelength , respectively . column
( 3 ) is the number of observations .
the standard deviation of the fluxes over all observations @xmath58 for each band is given in column ( 4 ) , and the average over individual photometric errors @xmath59err@xmath60 is given in column ( 5 ) . nominal extinction value for unit airmass @xmath61 , measured by intensive observations of standard stars on a few nights ,
are shown in column ( 6 ) .
the linear trend of flux decreasing with time during the period , derived from changes in telescope throughput , is corrected .
airmass correction for elevation in each observation is done with a constant value in the table .
the photometric errors @xmath59err@xmath60 are so small that they contribute little to @xmath58 .
therefore , the scatter @xmath58 mainly contains the day - to - day changes in extinction by the atmosphere or clouds . converting @xmath58 and @xmath61 to flux ratio , we show @xmath62 against wavelength @xmath63 in figure [ std ] .
the vertical bars with inverted triangles on top are @xmath58 , being corrected for @xmath59err@xmath60 .
filled squares are @xmath61 .
the solid line is linear fit to the filled squares except for the @xmath64-band which is particularly affected by water vapor .
the line shows wavelength - dependence of @xmath65 , which is consistent with a typical trend of a mixture of rayleigh - scattering by molecules and mie - scattering by small aerosols in the atmosphere ( cox 2000 ) .
a similar linear wavelength - dependence is seen in @xmath58 in the optical , and should be dominated by daily or seasonal variation in the extinction by the atmosphere . on the other hand , @xmath58 at longer wavelengths beyond
the @xmath66-band is near constant , regardless of wavelength - dependence of @xmath61 .
we consider that this flat component of @xmath58 could include variation in extinction by clouds missed by our cloud monitor , because the size of typical cloud particles is on the order of ten microns and their wavelength - dependence of extinction is white at a few microns or shorter .
therefore , photometric errors caused by extinction of the clouds are restricted to within a few percent .
bradley et al . ( 2006 ) overviewed meteorological characteristics at haleakala with respect to many types of weather data , such as humidity , temperature , wind speed , and cloud coverage . however , their analysis is based on compilation of various records with fairly large spatial and time resolution , including those taken by satellites .
our whole - sky cloud conditions are more straightforward and systematic , because our conditions are completely based on direct measurements of clouds that are projected onto the sky above the observatory .
table [ log_allsky ] shows the proportions of whole - sky cloud conditions averaged monthly between january 2001 and december 2005 .
several conditions in early 2001 are combined because of test operation of analysis software .
a total of 668,063 whole - sky images for 1,601 nights give the statistics in the table .
figure [ weather_freq ] presents the relative frequencies of the whole - sky cloud conditions combined over the data in table [ log_allsky ] .
the percentage of each condition is an average over the data weighted by the number of nights in which the data were obtained .
the monthly percentage for combined conditions between january 2001 and july 2001 is divided into respective conditions , according to their average proportions after august 2001 .
it should be noted that despite clear " and thinorpartial " conditions , observations were sometimes impossible due to other weather warnings such as high humidity and strong wind .
moreover , observations were not carried out when the wet sensor warned that a dome was not dried out after rainfall or moisture .
concerning haleakala , more than 50@xmath160 % of all night time is near photometric , and in about 75 % , it is feasible to execute particular observations .
haleakala is therefore not worse than mauna kea where the observable sky rate is 60@xmath180 % as one of the best locations for optical and near - infrared observations in the northern hemisphere , along with good access .
next , figure [ monthly_weather ] shows the monthly average relative frequency of the whole - sky cloud condition .
clear seasonal cycles over a year can be seen ; there are high observable rates in summer and winter , and low rates in early spring and late autumn , in agreement with bradley et al .
it has generally been said that there are a dry summer season and rainy winter season in hawaii .
our data , however , demonstrates that midwinter is not very bad at haleakala , as far as sky condition is concerned .
we developed an infrared cloud monitor weather system that has been most successfully supporting an unmanned robotic telescope .
it sees almost whole - sky in thermal infrared with no field deformation , sensitively detects thin high - altitude clouds , automatically evaluates sky conditions , and withstands outdoor environments for several months without maintenance .
owing to this system , the magnum observatory has been achieving unmanned automated observation at haleakala for more than four years .
its evaluation of the whole - sky cloud condition being photometric , observable , or non - observable seems to be mainly successful .
it also proves that for optical and near - infrared observations , haleakala is a site comparable to mauna kea .
copies of our cloud monitor are now used for many similar systems at sites all around the world , including the atacama region in the northern part of chile .
the development of our cloud monitor was supported by the advanced technology center , national astronomical observatory of japan .
we are grateful to the technical center of nagoya university for our use of their ultra - precise turning machine .
we thank h. takami and n. takato for helpful discussion and advice on development .
we also thank m. doi , k. motohara , and colleagues at the haleakala observatories for their help with facility maintenance .
this research has been supported partly by the grant - in - aid of scientific research ( 10041110 , 10304014 , 11740120 , 12640233 , 14047206 , 14253001 , 14540223 , 16740106 , and 17104002 ) and coe research ( 07ce2002 ) of the ministry of education , science , culture and sports of japan . ll wavelength & @xmath67 + optics & thermal imager and two aspherical reflective + & mirrors with cassegrain - like alignment + sensor & micro - thermal bolometer array of 320 x 240pix + field of view & circular field of 11 - 70 degrees at zenith angle + pixel scale & 0.5 deg / pixel + sampling rate & one - or two - minute interval + sensitivity & @xmath68 0.015 for blackbody temperature of 240k + ll detector & uncooled micro bolometer array + format & @xmath69 pixels + wavelength & @xmath70 m + lens & @xmath71 mm , _
f_/0.7 @xmath72 @xmath73_f_/1.4 + field of view & 18@xmath74 + nedt & @xmath75 k for blackbody temperature + & around 25@xmath76c + analog output & ntsc + digital output & 12 bit parallel ( ttl standard signals ) + remote control & rs232c interface + frame rate & 30 hz + lcccccccccccc original ....... & & ( -52 , 595 ) & 530 & 85.0 & 11.0 & 240 & 80 & & 300 & 85.0 & 1.0 & 80 + improved ..... & & ( @xmath77 - 2 , 488 ) & 360 & 89.6 & @xmath784.4 & 240 & 38 & & 340 & 89.6 & 2.3 & 100 ccrccc @xmath85 & 0.36 & 73 & 0.108 & 0.006 & 0.338 + @xmath86 & 0.44 & 113 & 0.058 & 0.004 & 0.223 + @xmath87 & 0.55 & 154 & 0.033 & 0.004 & 0.111 + @xmath66 & 0.65 & 110 & 0.023 & 0.004 & 0.072 + @xmath88 & 0.79 & 97 & 0.020 & 0.004 & 0.057 + @xmath89 & 1.25 & 86 & 0.028 & 0.010 & 0.010 + @xmath90 & 1.63 & 95 & 0.027 & 0.009 & 0.009 + @xmath64 & 2.20 & 54 & 0.024 & 0.012 & 0.044 + clrrrrrrr 2001 & jan & 20 & 6981 & & 3.6 & 6.0 + & feb & 14 & 5121 & & 5.7 & 26.0 + & mar & 31 & 11097 & & 2.3 & 2.6 & 12.1 + & apr & 28 & 8946 & & 5.5 & 6.3 & 25.7 + & may & 31 & 11732 & & 1.9 & 0.8 & 3.7 + & jun & 30 & 12351 & & 9.5 & 10.3 & 14.3 + & jul & 31 & 12328 & & 12.3 & 10.4 & 14.5 + & aug & 31 & 13051 & 67.0 & 25.3 & 2.4 & 2.7 & 2.6 + & sep & 30 & 13210 & 55.1 & 25.1 & 3.1 & 8.8 & 7.9 + & oct & 31 & 14540 & 47.1 & 22.7 & 2.4 & 8.8 & 19.0 + & nov & 30 & 14424 & 46.9 & 22.9 & 2.1 & 9.4 & 18.7 + & dec & 7 & 3427 & 0.0 & 76.3 & 3.6 & 10.5 & 9.6 + 2002 & jan & 0 & 0 & ... & ... & ... & ... & ... + & feb & 0 & 0 & ... & ... & ... & ... & ... + & mar & 0 & 0 & ... & ... & ... & ... & ... + & apr & 0 & 0 & ... & ... & ... & ... & ... + & may & 0 & 0 & ... & ... & ... & ... & ... + & jun & 10 & 3021 & 18.3 & 71.2 & 3.6 & 4.9 & 2.0 + & jul & 31 & 11703 & 59.8 & 19.2 & 4.3 & 9.0 & 7.7 + & aug & 31 & 9792 & 33.3 & 28.0 & 5.9 & 15.6 & 17.2 + & sep & 30 & 9071 & 60.0 & 20.9 & 3.4 & 7.7 & 8.0 + & oct & 31 & 9777 & 42.9 & 19.3 & 6.3 & 8.8 & 22.7 + & nov & 30 & 12246 & 70.3 & 20.5 & 1.5 & 4.9 & 2.8 + & dec & 31 & 14665 & 79.7 & 8.4 & 0.7 & 1.5 & 9.7 + 2003 & jan & 31 & 14500 & 73.6 & 11.4 & 2.0 & 2.4 & 10.6 + & feb & 28 & 12611 & 80.8 & 6.1 & 0.7 & 3.6 & 8.8 + & mar & 31 & 13318 & 67.6 & 11.0 & 1.4 & 4.9 & 15.1 + & apr & 30 & 12291 & 37.0 & 35.6 & 3.9 & 6.7 & 16.8 + & may & 27 & 10326 & 54.6 & 28.7 & 6.6 & 9.8 & 0.3 + & jun & 30 & 11761 & 64.9 & 22.9 & 1.4 & 4.6 & 6.2 + & jul & 31 & 12212 & 55.6 & 19.7 & 3.8 & 8.9 & 12.0 + & aug & 31 & 12025 & 48.9 & 24.4 & 9.0 & 6.1 & 11.6 + & sep & 29 & 11754 & 82.2 & 11.6 & 0.8 & 2.6 & 2.8 + & oct & 31 & 13856 & 43.9 & 27.6 & 4.0 & 11.6 & 12.9 + & nov & 30 & 13992 & 53.3 & 20.0 & 4.3 & 6.3 & 16.1 + & dec & 31 & 14755 & 55.5 & 8.7 & 2.3 & 4.4 & 29.1 + 2004 & jan & 31 & 14590 & 52.8 & 10.6 & 2.1 & 5.5 & 29.0 + & feb & 29 & 13095 & 62.4 & 11.2 & 1.7 & 9.0 & 15.7 + & mar & 31 & 13338 & 20.6 & 17.3 & 2.8 & 17.0 & 42.3 + & apr & 30 & 12444 & 50.5 & 12.8 & 2.5 & 5.8 & 28.4 + & may & 31 & 12569 & 39.1 & 15.1 & 3.4 & 9.5 & 32.9 + & jun & 30 & 11991 & 52.5 & 24.4 & 3.0 & 9.8 & 10.3 + & jul & 31 & 12490 & 80.0 & 9.7 & 1.1 & 3.3 & 5.9 + & aug & 31 & 12837 & 64.9 & 13.9 & 1.3 & 2.3 & 17.6 + & sep & 30 & 12877 & 49.6 & 13.9 & 2.4 & 7.9 & 26.2 + & oct & 31 & 13874 & 48.4 & 18.9 & 3.1 & 8.1 & 21.5 + & nov & 30 & 13994 & 39.6 & 23.8 & 2.5 & 4.9 & 29.2 + & dec & 31 & 14930 & 43.6 & 20.9 & 3.0 & 8.6 & 23.9 + 2005 & jan & 31 & 14748 & 56.1 & 12.9 & 1.7 & 3.1 & 26.2 + & feb & 28 & 12850 & 57.8 & 12.7 & 1.1 & 4.1 & 24.3 + & mar & 31 & 13590 & 43.6 & 16.8 & 3.0 & 10.8 & 25.8 + & apr & 30 & 12591 & 58.2 & 21.6 & 2.3 & 4.7 & 13.2 + & may & 31 & 12385 & 63.6 & 21.3 & 2.0 & 4.4 & 8.7 + & jun & 30 & 11855 & 57.0 & 16.1 & 1.8 & 5.5 & 19.6 + & jul & 31 & 12409 & 70.3 & 17.1 & 1.5 & 2.8 & 8.3 + & aug & 31 & 12737 & 77.0 & 14.8 & 1.1 & 2.5 & 4.6 + & sep & 30 & 12462 & 56.2 & 12.0 & 3.8 & 4.0 & 24.0 + & oct & 31 & 13579 & 52.0 & 15.3 & 4.5 & 10.4 & 17.8 + & nov & 30 & 14078 & 46.8 & 19.4 & 3.9 & 10.0 & 19.9 + & dec & 31 & 14866 & 79.5 & 12.9 & 0.7 & 3.1 & 3.8 + | we present the most successful infrared cloud monitor for a robotic telescope .
this system was originally developed for the magnum 2-m telescope , which has been achieving unmanned and automated monitoring observation of active galactic nuclei at haleakala on the hawaiian island of maui since 2001 . using a thermal imager and two aspherical mirrors
, it at once sees almost the whole sky at a wavelength of @xmath0 .
its outdoor part is weather - proof and is totally maintenance - free .
the images obtained every one or two minutes are analysed immediately into several ranks of weather condition , from which our automated observing system not only decides to open or close the dome , but also selects what types of observations should be done .
the whole - sky data accumulated over four years show that 50@xmath160 % of all nights are photometric , and about 75 % are observable with respect to cloud condition at haleakala .
many copies of this system are now used all over the world such as mauna kea in hawaii , atacama in chile , and okayama and kiso in japan . |
the problem of formation of a cylindrical beam of charged particles with a predefined cross - section arises in development of devices such as multi - aperture particle sources for plasma heating and diagnostics @xcite .
such devices compose a cylindrical cutting with a required cross - section from one - dimensional flow of charged particle produced in a planar diode .
the particles flow in an idealized endless planar diode is intrinsically rectilinear .
it is characterized by zero angular divergence and zero emittance , which means that the particles trajectories do not intersect .
cutting a piece of the flow gives rise to lateral electric field that increases the angular spread of the particles .
in many cases , the angular spread should be avoided . according to j.r .
pierce @xcite , the transversal field can be compensated in the entire beam interior by proper shaping of the external electrodes . from the point of view of pure mathematics , the synthesis of such ( pierce ) electrodes belongs to a class of ill - posed cauchy s problems for the laplace equation @xcite .
it assumes a computation of the electrostatic potential outside of the beam that provides the required compensation of the transversal electric field . for a particle beam in the form of endless belt , the pierce problem is two - dimensional .
it was solved by pierce himself .
he noted that required solution is given by an analytic continuation of the child - langmuir potential @xmath0 @xcite , which describes 1d electric field inside a space - charge - limited endless diode . for a cylindrical beam with an arbitrary cross - section ,
the pierce problem becomes three - dimensional and the method of analytical continuation fails .
a general solution of the pierce problem in 3d case was found by v.a .
syrovoi @xcite ( see also @xcite ) .
his algorithm involves elimination of the dependence of the potential on coordinate @xmath1 along the beam by means of the laplace transform , two analytical continuations , application of the riemann method to solve an intermediate 2d problem , and an inverse integral transform using the lipschitz
hankel integral . in this note , we apply a slightly modified method of syrovoi to a beam of charged particles with a cross - section , close to a polygon ; in particular , we present an exact solution of the pierce problem near the right angle of such a beam .
this solution is then extended to the case of an arbitrary angle .
the paper is organized as follows . in section [ s2 ]
we formulate a formal solution of the 3d pierce problem for a rectilinear beam of charged particles . in section [ s3 ]
we compute electric potential outside of the beam near the right angle of the beam cross - section . in section [ s4 ]
this solution is generalized to an arbitrary angle .
finally , in section [ s8 ] we summarize our results .
to begin with , we first formulate the pierce problem for a cylindrical beam with an arbitrary cross - section .
our goal is to find a solution of the laplace equation @xmath2 outside of a cylindrical domain assuming the cauchy boundary conditions @xmath3 on the side boundary of the cylinder , which is given in a parametric form as @xmath4 where @xmath5 is a formal parameter , and @xmath6 denotes a derivative , normal to the boundary .
the boundary conditions specify the cauchy problem for the laplace equation @xcite .
unlike neumann and dirichlet problems for elliptic equations ( see , eg , @xcite ) , it belongs to a class of ill - posed problems of mathematical physics .
as the hadamard example shows @xcite , a solution to the cauchy problem for the laplace equation is unique , but unstable with respect to small perturbations of the boundary conditions . from the point of view of physics , the space charge of a beam , bounded in transverse directions
, inevitably creates a lateral electric field that increases angular divergence of the flux ( ie , the emittance of the beam ) .
the lateral field can be compensated in the entire beam interior by a careful shaping of additional electrodes placed outside of the beam .
pierce @xcite found the shape of such electrodes in 2d problem by means of analytic continuation of the 1d solution @xmath0 to the complex plane @xmath7 .
pierce s solution @xmath8 refers to the case of a beam that occupies a half - space @xmath9 .
in particular , the shape of the electrode with zero potential is determined from the equation @xmath10 it has the shape of a plane , which forms an angle @xmath11 with the plane of the beam boundary .
exact solutions for the beams with circular and elliptical cross - sections were found in refs .
finally , v.a .
syrovoi found a general solution of the pierce problem for a cylindrical beam with a cross - section of arbitrary shape @xcite .
however , his work remains unappreciated , and other authors rarely refer to it . perhaps , this is due to the complexity of syrovoi s solution .
it includes double analytic continuation , and a final expression involves integration of the hypergeometric function whose argument involves analytic continuation of the functions @xmath12 , @xmath13 in a nontrivial way . in addition , syrovoi has not presented a clear evidence that the pierce solution can be obtained from his formulas .
equivalence of syrovoi s solutions for the beams with circular and elliptical cross - sections @xcite to solutions of other authors @xcite is not also verified . in this paper
, we use the method of syrovoi to find the electric potential near one individual right corner of the rectangle , assuming that the beam occupies a quarter of the space , namely @xmath14 in this case , a conformal mapping of the exterior of the beam on the complex half - plane is made by a power function , which greatly simplifies the calculations .
let the cylinder guide @xmath15 is defined by parametric equations .
it is easy to see that @xmath16 maps the real axis @xmath17 in the complex plane @xmath18 on @xmath15 in the plane @xmath19 .
specifically , the transformation @xmath20 maps the half - plane @xmath21 on the area immediately outside of the quadrant .
the transformation is unique in the plane @xmath22 with a cut along the negative real half - axis @xmath23 ; hereinafter we choose the principal value of the power function @xmath24 , which assumes that the argument @xmath5 is brought to the interval @xmath25 by the modulo @xmath26 . figure [ fig : map ] illustrates the mapping . ,
@xmath27 onto the plane @xmath28 .
, title="fig : " ] + separating the real and imaginary parts in eq .
, we find @xmath29 , \\
y(u , v ) & = \frac{1}{2}\left [ \left ( -iu+v \right)^{3/2 } + \left ( iu+v \right)^{3/2 } \right ] . \end{aligned } \end{gathered}\ ] ] the inverse transformation @xmath30 , \\
v(x , y ) & = \frac{1}{2 } \left [ ( y - i x)^{2/3}+(y+i x)^{2/3 } \right ] \end{aligned } \end{gathered}\ ] ] maps the exterior of the quadrant onto the upper half - plane @xmath22 .
we will further assume that the beam corner can be smoothed by choosing a coordinate line @xmath31 for the role of the beam boundary . in fig .
[ fig : map ] , the coordinate lines @xmath32 encircle the right angle , and the coordinate lines @xmath33 are orthogonal to them .
eqs . and
define the first analytic continuation of the two continuations used in syrovoi s theory .
the second analytic continuation is made by replacing @xmath34 in the laplace equation and the boundary conditions on @xmath15 .
such a substitution transforms the laplace equation into a hyperbolic equation that can be solved in a general form by the riemann method using laplace transform .
we give a final result in a form somewhat different from that obtained by syrovoi .
the electric potential at a point with coordinates @xmath35 , @xmath36 and @xmath37 is given by the expression @xmath38 where @xmath39 denotes the gauss hypergeometric function , parameters @xmath40 , @xmath41 are related to @xmath42 , @xmath43 by eq . , @xmath44^{2 } + \left [ y(u , v)-y(u+i\xi , v_{b } ) \right]^{2 } \\ = \left [ ( i u+v)^{3/2}-(iu+v_{b}-\xi)^{3/2 } \right ] \times \\
\left [ ( -iu+v)^{3/2}-(-iu+v_{b}+\xi)^{3/2 } \right ] , \end{gathered}\ ] ] and the equations @xmath45 , @xmath46 define a smoothed boundary of the beam . since @xmath47 is a multi - valued function , a particular branch of @xmath47 in the integral should be thoroughly chosen .
this is provided by making cuts in the complex plane @xmath48 . and what is more important , these cuts should conform to the rules used by computational algorithms when evaluating the power functions such as @xmath49 .
we note that writing down same expression in different forms leads to different results .
for example , @xmath49 is not the same as @xmath50 since these two forms can be evaluated to different complex numbers for the same values of @xmath40 , @xmath51 , and @xmath52 .
we carefully tuned the expressions , , and to achieve a satisfactory results so that @xmath47 has two cuts in the complex plane @xmath48 along the positive half of axis @xmath52 from @xmath51 to @xmath41 and along the negative half from @xmath53 to @xmath54 for any @xmath41 in the interval of integration over @xmath52 from @xmath55 to @xmath56 ( recall that @xmath57 ) .
the result of numerical integration of eq . near a smoothed beam corner
is shown in fig . [
fig : potential ] for @xmath58 .
pink lines indicate the position of the fractures on the equipotential surfaces .
the electric potential @xmath59 is continuous on the fractures but its normal derivative @xmath60 jumps ; it means that the pink surfaces bear a nonzero surface electric charge .
thus , despite the smoothness of the contour @xmath45 , @xmath46 with @xmath61 and the conformal mappings , eq
. defines a discontinuous solution . in the band
@xmath62 the solution is smooth , but further away from the boundary of the beam the equipotential surfaces experience a kink on the line @xmath63 .
this phenomenon is formally related to the fact that the imaginary part of the function is discontinuous at @xmath64 if @xmath65 . from the physical point of view
, the presence of the fracture means the impossibility of forming a cylindrical beam with zero emittance and relevant cross - section without introducing charged surfaces , which approach the boundary of the beam as the radius of rounded corners decreases .
numerical integration of eq . for the case @xmath66
yields a quite expected result .
the potential @xmath59 outside the beam is formed by the two 2d pierce solutions clued at the surface @xmath63 .
the solution , presented in sec .
[ s2 ] , can be readily generalized to the case of a beam edge with an angle of arbitrary magnitude .
assuming that the beam occupies the region @xmath67 we need to replace eq . by the generalized transformation @xmath68 then , eq
. reads @xmath69 , \\
y(u , v ) & = \frac{1}{2}\left [ \left ( -iu+v \right)^{(2\pi-2\phi)/\pi } + \left ( iu+v \right)^{(2\pi-2\phi)/\pi } \right ] , \end{aligned } \end{gathered}\ ] ] and the inverse transformation becomes @xmath70 , \\
v(x , y ) & = \frac{1}{2 } \left [ ( y - i x)^{\pi/(2\pi-2\phi)}+(y+i x)^{\pi/(2\pi-2\phi ) } \right ] . \end{aligned } \end{gathered}\ ] ] when the angle @xmath71 varies from @xmath72 ( cut angle ) to @xmath73 ( obtuse angle ) the factor @xmath74 increases from @xmath75 to @xmath76 ; @xmath77 corresponds to the right angle .
the main expression remains unchanged but eq .
takes the form @xmath78
\times \\ \left [ ( -iu+v)^{(2\pi-2\phi)/\pi}-(-iu+v_{b}+\xi)^{(2\pi-2\phi)/\pi } \right ] . \end{gathered}\ ] ] main conclusions made in sec .
[ s3 ] remain valid for arbitrary @xmath71 . in particular
, the fracture appears for any @xmath79 , ie .
, except for the case of planar beam boundary , considered by pierce
. the pierce solution for a beam with plane boundary can be recovered from eqs . and .
to proof this statement , it is sufficient to take @xmath80 in eq . .
then , @xmath81 and eq
. takes the form @xmath82 if @xmath83 . after the substitution @xmath84 ,
@xmath85 a computational software program _ mathematica _
@xcite computes the integral in an analytic form and returns the expression @xmath86 which is equivalent to eq . .
this proofs that syrovoi s theory contains the pierce solution for a planar beam in the form of endless belt as a particular case .
our study has confirmed validity of the theory developed by v.a .
syrovoi,@xcite which seems to be not properly evaluated by other researchers . in particular , we have shown that classical 2d pierce s solution @xcite can be deduced from that of syrovoi .
we found a solution of the 3d pierce problem near the sharp or rounded right angle of a rectilinear flow of charged particles .
this solution is then generalized to the angle of arbitrary magnitude .
although it demonstrates principal impossibility of constructing the pierce electrodes because of appearance of fractures in the spatial dependance of the electric potential for a beam with a cross - section that has sharp corners , this solution can be used as a benchmark for existing numerical codes when designing devices with minimized beam emittance . for practical needs
, it could also be sufficient to round the corners of the cross - section to achieve a smooth single - valued solution which would mean feasibility of electrodes with pierce s geometry in a vicinity of the beam boundary .
the work was supported by the ministry of education and science of russian federation ( project rfmefi61914x0003 ) .
the author is grateful to a.a .
ivanov , v.i .
davydenko , a.d .
beklemishev , and for useful discussions .
19ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty \doibase http://dx.doi.org/10.1063/1.2814248 [ * * , ( ) ] link:\doibase 10.1063/1.2149392 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.nima.2007.03.009 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2801550 [ * * , ( ) ] \doibase
[ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4792783 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , ed .
( , ) link:\doibase 10.1080/00207215808953831 [ * * , ( ) ] @noop * * , ( ) , link:\doibase 10.1134/s1064226912060125 [ * * , ( ) ] http://books.google.ru/books?id=u1x1zweacaaj[__ ] ( , , ) @noop * * , ( ) , @noop _ _ , ed .
( , , ) @noop _ _ ( , , ) @noop * * , ( ) , @noop * * , ( ) , link:\doibase 10.1103/physrevstab.8.014201 [ * * , ( ) ] http://www.wolfram.com/mathematica/ [ `` , '' ] ( ) | an ill - posed problem of synthesis of the pierce electrodes for a cylindrical beam with a polygonal cross - section is considered .
it is assumed that a beam of charged particles is extracted from a space - charge - limited planar diode and the pierce electrodes outside of the beam ensure its zero angular divergence .
a mathematical statement of the problem presumes a computation of the electrostatic potential outside of the beam that should match the child - langmuir 1d potential inside of the beam .
an exact solution is first obtained for the potential outside of the beam near its right angle .
the solution involves double analytic continuation and a numerical integration of the hypergeometric function and can be used as a benchmark for testing numerical codes .
it is shown that equipotential surfaces have fractures that can be pushed away from the corner of the beam by means of smoothing the beam corners .
this solution is then generalized to an angle of arbitrary magnitude . |
the largest and most powerful solar telescopes are ground - based facilities . however , compared to their counterparts in space such as the solar dynamics observatory ( ( * ? ? ?
* sdo ; pesnell et al .
2012 ) ) , they often lack in providing user - friendly data pipelines to reduce the raw data .
only few ground - based instruments have such a pipeline ( ( * ? ?
* e.g. , de la cruz rodrguez 2015 ) ) .
the data reduction process is often much more complex than for space missions and differs substantially from instrument to instrument .
the need for automatized data reduction pipelines to produce science - ready data becomes crucial in the era of big data .
therefore , in the framework of the eu - funded solarnet project , we developed a pipeline called `` stools '' to reduce and prepare data acquired with the gregor fabry - prot interferometer ( ( * ? ? ?
* gfpi ; puschmann 2012 ) ) , large - format imaging cameras , and the high - resolution fast imager ( hifi ) located at europe s largest solar telescope gregor ( ( * ? ? ?
* schmidt 2012 ) ) . in sect .
2 we will briefly describe the instruments supported by the pipeline .
section 3 is dedicated to the description of the pipeline , and sect .
4 reports on the data archive and provides an outlook .
the gfpi is the successor of the gttingen fabry - prot interferometer ( ( * ? ? ?
* bendlin 1992 ) , ( * ? ? ?
* puschmann 2006 ) ) which was attached to the vacuum tower telescope ( vtt ) at the observatorio del teide and is now installed at gregor .
it is an imaging spectropolarimeter which operates in the visible and near infrared spectral range .
the gfpi has two tunable etalons in a collimated setup .
the two ccd cameras , one for the broad - band and another for the narrow - band images , have the same image scale and a size of @xmath0 pixels . the spatial scale of gfpi was derived comparing gfpi broad - band images from the year 2014 with continuum images from sdo which results in 0@xmath10405pixel@xmath2 . hence , the field - of - view ( fov ) of the gfpi in spectroscopic mode is @xmath3 .
the blue imaging channel ( bic ) of the gfpi is fed by light in the visible below @xmath45300 .
a beamsplitter allows for simultaneous image acquisition at two different wavelengths with two cameras .
the observer aims for high - cadence with these cameras to minimize atmospheric seeing effects and assure high - quality image reconstruction .
initially , two pco.4000 ccd cameras with a size of @xmath5 pixels and a spatial sampling of 0@xmath10347pixel@xmath2 were used . in 2016 , these cameras were replaced by two synchronized scmos imagers and the instrument is now named high - resolution fast imager ( hifi ) .
both cameras write their images into the same file .
this requires the same exposure times for both cameras which can be achieved using neutral density filters .
the hifi chips have a size of @xmath6 pixels and a spatial sampling of 0@xmath10253pixel@xmath2 . of special interest
is the frame rate of these cameras which is 49hz or 98hz with the full or a @xmath7 pixels fov , respectively .
hifi can be used independently from the gfpi , e.g. , as the context image of the gregor infrared spectrograph ( ( * ? ? ?
* gris , collados 2012 ) ) .
a careful acquisition of calibration data each day is crucial to produce high - quality science data after the reduction process .
the following calibration data are required : dark , flat field , resolution target , and several pinhole images .
in addition , the spectroscopic mode of the gfpi requires a long scan to derive the prefilter curve of the interference filter .
it is crucial to have dark images with the same exposure times as all the acquired data .
all routines are written in the interactive data language ( idl ) and are documented in the header . moreover ,
the routines are strictly named with the prefix `` stools '' in order to avoid overlapping with other idl routines .
gfpi and hifi data are recorded in the native format of davis , a software package from the company lavision gmbh , gttingen . we will focus on the data reduction of the imaging spectroscopic mode . all computed images and parameters for the data calibration are stored in one single idl save file .
single variables are added to or are extracted from idl save files with special stools routines . to initialize the pipeline
, the user only needs to specify the in and output directories , the observed wavelength , and the on - chip binning of the cameras .
first , the average dark images are computed for both cameras . in the next step
the flat - field images are averaged for each position of the etalon .
other valuable information such as the positions of the etalon , accumulations , and step size is also saved .
the pinhole and resolution target images are corrected for dark and flat - field .
they are then used to derive the alignment parameters between both gfpi cameras such as rotation , displacement , pivot point , and magnification .
the output is stored in an idl structure .
the blueshift correction needs to be performed across the fov of the narrow - band images . for this task ,
flat - field images are taken and interpolated to a narrower and equidistant wavelength grid .
the resulting spectral profiles for each pixel are slightly smoothed by convolving them with a normalized kernel of @xmath8 units .
a polynomial fit of second order to the line core of the spectral line yields the central position on the wavelength axis .
we repeat this step for all spectra across the fov .
the average of all central positions is then computed and subtracted from each individual position . hence , we obtain the displacement in wavelength , i.e. , the blueshift correction , for all pixels within the fov .
the correction is applied by displacing the spectra using cubic spline interpolation . for the narrow - band camera , the aforementioned flat field
can now be used together with the blueshift correction to compute a normalized master flat , i.e. , a single two - dimensional gain table , where the spectral information is removed .
the user can choose either this master flat or the former spectral - dependent flat field to correct the data .
kms@xmath2.,scaledwidth=99.0% ] kms@xmath2.,scaledwidth=99.0% ] the last step before correcting the science observations is to derive the prefilter curve of the gfpi . to this end
, we use a long scan with a small step size covering a large fraction of the etalon s scan range .
this scan is ideally taken at the solar disk center under quiet - sun conditions to avoid spectral line shifts .
after dark , flat - field , and blueshift corrections a gfpi mean spectrum , which represents the average transmission profile of the narrow - band interference filter , is computed . to extract the trend of the prefilter curve we take advantage of the fourier transform spectrometer ( ( * ? ? ?
* fts ; neckel & labs 1984 ) ) spectrum from the kitt peak national observatory .
the fts spectrum is first convolved with the theoretical gfpi transmission profile and then matched to the observed gfpi mean spectrum by minimizing the wavelength offset .
it follows a resampling of the fts spectrum to the gfpi wavelength scale .
afterwards , the ratio of both spectra is computed .
finally , the ratio , with the exception of parts showing large variations , is fitted using a double gaussian .
linear least - squares fitting between model and observations is performed with the mpfit package ( ( * ? ? ? * markwardt 2009 ) ) .
all calibration data are now ready to be applied to the science data .
the pipeline splits now into two different branches . on the one hand ,
quick - look physical maps are being generated , e.g. , line - core intensity images , line - of - sight ( los ) velocity maps , equivalent width maps , seeing quality parameters , etc .
the results are stored into an idl structure together with the reduced broad- and narrow band images ( level 1.0 data ) . on the other hand ,
the data are prepared for image - restoration ( level 2.0 data ) with multi - object multi - frame blind deconvolution ( ( * ? ? ?
* momfbd ; lfdahl 2002 ) , ( * ? ? ?
* van noort 2005 ) ) .
the restored images are finally written in the flexible image transport system format ( ( * ? ? ?
* fits , wells 1981 ) ) .
one example of a restored line - core image and its associated los velocity map is shown in fig .
[ fig1 ] .
we follow a similar approach for the hifi data reduction .
the user needs to introduce the directory of the observations , telescope and user names , and the two wavelengths of the respective interference filters .
the average dark , flat field , pinholes , and resolution target images are then computed and saved as individual fits files .
the file name has a two - digit suffix which identifies the type of calibration data .
moreover , the fits header includes basic information such as date and time of the observations , wavelength , telescope , dimensions of the images , etc .
the target frames are used to align both cameras .
the obtained parameters such as displacement , rotation , and magnification are stored in an idl save file . due to the setup of the cameras , both have the same image scale , however , there might be a slight displacement in the vertical and horizontal axes .
the science observations are then corrected for dark and for artifacts on the chip by using the flat - field images .
if more than one burst of flat - field images was taken , the pipeline chooses the one closest in time to the science observations .
an image quality check to sort the images is performed at this point .
the reason for this is that hifi , with its high frame rate , acquires more images than actually needed for the image restoration .
usually 500 images are taken per burst and only the 100 best are used for the restoration .
the other images are dropped .
however , the user can change this criterion .
we take advantage of the median filter - gradient similarity ( mfgs ) introduced by ( * ? ? ?
* deng ( 2015 ) ) , to scrutinize the image quality .
the mfgs code is implemented in an independent idl routine and is applied to every single image .
each pixel has now a value between 0 and 1 .
this value and additional statistics related to the mfgs are stored in a separate idl save file and also added to the fits header .
finally , the average mfgs value over the whole image is used as a reference . note that the mfgs value is only calculated for the images of one of the cameras .
this is justified since both cameras strictly acquire images at the same time , hence , both were recorded under the same seeing conditions .
finally , the sorted images according to their mfgs value ( level 1.0 data ) are written into a fits file using standard routines of the idl astronomy user s library hosted at .
the two bursts of images are written alternately in a single fits file using image extensions ( ( * ? ? ?
* ponz 1994 ) ) .
odd extension numbers belong to images from one filter whereas even numbers belong to the other filter .
the file has a detailed primary header with all relevant information about the observations .
furthermore , there are extension headers for each image showing only image specific information , e.g. , the wavelength , image statistics , and mfgs information .
the data restoration is carried out using the speckle - interferometry code kisip ( ( * ? ? ?
* wger & von der lhe 2008 ) ) to produce science - ready data ( level 2.0 data ) as shown in fig .
gfpi and hifi data are stored at the data archive of aip ( http://gregor.aip.de[gregor.aip.de ] ) . in addition , the pipeline generates a webpage , which includes an overview of the acquired observations as well as quick - look images and statistics .
the data recorded since 2014 will be made public to the solar community .
the stools pipeline is under constant development .
an official version will be released in 2017 via a version control system on the abovementioned webpage under creative commons license .
this work was carried out as a part of the solarnet project supported by the european commission s 7th framework programme under grant agreement no .
the 1.5-meter gregor solar telescope was built by a german consortium under the leadership of the kiepenheuer - institut fr sonnenphysik in freiburg ( kis ) with the leibniz - institut fr astrophysik potsdam ( aip ) , the institut fr astrophysik gttingen ( iag ) , the max - planck - institut fr sonnensystemforschung in gttingen ( mps ) , and the instituto de astrofsica de canarias ( iac ) , and with contributions by the astronomical institute of the academy of sciences of the czech republic ( ascr ) .
sjgm is grateful for financial support from the leibniz graduate school for quantitative spectroscopy in astrophysics , a joint project of aip and the institute of physics and astronomy of the university of potsdam . | a huge amount of data has been acquired with the gregor fabry - prot interferometer ( gfpi ) , large - format facility cameras , and since 2016 with the high - resolution fast imager ( hifi ) .
these data are processed in standardized procedures with the aim of providing science - ready data for the solar physics community . for this purpose ,
we have developed a user - friendly data reduction pipeline called `` stools '' based on the interactive data language ( idl ) and licensed under creative commons license .
the pipeline delivers reduced and image - reconstructed data with a minimum of user interaction .
furthermore , quick - look data are generated as well as a webpage with an overview of the observations and their statistics .
all the processed data are stored online at the gregor gfpi and hifi data archive of the leibniz institute for astrophysics potsdam ( aip ) .
the principles of the pipeline are presented together with selected high - resolution spectral scans and images processed with stools . |
in recent years , the multisensor estimation fusion or data fusion has received significant attention for target tracking , artificial intelligence , sensor networks and big data ( see @xcite ) , since many practical problems involve information or data from multiple sources .
the problem of multisensor estimation fusion is that how to optimally fuse sensor data from multiple sensors to provide more useful and accurate information for the purpose of estimating an unknown process state @xcite .
currently , the estimation fusion technology has rapidly evolved from a loosely related techniques to an emerging real engineering discipline with standardized terminology @xcite .
generally speaking , there are two traditional architectures for estimation fusion , namely , centralized fusion structure and distributed fusion structure . the centralized architecture is sending the raw data of each sensor to the fusion center , theoretically , which is nothing but an estimation problem with distributed data
. moreover , the centralized fusion approach can usually reach optimal linear estimation in mean squared error ( mse ) sense @xcite . however , the distributed architecture is propagating the estimation of each sensor to the fusion center , which decreases computational burden in the fusion center , but it may not get the optimal linear estimation in mse sensse . due to its important practical significance , distributed estimation fusion has been studied extensively , see @xcite , @xcite , @xcite , @xcite , @xcite . for multisensor point estimation fusion in probabilistic setting ,
many results have been obtained ( see , e.g. , books @xcite , @xcite , @xcite ) .
@xcite provides the optimal linear estimation fusion method for a unified linear model .
@xcite proves that the distributed fusion algorithm is equivalent to the optimal centralized kalman filtering in the case of cross - uncorrelated sensor noises , and the one for the case of cross - correlated sensor noises is proposed in @xcite .
when there exists the limitation of communication bandwidth between a fusion center and sensors , @xcite achieves a constrained optimal estimation at the fusion center .
in addition , @xcite proposes lossless linear transformation of the raw measurements of each sensor for distributed estimation fusion .
most existing information fusion algorithms are based on the sequential estimation techniques such as kalman filter , information filter and the weighted least - squares methods @xcite , which need to know the accurate statistical knowledge of the process and measurement noises . since the limitation of human and material resources in real life
, we can not obtain the exact statistical characteristics of noise , which may lead to poor performance for the state estimation ( see @xcite , @xcite ) .
especially for the nonlinear target tracking systems , it is more sensitive to the precise distribution information of noise . in many engineering applications ,
it is easier to obtain the upper bound and lower bound of a unknown noise @xcite . in the unknown but bounded setting ,
the earliest work about the set - membership filter is proposed by @xcite at the end of 1960s , and it is later developed by @xcite and @xcite .
these robust filters are derived through set - membership estimate , usually a bounding ellipsoid of containing the true state . moreover , the set - membership filter for nonlinear dynamic system has also been investigated by @xcite , @xcite , @xcite and references therein
. for multisensor set - membership fusion in bounded setting , @xcite proposes a relaxed chebyshev center covariance intersection ( ci ) algorithm to fuse the local estimates , geometrically , which is the center of the minimum radius ball enclosing the intersection of estimated ellipsoids of each sensor . in order to account for the inconsistency problem of the local estimates
, @xcite proposes a covariance union method ( cu ) and it is more conservative than ci fusion .
however , the judgment and calculation about correlation may be difficult .
since the set - membership filter only needs to know the bound of the noises , rather than the statistical properties of noises , it does not require to judge the correlation between each sensor , which inspires us to consider set - membership information fusion .
when the dynamic system is linear dynamic systems , @xcite proposes some algorithms of multisensor set - membership information fusion to minimize euclidean estimation error of the state vector .
however , for nonlinear dynamic systems , the multisensor set - membership information fusion has not received enough research attention .
these facts motivate us to further research the more challenging set - membership fusion problem for nonlinear dynamic systems . in this paper ,
two popular fusion architectures are considered : centralized and distributed set - membership information fusion .
firstly , both of them can be converted into a semidefinite programming ( sdp ) problem which can be efficiently computed , respectively .
secondly , their analytical solutions can be derived surprisingly by using decoupling technique .
it is very interesting that they are quite similar in form to the classic information filter in mse sense @xcite . in the two analytical fusion formulae
, the information of each sensor can be clearly characterized , and the knowledge of the correlation among measurement noises across sensors are not required .
finally , multi - algorithm fusion is used to minimize the size of the state bounding ellipsoid by complementary advantages of multiple parallel algorithms .
a typical numerical example in target tracking demonstrates the effectiveness of the centralized , distributed , and multi - algorithm set - membership fusion algorithms .
in particular , it shows that multi - algorithm fusion performs better than both the centralized and distributed fusion .
the rest of the paper is organized as follows .
section [ sec_2 ] introduces the problem formulation for the centralized fusion and the distributed fusion . in section [ sec_3 ] ,
the centralized set - membership information fusion algorithm is derived by @xmath0-procedure , schur complement and decoupling technique .
section [ sec_4 ] provides the distributed set - membership information fusion algorithm . a typical example in target tracking
is presented in section [ sec_5 ] , while conclusion is drawn in section [ sec_6 ] .
consider the @xmath1-sensor _ centralized _ nonlinear dynamic system with unknown but bounded noises as follows : @xmath2 where @xmath3 is the state of system at time @xmath4 , @xmath5 is the measurement at the @xmath6th sensor , @xmath7 , @xmath8 is the nonlinear function of the state @xmath9 , @xmath10 is nonlinear measurement function of @xmath9 at the @xmath6th sensor , @xmath11 is the uncertain process noise and @xmath12 is the uncertain measurement noise .
assume that @xmath13 and @xmath14 are confined to specified ellipsoidal sets @xmath15 where @xmath16 and @xmath17 are the _ shape matrix _ of the ellipsoids @xmath18 and @xmath19 , @xmath20 , respectively .
both of them are known symmetric positive - definite matrices .
suppose that when the nonlinear functions are linearized , the remainder terms can be bounded by an ellipsoid , respectively .
specifically , by taylor s theorem , @xmath21 and @xmath22 can be linearized to @xmath23 where @xmath24 , @xmath25 , are jacobian matrices . @xmath26 and @xmath27 are high - order remainders , which can be bounded in an ellipsoid for @xmath28 , @xmath7 , respectively , i.e. , @xmath29 where @xmath30 and @xmath31 are the centers of the ellipsoids @xmath32 and @xmath33 , respectively ; @xmath34 and @xmath35 are the shape matrices of the ellipsoids @xmath32 and @xmath33 , respectively .
note that @xcite proposes the monte carlo methods for the bounding ellipsoids of the remainders , which can effectively take advantage of the character of the nonlinear functions , and it can obtain the tighter bounding ellipsoids @xmath32 and @xmath33 to cover the remainders on line . the corresponding centralized set - membership information fusion problem can be formulated as follows .
assume that the initial state @xmath36 belongs to a given bounding ellipsoid : @xmath37 where @xmath38 is the center of ellipsoid @xmath39 , and @xmath40 is the shape matrix of the ellipsoid @xmath39 which is a known symmetric positive - definite matrix . at time
@xmath4 , given that @xmath9 belongs to a current bounding ellipsoid : @xmath41 where @xmath42 is the center of ellipsoid @xmath43 , and @xmath44 is a known symmetric positive - definite matrix . at next time @xmath45 ,
the fusion center can obtain the measurements @xmath46 from the @xmath6th sensor , @xmath20 .
for the centralized fusion system , the goal of the fusion center is to determine a prediction ellipsoid @xmath47 and an estimation ellipsoid @xmath48 at time @xmath45 .
firstly , in prediction step , we look for @xmath49 and @xmath50 such that the state @xmath51 belongs to @xmath52 whenever i ) @xmath9 is in @xmath43 , ii ) the process noise @xmath53 , and iii ) the remainder @xmath54 .
secondly , in the fusion update step , we look for @xmath55 and @xmath56 such that the state @xmath51 belongs to @xmath57 whenever i ) @xmath58 is in @xmath47 , ii ) measurement noises @xmath59 , @xmath20 , and iii ) the remainders @xmath60 , @xmath20 .
moreover , we provide a state bounding ellipsoid by minimizing its size " at each time which is a function of the shape matrix @xmath61 denoted by @xmath62 . if we choose trace function , i.e. , @xmath63 , which means the sum of squares of semiaxes lengths of the ellipsoid @xmath64 , the other common size " of the ellipsoid is @xmath65 , which corresponds to the volume of the ellipsoid @xmath64 . in order to emphasize the importance of the interested state vector entry
, @xcite proposes an objective of the ellipsoid @xmath64 as follows @xmath66 where @xmath67 is the weight coefficient with @xmath68 , and @xmath69 denotes the element in the @xmath6th row and the @xmath6th column of the matrix @xmath61 , @xmath20 .
if the bound of the @xmath6th entry of the interested state vector is very important , we can give a larger weight to @xmath67 .
when @xmath70 , @xmath20 , which means that each entry of the state vector is treated equally , and it is also equivalent to the trace function .
therefore , we can use multi - algorithm fusion to obtain multiple bounding estimated ellipsoids , which squashed along each entry of the state vector as much as possible based on different weighted objective ( [ eqpre_170 ] ) , then the intersection of these bounding ellipsoids can derive a final state bounding ellipsoid with a smaller size . in this paper , we also consider @xmath1-sensor _ distributed _ estimation fusion for the nonlinear dynamic system ( [ eqpre_1 ] ) and ( [ eqpre_2 ] ) .
the problem is formulated as follows . at time @xmath45
, the @xmath6th local sensor can use the measurements @xmath71 to obtain the bounding ellipsoid @xmath72 by the single sensor recursive method @xcite .
then , the local estimated ellipsoids @xmath72 are sent to the fusion center without communication delay for @xmath7 .
suppose that the initial state @xmath36 belongs to a given bounding ellipsoid : @xmath73 where @xmath74 is the center of ellipsoid @xmath75 , and @xmath76 is the shape matrix of the ellipsoid @xmath75 which is a known symmetric positive - definite matrix . at time
@xmath4 , given that @xmath9 belongs to a current bounding ellipsoid : @xmath77 where @xmath78 is the center of ellipsoid @xmath79 , and @xmath80 is a known symmetric positive - definite matrix . at next time @xmath45 , the fusion center can receive the state bounding ellipsoid of the @xmath6th sensor @xmath81 firstly , in prediction step , the goal of the fusion center is to determine a state bounding ellipsoid @xmath82 , i.e. , look for @xmath83 and @xmath84 such that the state @xmath51 belongs to @xmath85 whenever i ) @xmath9 is in @xmath79 , ii ) the process noise @xmath53 , and iii ) the remainder @xmath54 . secondly , in the fusion update step
, we look for @xmath86 and @xmath87 such that the state @xmath51 belongs to @xmath88 whenever i ) @xmath51 is in @xmath82 , ii ) @xmath51 is in @xmath72 , @xmath20 .
moreover , we provide a state bounding ellipsoid by minimizing its size " in prediction and update step , respectively .
in this section , we discuss the centralized set - membership estimation fusion , which includes the prediction step and the fusion update step . by taking full advantage of the character of the nonlinear dynamic system and the recent optimization method proposed in @xcite for linear dynamic system
, the centralized set - membership estimation fusion can be achieved by solving an sdp problem , which can be efficiently computed by interior point methods @xcite and related softwares @xcite .
furthermore , the centralized set - membership information filter is derived based on the decoupling technique , which can make further to improve the computation complexity of sdp .
the analytical formulae of the state prediction and estimation bounding ellipsoid at time @xmath45 are proposed , respectively . in the prediction step ,
the state prediction bounding ellipsoid at time @xmath45 can be derived as follows .
[ thm_1 ] at time @xmath45 , based on the state bounding ellipsoid @xmath89 , the remainder bounding ellipsoid @xmath32 and the noise bounding ellipsoid @xmath18 , the state prediction bounding ellipsoid @xmath90 can be obtained by solving the optimization problem in the variables @xmath50 , @xmath49 , nonnegative scalars @xmath91 , @xmath92 \label{eqpre_19 } & & ~~\mbox{subject to}~~ -\tau^u\leq0,~ -\tau^w\leq0,~ -\tau^f\leq0,\\[5 mm ] \label{eqpre_20 } & & { { \matrixfont p}}_{k+1|k}^c\succ0,\\[5 mm ] \label{eqpre_21}&&\left[\begin{array}{cc } { { \matrixfont p}}_{k+1|k}^c&\phi_{k+1|k}(\hat{{{\vectorfont x}}}_{k+1|k}^c)\\[3 mm ] ( \phi_{k+1|k}(\hat{{{\vectorfont x}}}_{k+1|k}^c))^t & ~~\xi\\ \end{array}\right]\succeq0,\end{aligned}\ ] ] where @xmath93,\\[3 mm ] \label{eqpre_23}\xi & = & \diag(1-\tau^u-\tau^w-\tau^f,\tau^u{{\matrixfont i}},\tau^w{{\matrixfont q}}_k^{-1},\tau^f{{\matrixfont i}}),\end{aligned}\ ] ] @xmath94 is the cholesky factorization of @xmath95 , i.e , @xmath96 , @xmath30 and @xmath97 are denoted by ( [ eqpre_6 ] ) , and @xmath24 is jacobian matrix . * proof : * see appendix .
the objective function ( [ eqpre_18 ] ) is aimed at minimizing the shape matrix of the predicted ellipsoid , and the constraints ( [ eqpre_19])-([eqpre_21 ] ) ensure that the true state is contained in the the bounding ellipsoid @xmath98 .
interestingly , if the objective function is the trace of the shape matrix of the bounding ellipsoid , then the _ analytically optimal solution _ of the optimization problem ( [ eqpre_18])-([eqpre_21 ] ) can be achieved for the sate prediction step .
[ cor_1 ] if the objective function @xmath99 , then the analytically optimal solution for the state prediction is as follows : @xmath100 where @xmath101 @xmath24 is the jacobian matrix of the nonlinear state function @xmath21 denoted by ( [ eqpre_3 ] ) , @xmath30 and @xmath34 are the center and shape matrix of the bounding ellipsoid of the remainder denoted by ( [ eqpre_6 ] ) , respectively , and @xmath102 , @xmath103 , @xmath104 are the optimal solution of the decision variables @xmath105 , @xmath106 , @xmath107 , respectively . *
proof : * see appendix . when the state equation is linear , there is no the remainder constraint of the nonlinear state equation , i.e. , @xmath108 , it is easy to observe that the optimum ellipsoid derived by theorem [ cor_1 ] coincides with the classical schweppe bounding ellipsoid @xcite . in the fusion update step ,
the state bounding ellipsoid at time @xmath45 can be derived as follows .
* proof : * see appendix .
moreover , in order to reduce computation complexity , we can derive an explicit expression of @xmath123 .
in lemma [ thm_2 ] , note that a suitable form of the orthogonal complement of @xmath124 can be chosen as follows @xmath125,\end{aligned}\ ] ] where @xmath126^t,\\ \label{eqpre_118}\psi_{22}&=&\left [ \begin{array}{ccccc } ( { { \matrixfont e}}_{k+1|k}^c)^{-1 } & 0 & 0 & \cdots & 0 \\ -{{\matrixfont j}}_{h_{k+1|k}^1 } & { { \matrixfont i } } & 0 & \vdots & 0 \\ -{{\matrixfont j}}_{h_{k+1|k}^2 } & 0 & { { \matrixfont i } } & \vdots & 0 \\ \vdots & \vdots & \vdots & \vdots & \vdots \\ -{{\matrixfont j}}_{h_{k+1|k}^l } & 0 & 0 & \vdots & { { \matrixfont i}}\\ 0 & -{{\matrixfont b}}_{h_{k+1}^1}^{-1 } & 0 & \vdots & 0 \\ 0 & 0 & -{{\matrixfont b}}_{h_{k+1}^2}^{-1 } & \vdots & 0 \\ \vdots & \vdots & \vdots & \vdots & \vdots \\ 0 & 0 & 0 & \cdots & -{{\matrixfont b}}_{h_{k+1}^l}^{-1 } \\ \end{array } \right].\end{aligned}\ ] ] if we denote @xmath127 then equation ( [ eqpre_52 ] ) is equivalent to the following form by reordering of the blocks @xmath128\succeq0,\\ & & \label{eqpre_121}{{\matrixfont b}}=[{{\matrixfont i}}~\underbrace{0,\ldots , 0}_{l~ blocks}],\end{aligned}\ ] ] where @xmath129 and 0 have compatible dimensions .
moreover , the decoupled fusion update step is given in the following theorem .
* proof : * see appendix . here , we call the equations ( [ eqpre_62])([eqpre_180 ] ) _ centralized set - membership information filter _ , which has following characters : * similar to the information filter @xcite , @xmath147 and @xmath148 in ( [ eqpre_62])([eqpre_180 ] ) can be taken as the update information matrix and the gain matrix provided by the @xmath6-th sensor for the estimator , respectively .
@xmath149 are the fusion weights .
* @xmath150 is the nonlinear correction term of the state update estimation , which relies on the nonlinear measurement functions @xmath142 , @xmath20 .
* when the measurement equations are linear , there are no the remainder constraints , i.e. , @xmath151 , it is easy to observe that the optimum ellipsoid derived by the theorem [ cor_2 ] also similar to the classical schweppe bounding ellipsoid @xcite . if @xmath63 and @xmath124 is full - rank , then the optimization problem ( [ eqpre_49])-([eqpre_52 ] ) in lemma [ thm_2 ] is an sdp problem , the dimension of the constraint matrix ( [ eqpre_52 ] ) is @xmath152 and the number of decision variables is @xmath153 , where @xmath154 and @xmath1 are the dimensions of the state , the measurement and the number of sensors , respectively . moreover , if we use a general - purpose primal - dual interior - point algorithm to solve it , then the computation complexity of the problem is @xmath155 , see @xcite .
therefore , in our case , the computation complexity is @xmath156 if @xmath157 , otherwise , it is @xmath158 . as described in @xcite
, we can use a path - following interior - point method to solve ( [ eqpre_57])-([eqpre_59 ] ) in theorem [ cor_2 ] .
a tedious but straightforward computation shows the practical complexity can be assumed to be @xmath159 , which implies an @xmath160 dependence on the size of the state @xmath161 , and @xmath162 dependence on the number of the sensor .
therefore , for the @xmath163 number of sensors , the complexity of the decoupled problem ( [ eqpre_57 ] ) improves upon that of the coupled one ( [ eqpre_49 ] ) by a factor of @xmath160 .
the centralized set membership information fusion algorithm can be summarized as follows .
[ alg_1 ] + * step 1 : ( initialization step ) set @xmath164 and initial values @xmath165 such that @xmath166 .
* step 2 : ( bounding step ) take samples @xmath167 from the sphere @xmath168 , and then determine two bounding ellipsoids to cover the remainders @xmath169 by ( [ eqpre_5])-([eqpre_6 ] ) . *
step 3 : ( prediction step @xcite ) optimize the center and shape matrix of the state prediction ellipsoid @xmath170 such that @xmath171 by ( [ eqpre_18])-([eqpre_21 ] ) or ( [ eqpre_24])-([eqpre_25 ] ) .
* step 4 : ( bounding step @xcite ) take samples @xmath172 from the sphere @xmath173 , and then determine one bounding ellipsoid to cover the remainder @xmath174 , @xmath20 , by ( [ eqpre_7])-([eqpre_8 ] ) . *
step 5 : ( fusion update step ) optimize the center and shape matrix of the state estimation ellipsoid @xmath175 such that @xmath176 by solving the optimization problem ( [ eqpre_49])-([eqpre_52 ] ) or ( [ eqpre_57])-([eqpre_59 ] ) .
* step 6 : set @xmath177 and go to step 2 .
in this section , in order to reduce the computation burden of the fusion center and improve the reliability , robustness , and survivability of the fusion system @xcite , the distributed set - membership estimation fusion method is derived by fusing the state bounding ellipsoids , which are sent from the local sensors and using the character of the nonlinear state function .
since the state prediction step of the distributed fusion is completely same as that of the centralized fusion , we only discuss the fusion update step of the distributed fusion .
in addition , the distributed set - membership information fusion formula can also be achieved by the decoupling technique .
the main results are summarized to lemma [ thm_3 ] and theorem [ cor_3 ] .
the proofs are also given in appendix .
[ thm_3 ] at time @xmath45 , based on the prediction bounding ellipsoids @xmath82 and the estimation bounding ellipsoids of single sensors @xmath72 , @xmath20 , the distributed state bounding ellipsoid @xmath178 can be obtained by solving the optimization problem in the variables @xmath87 , @xmath86 , nonnegative scalars @xmath179 , @xmath20 , @xmath180 \label{eqpre_82}&&~~\mbox{subject to}~~ -\tau^u\leq0,~-\tau_i^y\leq0\\[5 mm ] \label{eqpre_83 } & & -{{\matrixfont p}}_{k+1}^d\prec0,\\[5 mm ] \label{eqpre_84}&&\left[\begin{array}{cc } -{{\matrixfont p}}_{k+1}^d&\phi_{k+1}^d\\[3 mm ] ( \phi_{k+1}^d)^t & ~~-\xi-\pi\\ \end{array}\right]\preceq0,\end{aligned}\ ] ] where @xmath181,\\[3 mm ] \label{eqpre_86 } \phi_{k+1}^i&=&[\hat{{{\vectorfont x}}}_{k+1|k}^d-\hat{{{\vectorfont x}}}_{k+1}^i,{{\matrixfont e}}_{k+1|k}^d ] , \\[3 mm ] \label{eqpre_87}\pi&= & \sum_{i=1}^l\tau_i^y(\phi_{k+1}^i)^t({{\matrixfont p}}_{k+1}^i)^{-1}\phi_{k+1}^i,\\ \label{eqpre_88}\xi & = & \diag(1-\tau^u-\sum_{i=1}^l\tau_i^y,\tau^ui),\end{aligned}\ ] ] @xmath182 is the cholesky factorization of @xmath87 , i.e , @xmath183 .
* proof : * see appendix . compared with the centralized fusion in lemma [ thm_2 ]
, it can be seen that the dimension of the constraint matrix ( [ eqpre_84 ] ) is @xmath184 independent of the number of the sensors and the number of decision variables is @xmath185 .
however , the dimension of the constraint matrix ( [ eqpre_52 ] ) is @xmath186 , and the number of the decision variables is @xmath153 .
therefore , the distributed fusion can decrease much more computation burden of the fusion center .
note that ( [ eqpre_84 ] ) can be rewritten to @xmath187 \label{eqpre_0106 } ( \hat{{{\vectorfont x}}}_{k+1|k}^d-\hat{{{\vectorfont x}}}_{k+1}^d)^t&\upsilon_{11 } & \upsilon_{12}\\[3 mm ] ( { { \matrixfont e}}_{k+1|k}^d)^t&\upsilon_{12}^t & \upsilon_{22}\\ \end{array}\right]\succeq0,\end{aligned}\ ] ] where @xmath188 moreover , we can derive an analytical formula for the shape matrix and the center of the bounding ellipsoid @xmath82 as follows .
the proof is similar to theorem [ cor_2 ] .
we call the equations ( [ eqpre_110])([eqpre_111 ] ) _ distributed set - membership information filter_. in ( [ eqpre_110])([eqpre_111 ] ) , @xmath197 and @xmath198 can be taken as the update information matrix and the gain matrix provided by the @xmath6-th sensor for the estimator , respectively , and @xmath199 are the fusion weights .
the distributed set membership information fusion algorithm can be summarized as follows .
[ alg_2 ] + * step 1 : ( initialization step ) set @xmath164 and initial values @xmath165 such that @xmath166 .
* step 2 : ( bounding step @xcite ) take samples @xmath167 from the sphere @xmath168 , and then determine a bounding ellipsoid to cover the remainders @xmath169 by ( [ eqpre_5])-([eqpre_6 ] ) . *
step 3 : ( prediction step ) optimize the center and shape matrix of the state prediction ellipsoid @xmath200 @xmath201 such that @xmath202 by solving the optimization problem ( [ eqpre_18])-([eqpre_21 ] ) or ( [ eqpre_24])-([eqpre_25 ] ) .
* step 4 : ( fusion update step ) optimize the center and shape matrix of the state estimation ellipsoid @xmath203 such that @xmath204 by solving the optimization problem ( [ eqpre_81])-([eqpre_84 ] ) or ( [ eqpre_107])-([eqpre_109 ] ) based on the state prediction bounding ellipsoids @xmath82 and bounding ellipsoids of single sensors @xmath72 , @xmath20 .
* step 5 : set @xmath177 and go to step 2 . in target tracking ,
whether it is the distributed fusion or the centralized fusion , if the measurement only contain range and angle , the boundary sampling method @xcite can be used to drive the bounding ellipsoid of the remainders with less computation complexity .
therefore , the bounding steps of algorithm [ alg_1 ] and algorithm [ alg_2 ] can be computed efficiently .
finally , the set - membership information fusion formulae are summarized in table [ tab_1 ] .
.set - membership information fusion formulae [ cols="^,^,^",options="header " , ] [ tab_1 ] as far as multi - algorithm fusion for nonlinear dynamic systems is concerned , the multiple bounding ellipsoids can be constructed to minimize the size of the state bounding ellipsoid by complementary advantages of multiple parallel algorithms .
specifically , one can use multiple parallel algorithm [ alg_1 ] or [ alg_2 ] with differently weighted objectives in ( [ eqpre_170 ] ) , where the larger @xmath205 emphasizes the @xmath206th entry of the estimated state vector , then the intersection of these bounding ellipsoids can achieve a tighter bounding ellipsoid that containing the true state in fusion center .
in this section , we provide an example to compare the performance of the centralized fusion with that of the distributed fusion .
moreover , we also use the multi - algorithm fusion to further reduce the estimation error bound based on the different weighted objective ( [ eqpre_170 ] ) . consider a common tracking system with bounding noise and there are two sensors track a same target in different position .
the state contain position and velocity of @xmath207 and @xmath208 directions . here
, the dynamic system equations is as follows @xcite : @xmath209{{\vectorfont x}}_k+{{\vectorfont w}}_k,\\[3 mm ] \label{eqpre_115}{{\vectorfont y}}_k^i&=&\left [ \begin{array}{c } \sqrt{({{\vectorfont x}}_k(1)-{{\vectorfont z}}_k^i(1))^2+({{\vectorfont x}}_k(2)-{{\vectorfont z}}_k^i(2))^2 } \\[3 mm ] arctan\left(\frac{{{\vectorfont x}}_k(2)-{{\vectorfont z}}_k^i(2)}{{{\vectorfont x}}_k(1)-{{\vectorfont z}}_k^i(1)}\right ) \\
\end{array } \right]+{{\vectorfont v}}_k^i,\\ \nonumber & & ~for~ i=1,2.\end{aligned}\ ] ] where @xmath210 is the time sampling interval with @xmath211 .
@xmath212^t$ ] is the position of the @xmath6th sensor , where @xmath213^t$ ] and @xmath214^t$ ] .
moreover , the process noise @xmath13 and measurement noise @xmath215 are taking value in specified ellipsoidal sets @xmath216 where @xmath217\\ \nonumber { { \matrixfont r}}_k^i&=&\left [ \begin{array}{cc } 0.01 & 0\\ 0 & 25\\ \end{array } \right].\end{aligned}\ ] ] the target acceleration is @xmath218 . in the example , the target starts at the point @xmath219 with a velocity of @xmath220 .
the center and the shape matrix of the initial bounding ellipsoid are @xmath221^t$ ] , @xmath222,\end{aligned}\ ] ] respectively . in order to simulate the performance of the center fusion and distributed fusion
, we assume the process noise measurement noise are truncated gaussian with zeros mean and covariance @xmath223 and @xmath224 on the ellipsoidal sets , respectively . from the description of the above
, we can use sensor 1 ( smf1 ) , sensor 2 ( smf2 ) , the centralized fusion ( csmf ) and distributed fusion ( dsmf ) to calculate the error bound with @xmath225 $ ] in ( [ eqpre_170 ] ) , respectively , moreover , we also use the multi - algorithm fusion ( msmf ) to produce the error bound based on the different weight coefficient with @xmath226 $ ] , @xmath227 $ ] , @xmath228 $ ] , @xmath229 $ ] , where the * error bound * of the @xmath6th entry of the state @xmath51 can be calculated by projecting the ellipsoid along the @xmath6th output direction .
the following simulation results are under matlab r2012a with yalmip
. figs .
[ fig_01]-[fig_04 ] present a comparison of the error bounds along position and velocity direction for sensors 1 , 2 using algorithm [ alg_1 ] ( l=1 ) and for the fusion center using the centralized fusion algorithm [ alg_1 ] ( l=2 ) and the distributed fusion algorithm [ alg_2 ] ( l=2 ) and the multi - algorithm fusion , respectively . from figs .
[ fig_01]-[fig_04 ] , we can observe the following phenomenon : * the performance of the centralized fusion and the distributed fusion is better than that of sensors . *
the performance of the centralized fusion is better than that of the distributed fusion along @xmath207 and @xmath208 position direction in figs .
[ fig_01]-[fig_02 ] , but the distributed fusion performs slightly better than centralized fusion along @xmath207 and @xmath208 velocity direction in figs .
[ fig_03]-[fig_04 ] .
the reasons may be that the optimal bounding ellipsoid can not be obtained for the nonlinear dynamic system , and the error bound of the state vector is calculated by minimizing trace of the shape matrix of the bounding state ellipsoid rather than minimizing the error bounds along position and velocity directions , respectively . *
the performance of the multi - algorithm fusion is significantly better than that of the other methods along position and velocity direction . since it extract the useful information of each entry of the state vector by the differently weighted objectives .
then the intersection fusion of these estimation ellipsoids can sufficiently take advantage of the information of each sensor , which yields a tighter state bounding ellipsoidal . to 6cmto to 6cmto to 6cmto to 6cmto
this paper has derived the centralized and distributed set - membership information fusion algorithms for multisensor nonlinear dynamic system via minimizing state bounding ellipsoid .
firstly , both of them can be converted into an sdp problem which can be efficiently computed , respectively .
secondly , their analytical solutions can be derived surprisingly by using decoupling technique .
it is very interesting that they are quite similar in form to the classic information filter in mse sense . in the two analytical fusion formulae
, the information of each sensor can be clearly characterized , and the knowledge of the correlation among measurement noises across sensors are not required .
finally , multi - algorithm fusion has been used to minimize the size of the state bounding ellipsoid by complementary advantages of multiple parallel algorithms . a typical example in target tracking
has showed that multi - algorithm fusion performs better than both the centralized and distributed fusion .
future work will include , in multisensor nonlinear dynamic system setting , multiple target tracking , sensor management and heterogeneous sensor fusion .
[ lem_1]@xcite let @xmath230 , be quadratic functions in variable @xmath231 @xmath232 with @xmath233 .
then the implication @xmath234 holds if there exist @xmath235 such that @xmath236 [ lem_2 ] schur complements @xcite : given constant matrices @xmath237 , @xmath136 , @xmath238 , where @xmath239 and @xmath240 , then @xmath241 if and only if @xmath242\preceq0\end{aligned}\ ] ] or equivalently @xmath243\preceq0\end{aligned}\ ] ] [ lem_3 ] decoupling @xcite : let @xmath244 be matrices of appropriate size , with @xmath245 square and symmetric . the problem ( in variable @xmath246 ) @xmath247\succeq 0\end{aligned}\ ] ] is feasible if and only if @xmath248\succeq 0.\end{aligned}\ ] ] in this case , problem ( [ eqpre_30])is equivalent to the problem ( in variable @xmath249 only ) @xmath250\succeq 0.\end{aligned}\ ] ] moreover , if the problem ( [ eqpre_32 ] ) is feasible , which means that @xmath251 suppose the objective function is either the trace function or log - det function , then @xmath252 whenever @xmath253 .
thus , ( [ eqpre_30 ] ) admits a unique optimal variable given by @xmath254 , where @xmath255 is the pseudo - inverse of @xmath256 . :
note that @xmath257 is equivalent to @xmath258 , @xmath259 , where @xmath94 is a cholesky factorization of @xmath95 . by the nonlinear state equations
( [ eqpre_1 ] ) and ( [ eqpre_3 ] ) , @xmath260 if we denote by @xmath261^t,\end{aligned}\ ] ] then ( [ eqpre_33 ] ) can be rewritten as @xmath262 where @xmath263 is denoted by ( [ eqpre_22 ] ) . moreover , the condition that @xmath264 , whenever , i ) @xmath265 , ii ) the process noise @xmath53 , iii ) the high - order remainders of state function @xmath266 , which are equivalent to @xmath267 whenever @xmath268 the equations ( [ eqpre_37])([eqpre_39 ] ) are equivalent to @xmath269 where @xmath129 and @xmath270 are matrices with compatible dimensions . from lemma [ lem_1 ] , a sufficient condition such that the inequalities ( [ eqpre_40])-([eqpre_42 ] ) imply ( [ eqpre_36 ] ) to hold
is that there exist nonnegative scalars @xmath271 , such that @xmath272 furthermore , ( [ eqpre_43 ] ) is written in the following compact form : @xmath273 where @xmath274 is denoted by ( [ eqpre_23 ] ) . applying lemma [ lem_2 ] , ( [ eqpre_44 ] ) is equivalent to @xmath275\succeq0\\ \label{eqpre_46}&&{{\matrixfont p}}_{k+1|k}^c\succ0.\end{aligned}\ ] ] therefore , if @xmath276 , @xmath50 satisfy ( [ eqpre_45 ] ) , then the state @xmath277 belongs to @xmath278 , whenever , i ) @xmath279 is in @xmath43 , ii ) the process noise @xmath53 , iii ) the high - order remainders of state function @xmath266 . : if we partition the left side of ( [ eqpre_21 ] ) by appropriate block , then it can be rewritten as @xmath281\succeq 0,\end{aligned}\ ] ] where @xmath282,\\ \nonumber{{\matrixfont x}}_{11}&=&1-\tau^u-\tau^w-\tau^f,\\ \nonumber{{\matrixfont x}}_{22}&=&\diag(\tau^u{{\matrixfont i}},\tau^w{{\matrixfont q}}_k^{-1},\tau^f{{\matrixfont i}}),\\ \nonumber{{\matrixfont x}}_{12}&=&0.\end{aligned}\ ] ] based on the decoupling technique in lemma [ lem_3 ] , the above matrix inequality is feasible if and only if @xmath283\succeq 0.\end{aligned}\ ] ] from the expression of @xmath284 , it is also equivalent to @xmath285 thus , the optimization problem of lemma [ thm_1 ] @xmath286 which , by lemma [ lem_3 ] , is equivalent to @xmath287 it is easy to see that @xmath256 is nonsingular according to ( [ eqpre_48 ] ) , then , the above optimization problem is equivalent to @xmath288 where @xmath289 .
therefore , based on lagrange dual function , the analytically optimal solution can be obtained in ( [ eqpre_24])-([eqpre_28 ] ) . :
note that we have get @xmath290 in prediction step , which is equivalent to @xmath291 , @xmath292 , where @xmath117 is a cholesky factorization of @xmath50 , then , @xmath293 and by the nonlinear measurement equations ( [ eqpre_2 ] ) and ( [ eqpre_4 ] ) @xmath294 if we denote by @xmath295^t,\end{aligned}\ ] ] then ( [ eqpre_63 ] ) and ( [ eqpre_64 ] ) can be rewritten as @xmath296 where @xmath297 and @xmath298 are denoted by ( [ eqpre_53 ] ) and ( [ eqpre_54 ] ) , respectively .
moreover , the condition that @xmath299 whenever i ) @xmath51 is in @xmath47 ii ) measurement noises @xmath300 are bounded in ellipsoidal sets , i.e. , @xmath59 , iii ) the high - order remainders of measurement function @xmath301 , , @xmath20 , which are equivalent to @xmath302 whenever @xmath303 the equations ( [ eqpre_69])([eqpre_71 ] ) are equivalent to @xmath304 where @xmath129 and @xmath270 are matrices with compatible dimensions . by @xmath0-procedure lemma [ lem_1 ] and ( [ eqpre_67 ] ) ,
a sufficient condition such that the inequalities ( [ eqpre_72])-([eqpre_74 ] ) imply ( [ eqpre_68 ] ) to hold is that there exist scalars @xmath305 and nonnegative scalars @xmath306 , such that @xmath307 furthermore , ( [ eqpre_75 ] ) is written in the following compact form : @xmath308 where @xmath274 and @xmath124 are denoted by ( [ eqpre_56 ] ) and ( [ eqpre_55 ] ) , respectively .
if we denote @xmath309 is the orthogonal complement of @xmath124 , then ( [ eqpre_76 ] ) is equivalent to @xmath310 using schur complements lemma [ lem_2 ] , ( [ eqpre_77 ] ) is equivalent to @xmath311\preceq0.\\ \label{eqpre_79}&&-{{\matrixfont p}}_{k+1}^c\prec0.\end{aligned}\ ] ] therefore , if @xmath312 , @xmath56 satisfy ( [ eqpre_78])-([eqpre_79 ] ) , then the state @xmath277 belongs to @xmath123 , whenever i ) @xmath51 is in @xmath47 ii ) measurement noises @xmath300 are bounded in ellipsoidal sets , i.e. , @xmath59 , iii ) the high - order remainders of measurement function @xmath301 , , @xmath20 . : in view of the optimization problem in lemma [ thm_2 ] , we can apply lemma [ lem_3 ] to the linear matrix inequalities ( [ eqpre_80 ] ) , with @xmath314 , and the rest of matrices defined appropriately .
thus , the problem @xmath315 which is equivalent to @xmath316 where @xmath317 , @xmath20 . if one of @xmath318 , @xmath20 , is zero , then the feasible sets of @xmath56 and @xmath55 become smaller from ( [ eqpre_76 ] ) , and the objective value becomes larger .
thus , the optimal @xmath318 , @xmath20 should be greater than zero , and @xmath319 be nonsingular .
if @xmath320 is the optimal value of the above optimization problem , then , by using lemma [ lem_3 ] again , the optimal ellipsoid @xmath137 is given by @xmath321 based on ( [ eqpre_201 ] ) and @xmath314 , we retrieve the center of the ellipsoid as @xmath322 by the definition of @xmath133 and @xmath323 in ( [ eqpre_118 ] ) and ( [ eqpre_120 ] ) , @xmath324 \ ] ] then @xmath325(\psi_{22}^t\xi_{22}\psi_{22})^{-1}[{{\matrixfont i}}~\underbrace{0,\ldots , 0}_{l~ blocks}]^t\\ \nonumber & = & \left(\tau^u{{\matrixfont p}}_{k+1|k}^{c^{-1}}+\sum_{i=1}^l{{{\matrixfont j}}_{h_{k+1|k}^i}^t}(\frac{{{\matrixfont r}}_{k+1}^{i}}{\tau_i^v}+\frac{{{\matrixfont p}}_{h_{k+1}^i}}{\tau_i^h})^{-1 } { { \matrixfont j}}_{h_{k+1|k}^i}\right)^{-1}.\end{aligned}\ ] ] thus , ( [ eqpre_62 ] ) can be obtained by ( [ eqpre_200 ] ) . moreover , substituting ( [ eqpre_122 ] ) , ( [ eqpre_118 ] ) and ( [ eqpre_120 ] ) into ( [ eqpre_202 ] ) , then ( [ eqpre_180 ] ) can be achieved . :
note that @xmath326 is equivalent to @xmath327 , @xmath328 , where @xmath182 is a cholesky factorization of @xmath87 , then @xmath329 if we denote by @xmath330^t,\end{aligned}\ ] ] then ( [ eqpre_89 ] ) can be rewritten as @xmath331 where @xmath332 is denoted by ( [ eqpre_85 ] ) .
similarly , we have @xmath333 where @xmath334 is denoted by ( [ eqpre_86 ] ) .
moreover , the condition that @xmath335 , whenever , i ) @xmath51 is in @xmath82 , ii)@xmath336 , for @xmath20 , is equivalent to @xmath337 whenever , for @xmath20 , @xmath338 the equations ( [ eqpre_94])([eqpre_97 ] ) are equivalent to @xmath339\xi&\leq & 0,\end{aligned}\ ] ] where @xmath340 and @xmath270 are matrices with compatible dimensions . by @xmath0-procedure lemma [ lem_1 ] , a sufficient condition such that the inequalities ( [ eqpre_98])-([eqpre_101 ] ) imply ( [ eqpre_93 ] ) to hold
is that there exist nonnegative scalars @xmath341 , @xmath20 , such that @xmath342 \preceq0\end{aligned}\ ] ] furthermore , ( [ eqpre_102 ] ) is written in the following compact form : @xmath343 where @xmath274 and @xmath344 are denoted by ( [ eqpre_87 ] ) and ( [ eqpre_88 ] ) , respectively .
therefore , if @xmath346 , @xmath87 satisfy ( [ eqpre_104])-([eqpre_105 ] ) , then the state @xmath277 belongs to @xmath347 , whenever , i ) @xmath51 is in @xmath82 , ii ) @xmath277 belongs to @xmath348 , for @xmath20 .
m. liggins , c. y. chong , i. kadar , m. g. alford , v. vannicola , and s. thomopoulos , `` distributed fusion architectures and algorithms for target tracking , '' _ proceeding of ieee _ , vol .
85 , pp . 95107 , january 1997 .
a. vempaty , y. s. han , and p. k. varshney , `` target localization in wireless sensor networks using error correcting codes , '' _ ieee transaction on information theory _ ,
vol . 60 , pp . 697712 , january 2014 .
x. shen , y. zhu , e. song , and y. luo , `` minimizing euclidian state estimation error for linear uncertain dynamic systems based on multisensor and multi - algorithm fusion , '' _ ieee transactions on information theroy _ , vol .
57 , pp .
71317146 , october 2011 . | the set - membership information fusion problem is investigated for general multisensor nonlinear dynamic systems . compared with linear dynamic systems and point estimation fusion in mean squared error sense , it is a more challenging nonconvex optimization problem .
usually , to solve this problem , people try to find an efficient or heuristic fusion algorithm .
it is no doubt that an analytical fusion formula should be much significant for rasing accuracy and reducing computational burden .
however , since it is a more complicated than the convex quadratic optimization problem for linear point estimation fusion , it is not easy to get the analytical fusion formula . in order to overcome the difficulty of this problem , two popular fusion architectures are considered : centralized and distributed set - membership information fusion .
firstly , both of them can be converted into a semidefinite programming problem which can be efficiently computed , respectively .
secondly , their analytical solutions can be derived surprisingly by using decoupling technique .
it is very interesting that they are quite similar in form to the classic information filter . in the two analytical fusion formulae
, the information of each sensor can be clearly characterized , and the knowledge of the correlation among measurement noises across sensors are not required .
finally , multi - algorithm fusion is used to minimize the size of the state bounding ellipsoid by complementary advantages of multiple parallel algorithms .
a typical numerical example in target tracking demonstrates the effectiveness of the centralized , distributed , and multi - algorithm set - membership fusion algorithms .
in particular , it shows that multi - algorithm fusion performs better than the centralized and distributed fusion .
* keywords : * nonlinear dynamic systems , multisensor fusion , target tracking , unknown but bounded noise , set - membership filter |
quantum control is aimed at designing external pulses in order to achieve efficient transfers between the states of the quantum system under study @xcite .
this task is crucial in atomic and molecular physics , and has many applications extending from photochemistry to quantum computation .
quantum control has attracted attention among the physics and chemistry communities @xcite , but also in applied mathematics for the development of new theoretical methods . in strong field molecular physics , antagonistic basic mechanisms like bond softening versus vibrational trapping @xcite , or barrier lowering versus dynamical dissociation quenching @xcite have been referred to in the control scenarios of molecular reactivity or even for alignment / orientation purpose @xcite .
an even more unexpected quenching mechanism is in relation with the so - called zero - width resonances ( zwr ) . for certain couples of laser parameters ( wavelength and intensity ) the photodissociation rate vanishes leading to , in principle , infinitely long lived resonances ( zwrs ) in both continuous wave @xcite and pulsed regimes @xcite . in a diatomic molecule ,
a destructive interference is in play taking place among fluxes contributing to the outgoing scattering amplitude of an initial vibrational state decaying through two laser - induced adiabatic channels .
when dealing with intramolecular couplings , the specific requirement for such destructive interference can only be fulfilled for some particular interchannel coupling amplitudes . despite strong interchannel coupling
, the observation of some narrow rotational lines in ibr predissociation provides such an example @xcite .
radiative interactions offer the possibility of a continuous tuning of laser parameters to reach a zwr .
our laser control objective is the vibrational cooling of diatomic molecules produced by photoassociation in translationally and rotationally cold , tightly bound states @xcite .
the filtration strategy we are referring to aims at protecting against photodissociation a given single vibrational state , while all others are decaying fast .
this is achieved through an adiabatic and continuous transport of this state population on its corresponding zwr all along a laser pulse adequately shaped and frequency chirped . starting from a given vibrational distribution ,
the efficiency of such a control remains on two crucial issues : ( i ) the best possible protection of the given vibrational state ; ( ii ) the highest decay rates for all the others .
the first issue deals with the adiabatic transport model .
recently , we built an intuitive transport scheme by chirping a laser pulse such as to merely combine , at each time within the pulse duration , the wavelength and the intensity corresponding to the following of a zwr originating , in field - free conditions , from the vibrational state we wish to protect .
although this simple scheme does not result from a formal adiabatic treatment , we obtained encouraging results in the case of a light diatomic like h@xmath1 , with 70@xmath2 of the initial population left at the end of the pulse thanks to the relatively slowly varying pulse envelope and modest chirp amplitude @xcite .
it turns out that the situation is drastically different for na@xmath0 where almost the whole initial population can photodissociate .
this prompted us to go through a full adiabatic transport control scheme based on the adiabatic floquet theory @xcite completely reformulated to account for zwrs in a pulsed regime .
the second issue concerns the anharmonicity and relatively high density of vibrational levels in heavier species like na@xmath0 .
the consequence is that laser pulses shaped so as to protect a given vibrational level can have characteristics close to the ones appropriate for at least partly protecting neighboring levels .
a semiclassical analysis helps in finding and interpreting , in the laser parameter plane , a complete map of zwrs originating from a field - free vibrational level .
we show how a choice can be done of a particular zwr path , among the complete series identified by the semiclassical analysis .
this aims at realizing the most efficient filtration , namely the best compromise for selectively protecting the single vibrational state in consideration .
the paper is organized in the following way : in section [ sec2 ] , zwrs are introduced within a two - state photodissociation model referring both to a time - independent close - coupled floquet hamiltonian formalism and a semiclassical interpretation .
two computational methods , either grid or global are presented .
section [ sec3 ] is devoted to a complete derivation of the adiabatic control theory .
the filtration strategy as applied to na@xmath0 is discussed in section [ sec4 ] , by introducing the model , the zwr maps , the adiabatic transport dynamics and the choice of the optimal zwr path .
in this section , we examine the rather unexpected property of a resonance state , originating from a bound state coupled to a dissociative continuum , to possibly acquire an infinite ( or quasi - infinite ) lifetime .
such exotic resonances have already been discussed in the literature , first in the context of accidentally narrow rotational lines in predissociation @xcite , then as bound sates in continuum ( bic ) @xcite , and more recently as zero - width resonances ( zwr ) @xcite . in the following , the physical context is molecular photodissociation and our illustrative example is na@xmath0 .
we briefly recall the model used to describe the photodissociation of a rotationless field - aligned diatomic molecule in a single spatial dimension ( the internuclear distance @xmath3 ) with only two electronic states labeled @xmath4 and @xmath5 in a born - oppenheimer approximation . reduced dimensionality with the frozen rotation assumption is validated by considering linearly polarized light and short pulse durations we refer to hereafter , as compared to the rotational periods of the molecular species .
the time - dependent wave function is written as : @xmath6 the nuclear dynamics is governed by the time dependent schrdinger equation ( tdse ) : @xmath7 = \left ( t_n + \left[\begin{array}{c c } v_1(r)&0 \\ 0&v_2(r ) \end{array}\right ] \right . & \nonumber \\ & - \left .
\mu_{12}(r)\mathcal{e}(t)\left[\begin{array}{c c } 0&1 \\ 1&0 \end{array}\right ] \right ) \left[\begin{array}{c } \phi_1 ( r , t ) \\ \phi_2 ( r , t ) \end{array}\right ] & \label{tdse}\end{aligned}\ ] ] @xmath8 represents the nuclear kinetic energy operator . @xmath9 and @xmath10
are the born - oppenheimer potentials .
@xmath11 is the electronic transition dipole moment between states @xmath4 and @xmath5 .
@xmath12 is the linearly polarized electric field amplitude .
we first consider the case of a continuous wave ( cw ) laser , @xmath13 with an intensity ( @xmath14 ) , a frequency @xmath15 and a wavelength @xmath16 , @xmath17 being the speed of light . in this strictly periodic case ,
the floquet ansatz being applied : @xmath18 = e^{-ie_v t/\hbar } \left [ \begin{array}{c } \chi_1 ( r , t ) \\ \chi_2 ( r , t ) \end{array}\right ] .
\label{eq : anzats}\end{aligned}\ ] ] due to the periodicity in time of @xmath19 these functions can be fourier expanded : @xmath20 where the fourier components satisfy a set of coupled differential equations , for any @xmath21 : @xmath22\varphi_{1,2}^n(r ) \nonumber \\ & -1/2 e \mu_{12}(r ) \left[\varphi_{2,1}^{n-1}(r ) + \varphi_{2,1}^{n+1}(r)\right]=0 \label{eq : closecoupledmulti}\end{aligned}\ ] ] for moderate field intensities inducing single - photon processes , these equations obviously simplify into two close - coupled equations : @xmath23\varphi_{1,v}(r)-1/2 e \mu_{12}(r ) \varphi_{2,v}(r)=0
\nonumber\ ] ] @xmath24\varphi_{2,v}(r)-1/2 e \mu_{12}(r ) \varphi_{1,v}(r)=0
\label{eq : closecoupled}\ ] ] where we have kept only the fourier component @xmath25 of @xmath26 and the zero - frequency fourier component @xmath27 of @xmath28 , denoted @xmath29 and @xmath30 , respectively .
it has to be noted that a specific solution ( @xmath31 ) has been identified in eq.([eq : closecoupled ] ) , by labeling both the eigenenergy @xmath32 and the corresponding eigenvector @xmath33 and their fourier components @xmath34 .
resonances are solutions with siegert type outgoing - wave boundary conditions @xcite and have complex quasi - energies of the form @xmath35 , where @xmath36 is the resonance width related to the decay rate . in the following ,
the label @xmath31 designates both the field - free vibrational level and the laser - induced resonance originating from this vibrational state . now relaxing the cw laser assumption
, we consider a chirped laser pulse with parameters @xmath37 involving slowly varying envelope and frequency .
the molecule , initially in a particular field - free vibrational state @xmath31 , is supposed to be adiabatically driven by such a pulse .
adiabaticity means here that a unique resonance @xmath38 , labeled @xmath31 according to its field - free parent bound state @xmath31 , is followed during the whole dynamics .
this resonance wave function involves , through its complete basis set expansion , a combination of both bound and continuum eigenstates of the field - free molecular hamiltonian .
but the important issue is that , at the end of the pulse , the molecule is again on its initial single vibrational state @xmath31 ( adiabaticity condition ) . for such open systems ,
contrary to dynamics involving bound states only , there is unavoidably an irreversible decay process , precisely due to the fact that vibrational continuum states are temporarily populated under the effect of the pulse .
a quantitative measure of such a decay could be given in terms of the overall fraction of non - dissociated molecules , assuming a perfect adiabatic following of the selected resonance @xcite : @xmath39 \;\;\ ] ] where the decay rate @xmath40 is associated with the relevant floquet resonance quasi - energy @xmath41 using the instantaneous field parameters @xmath42 at time @xmath43 .
there are two central control issues : ( 1 ) investigating how rates are changing with the field parameters and in particular find optimal combinations @xmath44 for which these rates are small enough ( or even ideally zero ) to insure the survival of the vibrational state @xmath31 to the laser excitation , that is @xmath45 , @xmath46 being the total pulse duration ; ( 2 ) predicting how well the adiabatic following is effectively realized ( this point will be analysed in more details in section [ sec3 ] ) . the occurrence of zwrs , at least under some particular circumstances , is rationalized by the semiclassical theory of predissociation of a diatomic molecule @xcite .
the formalism deals with the so - called adiabatic potentials @xmath47 resulting from the diagonalization of the molecule - field interaction matrix and predicts dissociation quenching resulting from a null value of the outgoing scattering amplitude in the lower ( open ) adiabatic channel @xmath48 , if the following two conditions are simultaneously fulfilled@xcite : @xmath49 and @xmath50 in eqs.([eq : semiclassical1],[eq : semiclassical2 ] ) the wavenumbers are @xmath51^{1/2}$ ] , @xmath52 is the reduced nuclear mass , @xmath53 are the left turning points of @xmath54 potentials , @xmath55 is the right turning point of @xmath56 and @xmath57 is the diabatic crossing point resulting from field - dressing . with integer @xmath58 and
@xmath59 these conditions lead to bohr - sommerfeld quantization involving a coincidence between two energies ; namely one @xmath60 of the upper adiabatic potential @xmath61 , with a phase correction @xmath62 , which in weak coupling is @xmath63 @xcite , and another @xmath64 of a potential made of two branches , namely , @xmath65 for @xmath66 , and @xmath61 otherwise . for a weak coupling ,
this is practically the diabatic attractive potential @xmath9 .
more precisely , the coincidence condition is implemented in the expression of the resonance width @xmath36 @xcite : @xmath67 ^ 3 } ( \varepsilon_{\tilde{v}}-\varepsilon_{\tilde{v}_{+}})^2 \label{eq : gammasem}\ ] ] @xmath68 and @xmath69 are the local energy spacings of the modified diabatic and adiabatic potentials respectively .
@xmath70 is the landau - zener coupling parameter : @xmath71 where @xmath72 and @xmath73 are the classical velocity and slope difference of the diabatic potentials at @xmath57 .
clearly , the two energies @xmath64 and @xmath60 , and therefore the width @xmath36 are dependent on field parameters , i.e. , both frequency ( or wavelength ) and amplitude ( or intensity ) .
this is in particular due to the ( @xmath74 , @xmath75)-dependence of the corresponding field - dressed adiabatic potentials @xmath47 .
all other factors building up @xmath36 in eq.([eq : gammasem ] ) , that is the coupling @xmath70 and the local energy spacings @xmath68 , @xmath69 , also depend on field parameters .
contrary to predissociation where such coincidences can only be accidental since there is no easy and continuous way to modify potentials and interstate couplings ( electronic or spin - orbit ) , for a diatomic molecule submitted to an electromagnetic field , a fine tuning of the wavelength and intensity produces them at will .
this explains the occurrence of zwrs in photodissociation .
moreover , for a wavelength @xmath74 which roughly brings into coincidence the levels @xmath59 ( corresponding to the field - free vibrational level @xmath31 in consideration ) and @xmath76 , a fine tuning of the intensity @xmath75 will result in an accurate determination of a zwr , that is @xmath77 . in some cases , a stronger field ( higher @xmath75 )
may also bring into coincidence @xmath59 with @xmath78 , producing thus a second zwr , for the same wavelength , and so on for @xmath79 but , one can also envisage slightly different wavelengths which build energetically close enough @xmath59 and @xmath76 levels in a field - dressed picture , such that a subsequent fine tuning of the intensity brings them into precise coincidence .
this flexibility offered by the field parameters that , in principle , can be continuously modified , is at the origin of not only quasi - zero width photodissociation resonances , but also for their multiple occurrence in the ( @xmath74 , @xmath75)-parameter plane @xcite .
we are referring to two classes of computational methods for an accurate determination of laser parameters @xmath44 producing a zwr : ( i ) grid methods and ( ii ) global methods .
to be more specific , for these calculations , @xmath43 has not to be considered as a time variable , but rather as a parameter .
to a fixed @xmath43 corresponds a set of field parameters @xmath37 or equivalently @xmath80 , which determine a specific cw laser , fulfilling the single period requirement of the floquet ansatz of eq.([eq : anzats ] ) . the time - independent system of close - coupled equations for the @xmath3-dependent fourier coefficients @xmath81 of eqs .
( [ eq : closecoupled ] ) are solved for these given field amplitude and frequency .
these methods are based on so - called shooting - matching techniques .
the fox - goodwin propagation algorithm @xcite is used in conjunction with a properly chosen set of imposed boundary conditions , which accounts for the correct behavior of the channel wave functions both at short and large distances .
the algorithm consists in an iterating sequence of two steps @xcite : ( i ) shooting : with an initial guess for the quasi - energy @xmath32 , properly initialized fox - goodwin matrices are constructed at each point of a grid ( along @xmath3 ) in terms of independent solution matrices propagated inward and outward ; ( ii ) matching : the criterion for convergence is a condition to be fulfilled for matching both the functions and their derivatives on two adjacent points of the grid , by properly changing the energy @xmath32 . while regularity ( @xmath82 ) is imposed at the origin for both open and closed channels , which are classically forbidden , the boundary conditions for large @xmath3 are different for these two types of channels .
zero inward initialization is still valid for the closed channel @xmath83 , whereas siegert - type outgoing boundary conditions should be adopted for the open one @xmath84 @xcite .
high accuracy can be achieved by using complex rotation of the coordinate @xmath3 @xcite which actually plays the role of an absorbing potential .
the most important observation is that the use of the complex coordinate brings the outgoing asymptotic behavior to regularity ( zero inward initialization ) @xcite .
such imposed boundary conditions are at the origin of quantization conditions leading to discrete complex resonance energies . in practice ,
a plausible laser wavelength is guessed from the semiclassical coincidence criterion involving the vibrational level @xmath59 of the quasi - diabatic potential @xmath9 at low intensity eq.([eq : semiclassical2 ] ) and one of the vibrational levels @xmath85 of the upper adiabatic potential @xmath61 eq.([eq : semiclassical1 ] ) .
this is convenient for the search of a zwr originating from @xmath59 , which actually is nothing but @xmath31 at the zero field intensity limit .
an important point to consider is that , due to the possibility to bring into coincidence several vibrational levels of the upper adiabatic potential @xmath85 with a given @xmath59 , there are several wavelengths which may be convenient for a zwr search .
such a wavelength being fixed for a given field - free vibrational level @xmath31 , the field strength @xmath75 is progressively increased . for each discrete value of @xmath75 ,
starting from an initial guess for the energy coming from the previously considered value of @xmath75 , the propagation - matching procedure leads to an accurate eigenenergy @xmath86 when properly converged .
all calculations are conducted within the single photon absorption approximation frame of eqs.([eq : closecoupled ] ) , which is valid for low enough intensities , neglecting the mixing of neighboring floquet blocks occurring within the period @xmath87 @xcite .
a typical behavior of the resonance width @xmath36 as a function of the increasing field strength @xmath75 displays a linear increase , and a saturation , followed by a natural decrease of the decay rates which is interpreted in terms of spatial non - adiabatic effects @xcite .
but more interesting is the observation of very fast and local collapse of the width , for specific intensities .
one can finely adjust the intensity , such as to obtain resonance widths close to zero within several figures of accuracy .
this shows that , at least in the single floquet block approximation of eqs.([eq : closecoupled ] ) , zwrs do exist as also predicted by the semiclassical theory .
introducing multiphoton processes will offer additional dissociation channels from which the outgoing flux could be evacuated .
this may give rise to ( small but still non zero ) partial widths contributing to @xmath36 @xcite which may no more strictly collapse to zero .
inserting the floquet ansatz eq . into the schrdinger equation
leads to an eigenvalue problem for the floquet states @xmath88 and the associated quasienergies @xmath32 , @xmath89 \left [ \begin{array}{c } \chi_{v,1 } ( r , t ) \\ \chi_{v,2 } ( r , t ) \end{array } \right ] = e_v \left [ \begin{array}{c } \chi_{v,1 } ( r , t ) \\ \chi_{v,2 } ( r , t ) \end{array } \right ] , \label{floquetev1}\ ] ] where @xmath90 is the hamiltonian introduced in eq . .
this eigenvalue problem can be solved using a global iterative method .
the operator @xmath91 appearing in the left - hand side of eq .
is called the floquet hamiltonian and can be represented within a direct product basis set made of two fourier basis sets , one for the radial coordinate and another for the time coordinate ( as an additional quantum coordinate ) .
we typically use 4 to 8 floquet blocks with both open channels ( describing multi photon absorption ) and closed ones ( describing photon emission ) . with the moderate intensities used hereafter , the floquet block describing a single photon absorption remains the dominant one .
it is to be noticed that grid method calculations can also be extended to perform calculations including a larger number of floquet blocks and the convergence with respect to the number of channels has been checked independently with both numerical approaches . in any case , if the intensity is small enough , the difference between two - channel calculations and calculations using a larger number of floquet blocks remains very small . using the global representation of the floquet hamiltonian described above , we proceed as follows : for each value of the field parameters @xmath92 , one particular solution of eq .
( [ floquetev1 ] ) is obtained by using an initial guess for the eigenvector and modifying it by using an iterative method .
we have taken advantage of the wave operator method which can handle dissipative problems and is well - adapted for searching small numbers of complex eigenvalues of a large matrix @xcite .
the main idea is to define an active space of interest chosen so as to possess a strong overlap with the expected unknown eigenvector .
then we find the wave operator @xmath93 which transforms the floquet hamiltonian matrix @xmath94 into a smaller , effective matrix within the active space , @xmath95 p_o , \label{waveop}\ ] ] the eigenvalues of which are exact eigenvalues of the initial problem .
@xmath96 is the projector on this active space and @xmath97 is a non - invertible matrix with a special structure , such that all columns corresponding to states out of the active space be zero .
the wave operator can be found by effectively solving the non - linear bloch equation @xcite , @xmath98 the above eq . , together with the special structure imposed to @xmath97 ,
completely define the wave operator able to reduce the eigenvalue problem to an effective , small - dimensional problem .
eq . can be solved in an iterative way .
we have used a variant of the recursive distorded wave approximation algorithm @xcite . here , the active space can be chosen as one - dimensional ( the space spanned by the vibrational state from which the resonance is formed ) .
this choice is simple and sufficient to attain convergence .
this means that we calculate one resonance at a time .
the same approach has been previously used in a slightly different context within the constrained adiabatic trajectory method ( catm ) and more methodological details about the iterative equations can be found in ref .
it should be noted that the wave - operator equation ( [ waveop ] ) can give either solutions to the tdse ( if suitable time - dependent absorbing potentials are included in the model to constrain the initial conditions , as in ref .
@xcite ) , or quasienergies and the corresponding floquet eigenstates defined in eq . ( [ floquetev1 ] ) ( if there are no constraints on the initial conditions , as is the case here ) . in comparison with grid methods described in section [ gridmethods
] , there is no need for an initial energy guess but we do need an initial guess for the eigenvector . at low intensities ,
the algorithm converges well by using a field - free vibrational state , constant over the time period , as an initial guess .
then , when the laser parameters are progressively varied it is possible to use the result obtained in the previous calculation ( with slightly different parameters ) as the initial guess for calculating the next eigenvector corresponding to the next laser parameters .
each new calculation costs only a few more iterations to converge toward the new quasienergy ( only one iteration is often sufficient ) .
this is an advantage of using a global iterative method , the exploration of the parameter plane is made easier .
another difference with grid methods is that the calculations with the global algorithm have been done using a radial complex absorbing potential to discretize the continuum , instead of using a complex rotation of the coordinate .
this different implementation has very little consequence on the results , since the absorbing potential is acting over a large domain .
the optimal zwr paths in the @xmath92 parameter plane are obtained following a two - step strategy . in a first step
, we fix a low intensity and we study exhaustively the variations of the width @xmath36 with respect to the wavelength , for different resonances originating from different values of @xmath31 .
this allows us to identify the wavelengths of interest
. then complete zwr paths are obtained by gradually varying the laser intensity @xmath75 and finding for each intensity the optimal @xmath74 corresponding to a minimum resonance width @xmath36 .
this gives a critical set of parameters @xmath99 which forms an almost continuous numerical path ( see section [ zwrmaps ] ) .
the purpose of optimizing laser parameters @xmath100 such that the survival probability of a resonance state originating in field - free conditions from a given vibrational state @xmath31 be maximized , while all other resonances ( originating from @xmath101 ) are decaying fast , can be conducted within the frame of the adiabatic floquet formalism @xcite .
the methodology goes through the following steps : ( a ) introduction of the adiabatic floquet hamiltonian acting on an extended hilbert space together with the conditions of its equivalence with the original tdse of the physical space ; ( b ) adiabatic approximation aiming at the tracking of a single resonance in the extended space ; ( c ) zero - width resonance ( zwr ) strategy in the extended space ; ( d ) back transformation to the physical space .
we are hereafter developing these steps . starting from the tdse : @xmath102 with @xmath103 .
\nonumber\ ] ] the first goal to achieve towards adiabaticity is to fix the rapidly growing phase @xmath104 leading to non - adiabatic fast oscillations of the control field .
this is precisely done through the floquet hamiltonian @xmath105 acting on an extended hilbert space given as a tensorial product @xmath106 of the physical hilbert space @xmath107 times the space @xmath108 of square integrable functions of @xmath109 defined as an additional phase , varying within the interval @xmath110 $ ] and describing a field degree of freedom .
more precisely , this adiabatic floquet hamiltonian involves a so - called effective frequency @xcite : @xmath111 and reads : @xmath112 the resulting time evolution equation is : @xmath113 we will now examine the relationship between the two wavepackets @xmath114 and @xmath115 , solutions of eq.([eq : tdse ] ) and eq.([eq : floquetevolution ] ) , respectively . for doing this ,
@xmath115 is expanded on the complete basis set of square integrable eigenfunctions @xmath116 of @xmath117 with the straightforward possibility to extend to energy - normalized continua .
more precisely , @xmath118 designates the complete set of eigenvectors associated with the square integrable vibrational eigenfunctions , augmented by the enumerable set of the continuum eigenfunctions sub - space .
the remaining part of the expansion concerns the basis set @xmath119 of @xmath120 , with time - dependent unknown coefficients @xmath121 : @xmath122 recasting eq.([eq : psi ] ) into eq.([eq : floquetevolution ] ) results in : @xmath123 the notation @xmath124 is for the time - derivative of @xmath17 .
it is important to notice that eq.([eq : cvn ] ) holds for any time @xmath43 and for any phase @xmath109 , taken up to here as two independent variables .
more specifically , from now on @xmath109 is taken as : @xmath125 this choice being done for @xmath109 , eq.([eq : cvn ] ) reads : @xmath126 or equivalently : @xmath127 which shows that if @xmath115 is a solution of eq.([eq : floquetevolution ] ) , then @xmath128 is in turn a solution of eq.([eq : tdse ] ) , with @xmath109 defined by eq.([eq : theta ] ) .
the fast oscillating behavior of @xmath129 is now recast in @xmath130 , the fourier components @xmath121 of @xmath131 being henceforth slowly varying functions of time . as a conclusion , the two control strategies aiming at a vibrationally selective survival through an adiabatic tracking of a zwr , namely : ( i ) finding the laser parameters @xmath37 such that @xmath114 , solution of eq.([eq : tdse ] ) follows this zwr , or , ( ii ) finding the laser parameters @xmath132 such that @xmath115 , solution of eq.([eq : floquetevolution ] ) follows the zwr , are formally identical , with the choice of eq.([eq : theta ] ) for @xmath109 . in the following
it is the adiabatic strategy ( ii ) which is retained .
we are now looking for an adiabatic evolution such that at all times @xmath43 , the solution @xmath115 follows a specific single resonance eigenvector @xmath133 of the adiabatic floquet hamiltonian labeled by its corresponding field - free parent state @xmath134 : @xmath135 this equation should be understood in the following way : at each time @xmath43 , the field parameters @xmath132 give rise to a hamiltonian @xmath136 with resonance eigenfunctions @xmath137 and eigenvalues @xmath138 .
the adiabatic approximation for @xmath115 is nothing but : @xmath139~ |\chi_v\{\theta , \epsilon(t)\}\rangle .
\label{eq : psiad}\ ] ] we note , for completeness , that the exact wavefunction @xmath115 could be given as an expansion , on the basis set of the resonance eigenvectors @xmath140 or their analogs @xmath141 including the corresponding phase @xmath142 $ ] : @xmath143 the initial condition describing the system in its vibrational state @xmath31 , implies : @xmath144 with the consequence : @xmath145 which actually is the initial field - free vibrational wavefunction .
the full control strategy is built in two steps : ( i ) trapping the system into a single eigenvector @xmath133 of the adiabatic floquet hamiltonian ; and ( ii ) designing a pulse with field parameters such that this eigenstate presents the lowest ( zero , if possible ) dissociation rate .
the requirement of step ( i ) leading to the approximation of eq.([eq : psiad ] ) , is expected to be reached for slowly enough ( adiabatically ) varying parameters @xmath146 avoiding any degeneracy between complex eigenvalues @xmath147 , at all times @xmath43 @xcite .
the second step ( ii ) involves an optimal choice for the field parameters : @xmath148 such that : @xmath149=0 ~~ \forall t \label{eq : imag}\ ] ] @xmath150 being the imaginary part of the energies of these field - induced resonances .
we note that eq.([eq : imag ] ) stands precisely for the search of a zwr path ( originating from @xmath134 ) in the amplitude / frequency parameter plane ( or equivalently , intensity @xmath75 / wavelength @xmath74 ) as a function of @xmath43 . in other words , the star * in eq.([eq : imag ] ) could be replaced by the superscript zwr , such that : @xmath151^{1/2 } , \\
\omega_{\text{eff}}^{\text{zwr}}(t ) = 2\pi c/\lambda^{\text{zwr}}(t ) \nonumber \end{array } \right\}\ ] ] be the control field of the evolution monitored by the adiabatic floquet hamiltonian of the extended hilbert space eq.([eq : floquetevolution ] ) .
following such a resonance associated with a quasi - infinite lifetime is very favorable for obtaining a good adiabatic approximation in this dissipative context .
following the less dissipative eigenstate is actually a requirement of the adiabatic theorem in presence of dissipation @xcite . in any other situation , there is a risk of loosing population in the selected resonance and to magnify non - adiabatic transitions due to this relative population loss with respect to other eigenstates . following
a zwr automatically avoids this magnification of non - adiabatic contaminations , which could occur if the adiabatic approximation is done on a resonance associated with a non - zero decay rate . on mathematical grounds , it is also to be noticed that fulfilling the requirement of eq.([eq : imag ] ) leads to the possibility of solving the challenging issue of adiabatic transport involving passages through continuum spectra .
in other words , zwrs are good candidates for a full adiabatic floquet treatment as initially derived for pure bound states .
the purpose is now to obtain the optimal field parameters acting in the original hilbert space where the evolution is monitored by the tdse displayed in eq.([eq : tdse ] ) .
this is achieved by solving the first order differential equation eq.([eq : omegaeff ] ) , with @xmath152 as the unknown function : @xmath153 here again the star * stands for the optimal control solution . a particular solution of eq.([eq : domega ] )
is obtained as : @xmath154 which satisfies the expected initial condition , i.e. : @xmath155 finally , the external laser field which fulfills the requirements of the control objective , in the physical hilbert space is : @xmath156 } \nonumber \\ & = & e^{\text{zwr}}(t ) \cos \left({\int_0^t \omega_{\text{eff}}^{\text{zwr}}(t ' ) dt ' } \right ) \label{adiab_pulse1}\end{aligned}\ ] ] which can be written in terms of intensity / wavelength parameters as : @xmath157^{1/2 } \cdot \cos \left({\int_0^t 2\pi c/\lambda^{\text{zwr}}(t ' ) dt ' } \right ) \label{adiab_pulse}\ ] ] all the dynamical results presented below have been obtained by rigorously solving the tdse .
we have used two independent methods to validate the numerical results .
we have first used the split - operator scheme @xcite which is based on a differential propagation over small time steps .
split - operator calculations have been done either with a grid fourier dvr basis set associated with the internuclear distance coordinate or within the bare vibrational eigenbasis , the splitting being done between kinetic and potential operators or between the diagonal and dipolar coupling terms , respectively .
we have also confirmed all the results by using a recently developed iterative algorithm based on the time - dependent wave operator formalism in which fourier decomposition is also used to describe the time coordinate @xcite , allowing a much smaller number of discretization points . in this framework ,
the time - dependent wave operator @xmath158 is calculated and used to deduce the true dynamics from the dynamics within a smaller dimensional subspace which includes only a small part of the bare vibrational ( bound and possibly continuum ) states . the evolution operator issuing from the small - dimensional subspace is calculated as @xmath159 where @xmath160 is a time - dependent effective hamiltonian governing the dynamics within the active subspace of projector @xmath96 . this effective hamiltonian is similar to the one of eq.([waveop ] ) but is now defined over the entire duration of the chirped pulse , while eq.([waveop ] ) is valid over a single optical period , at fixed frequency and intensity .
we shall point out that this concept of effective hamiltonian is not only a computational intermediate , it also naturally gives an indication about the adiabatic or non - adiabatic character of the dynamics during the interaction .
if the active subspace is one - dimensional ( the space spanned by the vibrational state @xmath161 associated with the selected zwr @xmath162 ) , then the effective hamiltonian leads to an effective energy @xcite , @xmath163 where @xmath164 is a solution of the tdse ( [ eq : tdse ] ) .
this quantity follows some trajectory in the complex plane which illustrate the more or less adiabatic character of the dynamics .
assuming a perfect adiabatic dynamics and using eq .
( [ eq : psiad ] ) , the effective energy can also be expressed as the sum of two terms @xcite @xmath165 where @xmath166 and @xmath167 are the selected instantaneous floquet state and eigenvalue , corresponding to the field parameters @xmath146 ( cf .
( [ floquetev2 ] ) ) . following a zwr path
, the first term @xmath168 will approximately follow a straight line along the real axis during the pulse , starting from the initial field - free vibrational energy . the second term in eq .
( [ eeff2 ] ) depends on @xmath166 which is time - periodic with a chirped period , @xmath169 .
this term may show rapid oscillations in the complex plane . for a given zwr path ,
two different trajectories can be calculated : @xmath170 , using the exact wavefunction ( direct use of eq .
( [ eeff1 ] ) with the solution of eq .
( [ eq : tdse ] ) including all the possible non - adiabatic transfers ) , and @xmath171 , expected from the adiabatic approximation using eq .
( [ eeff2 ] ) . the quality of the matching between these two complex trajectories will be used in section [ adiatransp ] as an indicator for the quality of the adiabatic following .
this section is devoted to the potentiality of zwrs implemented in an adiabatic control scenario to reach a filtration strategy for molecular vibrational cooling with na@xmath0 as an illustrative example .
the following application deals with a realistic situation since the formation of ultracold sodium molecules has already been demonstrated @xcite .
the transposability to other diatomic molecules , especially to alkali dimers , requires only small changes due to the similarities of their potentials .
transitionally and rotationally cold , tightly bound and vibrationally hot na@xmath0 species are experimentally produced by photoassociation in a metastable bound state @xmath172 , considered as a ground state ( referred to as state u ) .
typically , vibrational levels with quantum numbers @xmath173 are prepared .
the laser controlled filtration strategy consists in applying an electromagnetic field with wavelengths around 570@xmath174 which couples state u with a repulsive , thus dissociating excited ( 1)@xmath175 electronic state ( referred to as state g ) . the corresponding born - oppenheimer potential energy curves @xmath176 and the electronic transition dipole moment @xmath11 between states g and u are taken from the literature @xcite and references therein .
our specific model refers to a rotationless field - aligned molecule in a single spatial dimension ( namely , the internuclear distance @xmath3 ) .
such a frozen rotation approximation is validated when comparing the short pulse durations in consideration ( 12@xmath177 ) with the long rotational periods of na@xmath0 ( estimated as hundreds of @xmath177 ) .
finally , na@xmath0 reduced mass which is involved in the kinetic energy operator @xmath8 of eq.(2 ) is taken as 20963.2195@xmath178 . solving time - independent
coupled equations eq.([eq : closecoupled ] ) with siegert boundary conditions for a set of continuous wave cw laser parameters @xmath179 or equivalently @xmath92 , gives rise to resonances with complex eigenvalues @xmath32 correlating , in field - free conditions , with the real vibrational eigenenergies @xmath32 .
we are actually interested in finding specific couples of field parameters for which the imaginary part of resonance eigenvalues are close to zero .
the general morphology of zwrs is illustrated in fig.([morphology_zwr ] ) which displays an exploratory calculation in the @xmath92 parameter plane .
the left panel shows the imaginary part of @xmath180 ( in logarithmic scale ) as a function of laser intensity and wavelength covering a 5@xmath174 wide window in the vicinity of @xmath181 .
the right panel displays the imaginary parts of @xmath182 as a function of the laser wavelength for a fixed intensity @xmath183 w.@xmath184 . within numerical inaccuracies inherent to the evaluation of such very small width resonances ( typically less than @xmath185 )
, we observe that there are several couples of critical wavelengths and intensities producing zwrs originating from a single vibrational level @xmath31 .
referring to the semiclassical analysis , the quantum destructive interference pattern from phase ( or energy ) coincidences in eq.([eq : gammasem ] ) leads to the choice of photon frequencies picking out the couple of electronic states which take part in the process by appropriately dressing the corresponding adiabatic potentials .
the field intensity is the remaining parameter which allows an accurate phase coincidence , aiming at reaching zwrs . due to the relatively moderate intensities we are referring to , a single photon model with only two electronic states offers enough accuracy for converged calculations , obtained using the grid and global methods presented in section [ computmethods ]
clearly , for a given field - free vibrational state , the destructive interference leading to a zwr can not take place , whatever the intensity be , if an avoided curve crossing position exceeds the right turning point of the vibrational level in consideration .
in particular , for @xmath186 this fixes a maximum value for @xmath74 at about 577@xmath174 .
guided by the semiclassical analysis , a tentative wavelength of @xmath181 brings roughly into coincidence the @xmath186 level with the @xmath187 level of the upper adiabatic potential .
the fine adjustment is obtained referring to the full quantum floquet approach by tuning the remaining field parameter , i.e. , the intensity @xmath75 , such as to produce the zwr within accuracies of typically @xmath188 concerning the widths . in the laser parameter plane @xmath189 .
the labels ( @xmath186 , @xmath190=0,1,2,3 ) refer to energy coincidences between @xmath186 on one hand , and on the other hand , the levels @xmath190 of the upper adiabatic potential of the semiclassical picture . ] as has previously been observed @xcite , within an intensity window not exceeding @xmath191 , zwrs are very close to being positioned along a straight line in the @xmath189 parameter plane . figure ( [ loci_zwr ] )
displays such a behavior which seems generic for all zwr paths .
these are labeled as ( @xmath192 ) referring to the field - free initial vibrational state @xmath186 and the successive @xmath193 levels of the upper field - dressed adiabatic potential of the semiclassical picture brought into coincidence using different wavelengths .
it is worthwhile noting that the semiclassical picture is merely referred to for an initial guess of laser parameters and also for a systematic labeling of zwrs paths .
figure ( [ loci_zwr ] ) illustrates the general morphology of zwr in the laser parameter plane .
it could obviously be extended to additional paths based on coincidences involving @xmath186 with @xmath190=4,5 , ... we are now looking for the dynamical signature of a zwr on a vibrational population transfer using the adiabatic transport strategy of section iii .
we proceed by shaping a laser pulse envelope optimized for such an adiabaticity with a total duration not exceeding 12@xmath177 to fulfill the frozen rotation requirement .
the remaining part of the strategy consists in chirping the pulse such as to combine at each time the wavelength and intensity corresponding to the following of a zwr path as obtained from the floquet theory and illustrated in figure ( [ loci_zwr ] ) . for convenience
a linear fit is used for representing these paths : @xmath194 the coefficients @xmath195 and @xmath196 are separately calculated for each of the different paths ( @xmath192 ) . taking ( @xmath197 ) as an illustrative example we get @xmath198 and @xmath199 .
the additional parameters involved in the pulse envelope could be obtained by optimal control algorithms aiming at the minimization of the population decay due to non - adiabatic contamination in the transport .
we simplified the procedure and based on our previous experience , after a few attempts , we have used linear ramps of intensity , @xmath200 \\ i_{\text{max } } \left ( 2 - \frac{t}{t_{1/2 } } \right ) & \text { for } t \in [ t_{1/2 } , t_{\text{tot } } ] \end{array } \right .
\label{eq : enveloppe}\ ] ] where @xmath201 is half the total pulse duration .
the wavelenght is continuously chirped according to eq .
( [ eq : linear ] ) to follow the zwr path . as a function of time .
( a ) for the adiabatic transport based on a laser pulse resulting from eqs.([eq : enveloppe ] ) and ( [ adiab_pulse ] ) , following the zwr path ( @xmath186 , @xmath187 ) of fig .
( [ loci_zwr ] ) : @xmath186 , solid black line , @xmath202 solid green , grey and red lines respectively ( b ) for the instantaneous zwr frequency strategy of eq.([eq : ins_pulse ] ) . only the populations of @xmath186 are displayed for three zwr paths ( @xmath186 , @xmath187 ) thick solid line , ( @xmath186 , @xmath203 ) dashed line and ( @xmath186 , @xmath204 ) thin solid line , following the notations of fig .
( [ loci_zwr ] ) . ] as a function of time .
( a ) for the adiabatic transport based on a laser pulse resulting from eqs.([eq : enveloppe ] ) and ( [ adiab_pulse ] ) , following the zwr path ( @xmath186 , @xmath187 ) of fig .
( [ loci_zwr ] ) : @xmath186 , solid black line , @xmath202 solid green , grey and red lines respectively ( b ) for the instantaneous zwr frequency strategy of eq.([eq : ins_pulse ] ) .
only the populations of @xmath186 are displayed for three zwr paths ( @xmath186 , @xmath187 ) thick solid line , ( @xmath186 , @xmath203 ) dashed line and ( @xmath186 , @xmath204 ) thin solid line , following the notations of fig .
( [ loci_zwr ] ) . ] figure [ adiab_ins ] ( upper panel ) illustrates our strategy with a total pulse duration of 12@xmath177 , long enough to produce almost complete depletion of neighboring level populations @xmath202 . during the transport dynamics
, the initial population of @xmath186 is partially shared with @xmath205 and 9 ( in decreasing importance ) . as expected from adiabaticity , at the end of the pulse ,
the major part of shared populations is recovered back by @xmath186 which ends up with a final population of 95@xmath2 .
the quality of adiabaticity which is achieved can be measured by less than 5@xmath2 of the @xmath2068 population escaped towards the dissociative continuum . at that respect , it is worthwhile comparing the efficiency of the adiabatic transport control we are presenting in this work eq.([adiab_pulse ] ) to the one much more naive , we have previously discussed about @xcite , where the pulse was shaped referring to an instantaneous zwr frequency strategy : @xmath207^{1/2 } \cdot \cos [ ( 2\pi c/\lambda^{\text{zwr}})t ]
\label{eq : ins_pulse}\ ] ] figure [ adiab_ins ] ( lower panel ) displays the results for the same @xmath2068 and the zwr path identified in eq.([eq : linear ] ) , but with the field of eq.([eq : ins_pulse ] ) with the envelope given by eq.([eq : enveloppe ] ) . the efficiency with respect to the protection against dissociation is completely washed out , with only 5@xmath2 of the population remaining undissociated .
similar calculations done with the other zwr paths , taken from fig.([loci_zwr ] ) , lead to intermediate efficiency , namely , 50@xmath2 for (
@xmath186 , @xmath203 ) and 75@xmath2 for ( @xmath186 , @xmath204 ) , but still clearly illustrating that the instantaneous strategy is far being an adiabatic transport , as has been previously proposed @xcite .
[ fig_eeff ] is another way of checking the adiabaticity of the dynamical process . following the discussion of subsection [ nummethod ]
we have drawn the trajectories of the effective energies given by eqs .
( [ eeff1 ] ) and ( [ eeff2 ] ) . the top frame of fig .
[ fig_eeff ] shows short extracts of the complex trajectory @xmath208 given in eq .
( [ eeff2 ] ) , deduced from the ideal adiabatic approximation of eq .
( [ eq : psiad ] ) . in this case , only one floquet eigenvector is present .
each ellipse represents the complex trajectory followed by @xmath209 during one optical cycle , at three intermediate times corresponding to given field parameters during the rise of the pulse , @xmath210 , @xmath211 and @xmath212 .
@xmath213 is the maximum field amplitude reached during the pulse .
the dotted line represents the trajectory followed by the first term of eq .
( [ eeff2 ] ) , i.e. the instantaneous floquet eigenvalue @xmath214 .
this is also the trajectory followed by the instantaneous centers of the ellipses resulting from the second term in eq .
( [ eeff2 ] ) .
four black circles also indicate the positions of @xmath214 at four different times , @xmath215 ( when there is no electric field ) and at the intermediate times @xmath216 , @xmath217 , @xmath218 defined above .
this dotted trajectory stays very close to the real axis because the control pulse is designed to follow a zwr path .
this frame should be compared with the middle and bottom frames of fig .
[ fig_eeff ] which contains extracts of the effective energy trajectory @xmath219 obtained by using eq .
( [ eeff1 ] ) with the exact solution of the tdse .
the ellipses have been recorded during short time intervals @xmath220 $ ] around the same intermediate times @xmath216 , @xmath217 , @xmath218 , with @xmath221 0.12 ps corresponding to a few tens of optical cycles ( this explains the broadening of the curves ) .
we have also shown the instantaneous center , i.e. the rolling mean over one optical cycle , @xmath222 , during the whole pulse .
if the adiabatic approximation is satisfied , this instantaneous center should be close of the eigenvalue trajectory ( dotted line in the top frame ) .
the more adiabatic is the dynamics , the closer from the top frame should be the trajectories .
the middle frame has been obtained by using the pulse of eq .
( [ adiab_pulse1 ] ) and ( [ eq : enveloppe ] ) , with the phase @xmath223 .
we observe a nice agreement with the ideal adiabatic trajectory shown in the top frame , with ellipsoids of growing radii whose centers closely follows the instantaneous resonance trajectory .
the deviations from the real axis are less than 0.2 @xmath224 . as a comparison ,
the bottom frame shows the trajectories obtained by using the naive pulse of eq .
( [ eq : ins_pulse ] ) with the phase @xmath225 .
we observe very different results : adiabaticity is rapidly lost , in agreement with the explanations developed in section [ sec3 ] .
the deviation from adiabaticity is even more marked during the falling - off of the control field , leading to an erratic behavior in the instantaneous center trajectory .
the chaotic path on the right side of panel ( c ) corresponds to the second half of the pulse and indicates that the adiabatic character is completely lost .
these results are consistent with the direct population analysis of fig .
[ adiab_ins ] .
equations ( [ eeff1 ] ) and ( [ eeff2 ] ) ) during the dynamics following the zwr path .
( a ) ideal trajectories expected from the adiabatic approximation following a single floquet eigenstate .
( b ) results from the tdse using the pulse of eq .
( [ adiab_pulse1 ] ) and ( [ eq : enveloppe ] ) with adiabatic phase @xmath226 .
( c ) results from the tdse using the pulse of eq .
( [ eq : ins_pulse ] ) with the naive phase @xmath227 .
the ellipses correspond to short extracts of the complex parametric curve @xmath228 around three different times corresponding to given intermediate values of the field amplitude during the rise of the pulse : @xmath229 , @xmath211 and @xmath212 . for panels ( b ) and
( c ) the instantaneous center is the complex trajectory obtained by performing a rolling mean of @xmath230 over one optical cycle ( see text ) .
note that the scales are slightly different in panel ( c ) .
, title="fig : " ] ( cf . equations ( [ eeff1 ] ) and ( [ eeff2 ] ) ) during the dynamics following the zwr path .
( a ) ideal trajectories expected from the adiabatic approximation following a single floquet eigenstate .
( b ) results from the tdse using the pulse of eq .
( [ adiab_pulse1 ] ) and ( [ eq : enveloppe ] ) with adiabatic phase @xmath226 .
( c ) results from the tdse using the pulse of eq .
( [ eq : ins_pulse ] ) with the naive phase @xmath227 .
the ellipses correspond to short extracts of the complex parametric curve @xmath228 around three different times corresponding to given intermediate values of the field amplitude during the rise of the pulse : @xmath229 , @xmath211 and @xmath212 . for panels ( b ) and
( c ) the instantaneous center is the complex trajectory obtained by performing a rolling mean of @xmath230 over one optical cycle ( see text ) .
note that the scales are slightly different in panel ( c ) .
, title="fig : " ] ( cf . equations ( [ eeff1 ] ) and ( [ eeff2 ] ) ) during the dynamics following the zwr path .
( a ) ideal trajectories expected from the adiabatic approximation following a single floquet eigenstate .
( b ) results from the tdse using the pulse of eq .
( [ adiab_pulse1 ] ) and ( [ eq : enveloppe ] ) with adiabatic phase @xmath226 .
( c ) results from the tdse using the pulse of eq .
( [ eq : ins_pulse ] ) with the naive phase @xmath227 .
the ellipses correspond to short extracts of the complex parametric curve @xmath228 around three different times corresponding to given intermediate values of the field amplitude during the rise of the pulse : @xmath229 , @xmath211 and @xmath212 . for panels ( b ) and
( c ) the instantaneous center is the complex trajectory obtained by performing a rolling mean of @xmath230 over one optical cycle ( see text ) .
note that the scales are slightly different in panel ( c ) .
, title="fig : " ] the results obtained using the two kind of pulses , corresponding to eq .
( [ eq : ins_pulse ] ) ( naive pulse ) or eq .
( [ adiab_pulse ] ) ( adiabatic strategy ) , are very different , only the adiabatic strategy giving rise to an efficient protection against dissociation of the selected state .
the two pulses are compared in fig .
[ fig : spectres ] .
a direct comparison is difficult due to the very large number of oscillations ( more than 6000 optical cycles ) , therefore we have shown the power spectrum of each pulse on the left panel , together with the relative frequency shifts occuring during the pulses ( right panel ) . the spectrum corresponding to the pulse of eq .
( [ adiab_pulse ] ) displays three major peaks close to the central frequency ( corresponding to lambda = 576 nm ) , whereas the one resulting from eq.([eq : ins_pulse ] ) covers a much larger frequency range . at the selected scale ,
the difference between the two pulses is clearly noticeable . on the right panel of fig .
[ fig : spectres ] , we have also drawn the relative frequency shifts @xmath231 ( for the effective frequency ) and @xmath232 ( using the actual instantaneous frequency ) .
it is worthwile noting that the shifts remain very small all along the pulses .
this is certainly an advantage for a future experimental implementation .
the various numerical results previously shown in fig . [ adiab_ins ] and fig . [ fig_eeff ] can be better understood by noting that a strong difference piles up during the pulse between the effective and the instantaneous frequencies , explaining why the dynamics progressively gets away from an optimal adiabatic dynamics if we use one frequency instead of the other . using the adiabatic strategy of eq .
( [ adiab_pulse ] ) for shaping the pulse is thus very important . ) , solid line , and the naive strategy of eq .
( [ eq : ins_pulse ] ) , dashed line .
( b ) frequency shift during the pulses , for the effective frequency @xmath231 ( solid line ) and the actual instantaneous frequency @xmath232 ( dashed line ) .
, title="fig : " ] ) , solid line , and the naive strategy of eq .
( [ eq : ins_pulse ] ) , dashed line .
( b ) frequency shift during the pulses , for the effective frequency @xmath231 ( solid line ) and the actual instantaneous frequency @xmath232 ( dashed line ) .
, title="fig : " ] as graphically shown in fig .
( [ loci_zwr ] ) , there are several zwr paths originating from @xmath2068 and labeled by the corresponding quasi - degenerate upper adiabatic potential levels @xmath190=0,1,2,3 of the semiclassical analysis .
the question which is addressed now is the choice of the path which better fulfills the requirement of the filtration strategy .
more precisely , all of these paths are well adapted to protect @xmath2068 population from decaying .
but , due to the anharmonicity and the density of the vibrational levels , a couple of laser parameters ( @xmath233 ) which are well identified as producing a zwr of @xmath2068 , can also be rather well suited in producing a zwr originating from a neighboring vibrational state @xmath2067 , 9 or 10 .
in such a situation , when following this zwr path , not only @xmath2068 population is protected against dissociation as it should be , but the dissociation of neighboring levels could also be severely slowed down . as a consequence the filtration strategy would loose its efficiency .
only very long duration pulses would achieve the complete decay of @xmath2067,9 and 10 , inducing the unwanted effect of rotational degrees of freedom coming into play producing rotational heating .
the best choice among zwr paths targets the highest contrast that could be expected between the widths of the resonances originating from @xmath2068 ( ideally zero ) on one hand , and from @xmath2067,9,10 ( largest possible values , that is shortest lifetimes ) on the other hand .
once again the semiclassical analysis reveals helpful for such a choice .
( in blue ) and @xmath234 ( in green ) for the different zwr paths ( @xmath186 , @xmath193 ) of fig . ( [ loci_zwr ] ) . for the definition of @xmath235 , see eq.([eq : d_def ] ) .
] as indicated in eq.([eq : gammasem ] ) , the semiclassical estimate of resonance widths is proportional to the square of the difference @xmath236 of the energies of the corresponding levels accommodated by the adiabatic potentials @xmath47 .
they are obtained by the numerical solution of eqs.([eq : semiclassical1],[eq : semiclassical2 ] ) including the phase correction @xmath62 . for field intensities close to zero
, @xmath237 is nothing but the energy @xmath238 , the energy of the ground state vibrational level @xmath31 , whereas @xmath239 is the semiclassical estimate @xmath240 of the upper adiabatic potential vibrational levels with a @xmath241 phase correction . restricting the analysis to @xmath2068,9 and 10 , the field free values of @xmath238 are calculated once for all , with the results : @xmath242 , @xmath243 , @xmath244 , the zero - energy being chosen as the energy of two infinitely separated na@xmath245 ions , as in @xcite .
the values of @xmath240 depend on the specific wavelength @xmath74 appropriately shifting the @xmath61 potential in a field - dressed picture . as an example , for @xmath246 corresponding to the almost zero intensity field value of the zwr ( @xmath186 , @xmath187 ) of fig .
( [ loci_zwr ] ) , one gets : @xmath247 , @xmath248 , @xmath249 . remaining in the almost zero - field regime , the different zwr paths displayed in fig .
( [ loci_zwr ] ) correspond to a quasi coincidence of @xmath250 with one of the @xmath240 ( @xmath193 ) depending on the laser wavelength . due to important anharmonicities and relatively dense vibrational distribution ( for @xmath2518 )
, it may happen that , for some wavelengths , in addition to the quasi - degenerate couple of levels ( @xmath2068 with @xmath190=0,1,2,3 ) other couple of levels ( @xmath2528 with @xmath253 ) be in close proximity , leading unwanted secondary trapping processes , acting against the efficiency of the filtration strategy . ) .
( b ) estimate of the population decay as a function of time , using the adiabatic approximation of eq .
( [ p - undiss ] ) .
color code for both panels : @xmath2069 in red and @xmath20610 in blue , thick solid lines for zwr(8 , @xmath190=0 ) , dashed lines for zwr(8 , @xmath190=1 ) , thin solid lines for zwr(8 , @xmath190=2,3).,title="fig : " ] ) .
( b ) estimate of the population decay as a function of time , using the adiabatic approximation of eq .
( [ p - undiss ] ) .
color code for both panels : @xmath2069 in red and @xmath20610 in blue , thick solid lines for zwr(8 , @xmath190=0 ) , dashed lines for zwr(8 , @xmath190=1 ) , thin solid lines for zwr(8 , @xmath190=2,3).,title="fig : " ] a quantitative analysis can be provided based on the definition , for a given wavelength , of a dimensionless parameter @xmath254 which measures the energy proximity among a given couple @xmath238 and @xmath240 : @xmath255 @xmath256 labels one of the two anharmonicity windows of the vibrational levels sequence , with an energy extension @xmath257 .
more precisely , for a given @xmath31 , @xmath258 depicts a situation where a given @xmath240 is within the window @xmath259 ( for example @xmath260 ) , whereas @xmath261 corresponds to @xmath262 ( for example @xmath263 ) .
obviously @xmath235 may vary between 0 ( exact degeneracy @xmath264 ) to 1 .
in practical calculations done for the different zwr paths originating from @xmath2068 , we obtain semiclassical estimates for @xmath265 typically less than 0.04 .
the additional important information emerges from the analysis of @xmath266 and @xmath267 addressing the following question : when a given @xmath190 is almost degenerate with @xmath2068 , how much is ( @xmath268 ) close to @xmath269 , or ( @xmath270 ) to @xmath271 ?
the histograms of fig.([histogram ] ) , displaying the corresponding values of @xmath235 , for each of the zwr paths , answer this question and help in the choice of the best adapted path leading to the maximum contrast , that is for an almost zero estimate for @xmath272 , the largest possible values for @xmath273 and @xmath274 , in order for the filtration be the most efficient .
this is obviously obtained for zwr(@xmath186 , @xmath187 ) on which the filtration will ultimately be based .
this choice is also confirmed by looking at the floquet results displayed in fig .
( [ morphology_zwr].b ) .
there are three local minima for the resonance width @xmath275 .
the third minimum ( wavelenght around 576.4@xmath174 ) , which corresponds to the semi - classical coincidence of @xmath276 , is the one which faces simultaneously the two best - marked maxima for both @xmath277 and @xmath278 .
the contrast is then optimal in comparison to the two other possible minima of @xmath275 appearing around @xmath279 and @xmath280 .
this semiclassical analysis conducted on near zero - field limit , can be confirmed by the time - dependent quantum behavior of resonance widths during their adiabatic transport all along the zwr paths of fig.([loci_zwr ] ) .
the result is displayed in fig.([gammas910]a ) for resonances originating from @xmath269 and @xmath271 with laser pulses shaped and chirped as to follow the zwr paths ( @xmath186 , @xmath193 ) .
a much more pronounced decay rate ( larger imaginary parts of the energies ) is observed when following zwr ( @xmath186 , @xmath190=0 ) . within a purely adiabatic hypothesis of a single ( not contaminated ) resonance transport ,
these rates can be used to calculate the resulting vibrational population decays during the pulse through eq.([p - undiss ] ) .
the results are gathered in fig.([gammas910]b ) which clearly shows that the most important @xmath2069 and 10 populations depletion are observed when the transport is along the zwr ( @xmath186 , @xmath190=0 ) path .
this is a complete confirmation of the semiclassical choice of this particular path as producing the highest possible decay of @xmath2069 , 10 populations , while quenching the one of @xmath2068 . ) .
thick solid lines for the path corresponding to zwr ( @xmath281 ) , dashed lines for the path ( @xmath282 ) , thin solid lines for the path ( @xmath283 ) .
color code : black for @xmath2068 , red for @xmath2069 , blue for @xmath20610 . ]
the implementation of the filtration scheme , within the hypothesis of an initial ensemble of vibrational states , or of an initial coherent vibrational superposition , is summarized in fig.([populations ] ) .
we assume an equally populated field - free initial vibrational distribution among levels @xmath186 ( in black ) , @xmath269 ( in red ) and @xmath271 ( in blue ) .
three laser pulses are considered , which adiabatically follow the zwr originating from @xmath186 as depicted in fig.([loci_zwr ] ) and shaped such as to quench @xmath186 dissociation .
zwr ( @xmath281 ) is indicated by thick solid lines , zwr ( @xmath282 ) by dashed lines , and zwr ( @xmath283 ) by thin solid lines .
full quantum dynamical simulations based on wavepacket propagation are used for solving the tdse eq.([tdse ] ) as explained in section [ nummethod ] .
the first observation is that for all zwr paths the initial population of @xmath2068 is well protected against dissociation , as is expected from an adiabatic transport .
the final @xmath2068 undissociated populations are all above 95@xmath2 , which remains a remarkable result for the zwr quenching mechanism itself .
the second important information concerns the overall filtration process efficiency . while for zwr ( @xmath283 ) a large remaining population on @xmath2069 , 10 levels ( more than 80@xmath2 and 50@xmath2 respectively ) renders the filtration inefficient , only the choice of zwr ( @xmath186 , @xmath187 ) leads to acceptable results , with less than 10@xmath2 of undissociated population on @xmath284 .
the vibrational cooling control is achieved as the system is left on a single vibrational level , namely @xmath186 .
a final stirap process @xcite could optionally be referred to for transferring the population from @xmath186 to @xmath285 .
we propose a laser control strategy for the vibrational cooling of diatomic molecules based on an efficient filtration scheme.the basic mechanism which is referred to is a destructive interference between outgoing fluxes from the two field - dressed electronic states involved in the photodissociation process . in principle , infinitely long lived resonances ( zwrs ) result from such interference , for some well tuned couples of laser parameters ( wavelength and intensity ) .
starting from an initial vibrational population distribution , filtration aims at a selective decay among them ; namely , all levels , except one , should dissociate .
the proposed scheme goes through the optimization of a laser pulse envelop shaped and frequency chirped in such a way as to protect against dissociation a given initial vibrational level @xmath31 .
this is achieved by an adiabatic transport of the @xmath31 population on its corresponding zwr which is continuously followed all along the pulse duration .
two important issues that turn out to be crucial for the success of the process when dealing with heavy systems like na@xmath0 , and that have not been addressed previously , are emphasized in this work .
the more fundamental first one concerns the challenging problem of adiabaticity in decaying system dynamics involving a dissociative continuum .
actually this is solved by adapting the adiabatic floquet approach ( originally derived for bound states ) to the special case of zwrs , taking advantage of their non - decaying peculiarity , although being coupled to a dissociative continuum ( bics , bound states in continuum ) .
it is only through such an adiabatic transport that we can protect the population of a vibrational level in a robust way .
the second issue concerns more specifically the application in mind .
we have shown that the family of zwrs , originating from the field - free initial vibrational level @xmath31 to be protected , is organized in terms of several paths in the laser parameter plane .
if all the members of this family are convenient to quench the dissociation from @xmath31 , some of the laser pulses built to follow such paths may have characteristics close to be appropriate for partly protecting neighboring vibrational populations as well .
this situation arises when transposing the control scheme to heavier diatomics like na@xmath0 , basically due to anharmonicity and relatively high vibrational levels density .
its occurrence is obviously against the efficiency of the filtration process .
a semiclassical analysis developed in detail helps the choice of the pulse offering the best compromise between quenching @xmath31 and providing highest possible decay rates for the neighboring vibrational states @xmath286 ( @xmath287 ) .
a full quantum wavepacket propagation using the optimal pulse confirms the expectations of both the adiabatic floquet theory and the semiclassical analysis . in the hypothesis of an initial equal vibrational population partition on @xmath2068,9 , and 10 for na@xmath0 prepared by photoassociation
, @xmath2068 is protected up to 95@xmath2 of its initial population , whereas only less than 5@xmath2 of the @xmath2069 and 10 populations are left .
it is worthwhile noting that such a control is achieved with a total pulse duration not exceeding 12@xmath177 , thus avoiding rotational heating of the molecule .
our final claim is that the proposed laser pulse characteristics , namely rather modest frequency chirp amplitudes around @xmath288 , moderate intensities ( less than @xmath289 ) , realistic pulse duration ( less than 12@xmath177 ) , remain experimentally feasible .
moreover , within the adiabatic transport frame that has been developed , the model becomes finally generic enough to be transposed to the vibrational cooling of other alkali dimers .
o. a. acknowledges support from the european union ( project no .
itn-2010 - 264951 , corinf ) .
part of the simulations have been executed on computers of the utinam institute of the universit de franche - comt , supported by the rgion de franche - comt and institut des sciences de lunivers ( insu ) . | in molecular photodissociation , some specific combinations of laser parameters ( wavelength and intensity ) lead to unexpected zero - width resonances ( zwr ) , with in principle infinite lifetimes .
their interest in inducing basic quenching mechanisms have recently been devised in the laser control of vibrational cooling through filtration strategies [ o. atabek et al .
, phys .
rev .
a87 , 031403(r ) ( 2013 ) ] . a full quantum adiabatic control theory based on the adiabatic floquet hamiltonian is developed to show how a laser pulse could be envelop - shaped and frequency - chirped so as to protect a given initial vibrational state against dissociation , taking advantage from its continuous transport on the corresponding zwr , all along the pulse duration . as compared with previous control scenarios actually suffering from non - adiabatic contamination , drastically different and much more efficient filtration goals are achieved .
a semiclassical analysis helps in finding and interpreting a complete map of zwrs in the laser parameter plane .
in addition , the choice of a given zwr path , among the complete series identified by the semiclassical approach , amounts to be crucial for the cooling scheme , targeting a single vibrational state population left at the end of the pulse , while all others have almost completely decayed .
the illustrative example , offering the potentiality to be transposed to other diatomics , is na@xmath0 prepared by photoassociation in vibrationally hot but translationally and rotationally cold states . |
low energy electromagnetic properties of baryons , such as mass , charge radius , magnetic and quadrupole moments are important observables for any model of the nucleon structure . in various constituent
quark models a tensor force in the inter quark hyperfine interaction , introduced first by de rujula , georgi and
glashow @xcite , leads to a @xmath10state admixture in the baryon ground state wave function . as a result
the tensor force induces a small violation of the becchi
morpurgo selection rule @xcite , that the @xmath2 excitation is a pure @xmath11 ( magnetic dipole ) transition , by introducing a non vanishing @xmath12 ( electric quadrupole ) amplitude . for chiral quark models or in the skyrmion picture of the nucleon ,
the main contribution to the @xmath12 strength stems from tensor correlations between the pion cloud and the quark bag or meson exchange currents between the quarks . to observe a static deformation ( @xmath10state admixture ) a target with a spin of at least @xmath13 ( e.g. @xmath14 matter ) would be required .
the only realistic alternative is to measure the transition @xmath12 moment in the @xmath15 transition at resonance , or equivalently the @xmath4 partial wave amplitude in the @xmath16 decay .
the amplitudes in the @xmath17 final state are usually denoted by @xmath18 and @xmath19 , where @xmath20 and @xmath21 are the electric and magnetic multipoles , @xmath22 is the isospin and @xmath23 is the orbital angular momentum of the @xmath17 system and the @xmath24 sign refers to its total angular momentum @xmath25 .
the experimental quantity of interest to compare with the different nucleon models is the ratio @xmath26 of the electric quadrupole @xmath12 to the magnetic dipole @xmath11 amplitude in the region of the @xmath27 resonance . in quark models with @xmath28 symmetry , for example the mit bag model ,
@xmath29 is predicted .
depending on the size of the hyperfine interaction and the bag radius , broken @xmath28 symmetry leads to @xmath30 @xcite .
larger negative values in the range @xmath31 have been predicted by skyrme models @xcite while results from chiral bag models @xcite give values in the range @xmath32 to @xmath33 .
the first lattice qcd result is @xmath34 @xcite and a quark model with exchange currents yields values of about @xmath35 @xcite .
the determination of the quadrupole strength @xmath12 in the region of the @xmath27 resonance has been the aim of a considerable number of experiments and theoretical activities in the last few years .
very recently , new experimental results have been published for the differential cross section and photon asymmetry of pion photoproduction off the proton from the mainz microtron mami and the laser backscattering facility legs at brookhaven national laboratory , with the results @xmath36 from the mainz group @xcite and @xmath37 from the legs group @xcite .
these new @xmath38 results have started intense discussions about the correct way to extract the @xmath7 ratio from the new experimental data .
in particular the large variation in the @xmath38 values obtained in theoretical analyses of these data at rpi @xcite ( @xmath39 ) , vpi @xcite ( @xmath40 ) and mainz @xcite ( @xmath41 ) was quite unsatisfactory .
in this paper we present the mami @xmath0 and @xmath1 differential cross sections and photon asymmetries and discuss in detail different analyses to extract the @xmath42 and @xmath43 partial wave amplitudes and the @xmath7 ratio . in section ii we briefly describe the experimental setup and show a selection of the measured cross sections and photon asymmetries for both pion production channels from the proton in section iii .
the essential ingredients of a multipole analysis are outlined in section iv .
three different analyses of our data to extract the @xmath42 and @xmath43partial wave amplitudes and the @xmath7 ratio are discussed in sections v and vi .
we conclude with a summary and outlook in section vii .
since the experimental setup used for this measurement was described in detail in ref .
@xcite , we will restrict the present discussion to the main features of the experiment .
linearly polarized photons were produced by coherent bremsstrahlung in a @xmath44 thick diamond crystal @xcite .
the photon energy was determined by the glasgow tagging spectrometer at the mainz microtron mami @xcite , which in a 352 channel focal plane detector analyses the momentum of the electron that has radiated the bremsstrahlung photon @xcite .
this detector system is able to energy tag photons in the range from @xmath45 to @xmath46 with a resolution of about @xmath47 @xcite .
the collimation of the photon beam yielded a tagging efficiency of about @xmath48 for incoherent bremsstrahlung . to continuously monitor the tagging efficiency and the photon polarization ,
a pair detector was used downstream of the hadron detector daphne .
this pair detector consists of a @xmath49 mm thick cu converter followed by two @xmath50 mm thick plastic scintillators operated in coincidence . its efficiency ( @xmath51 for photons )
was regularly checked with a lead glass detector ( @xmath52 for photons ) in calibration runs at low beam intensity .
the photon polarization was determined from the photon spectrum measured by the tagging spectrometer in coincidence with the pair detector and with the aid of theoretical calculations @xcite .
the quality of these calculations was tested by an absolute measurement of the photon polarization using coherent @xmath53 photoproduction on @xmath54 as a polarimeter reaction with an analyzing power @xmath55 .
excellent agreement between calculations and experiment was found @xcite and in this way , both the photon polarization and the photon flux could be determined with an absolute precision of better than @xmath56 .
the liquid hydrogen cryogenic target was contained in a @xmath57 mm diameter , @xmath58 mm long mylar cylinder with a wall thickness of @xmath59 mm .
the target density was stabilized and determined to an accuracy of @xmath60 by means of an an automatic pressure and temperature control system .
the reaction products , the recoil proton from @xmath61 and the pion from @xmath62 , were detected using the large acceptance detector daphne @xmath63 built by the cea / dapnia sphn at saclay and the infn - sezione di pavia @xcite .
good definition of the charged particle tracks was obtained from the central vertex detector consisting of @xmath64 coaxial cylindrical multiwire proportional chambers providing a polar angular resolution of @xmath65 fwhm and an azimuthal resolution @xmath66 fwhm .
this vertex detector is surrounded by a segmented @xmath67 plastic scintillator telescope with successive thicknesses of @xmath68 mm , @xmath69 mm and @xmath70 mm respectively .
the outermost layer is a lead aluminium scintillator sandwich designed to enhance the @xmath53 detection efficiency and to provide additional energy loss information for charged particles . in the first step of the analysis
those events were selected , that had only one charged trajectory with its polar angle in the range @xmath71 .
after this cut , the basic task of the data analysis was to identify the pions and protons .
this separation was performed by using a range method as described in detail in ref .
@xcite , which simultaneously uses all of the measured energy losses in the scintillator layers of daphne to identify charged hadrons ( @xmath72 , @xmath43 ) and to determine their energy . a restriction made on the vertex position , defined as the point on the reconstructed track , which lies closest to the detector axis , ensured a complete rejection of particles coming from the target windows and walls . as a result
no empty target subtraction was needed .
the main contributions to the systematic error in the determination of the unpolarized differential cross section are due to uncertainties in the photon flux ( @xmath56 ) , the target density ( @xmath73 ) and ( @xmath56 ) for the proton / pion separation with the range method .
in single pion photoproduction , the differential cross section for linearly polarized photons and unpolarized targets is given by the expression @xmath74 where @xmath75 is the photon asymmetry and @xmath76 and @xmath77 are the polar and azimuthal angles of the pion with respect to the beam direction . since daphne has full 2@xmath78 azimuthal coverage it allows a direct measurement of the @xmath77 dependence of the differential cross section , and therefore , the determination of @xmath75 and the unpolarized cross section @xmath79 at the same time . a selection of the measured cross sections and photon asymmetries in the cm - frame for the @xmath61 and @xmath80 reaction channels @xcite are shown in figs .
[ fig : pi0d][fig : pips ] .
this is the first data set for which the two observables @xmath79 and @xmath75 have been measured simultaneously at all angles and photon energies for both pion production channels from the proton .
the measurement covers the whole @xmath27 resonance region in @xmath81 energy bins between @xmath82 and @xmath6 , where @xmath83 is the tagged photon energy in the lab - frame .
the angular distributions for @xmath79 and @xmath75 were measured from @xmath84 up to @xmath85 for the @xmath86 channel . in the @xmath87 case
the angular range varied from @xmath88 to @xmath89 at @xmath82 to @xmath90 to @xmath89 for the highest energy point at @xmath91 . in figs .
[ fig : pi0d][fig : pips ] the result of a fixed
t dispersion relation analysis @xcite , which will be discussed below , is shown as a solid line .
our differential cross sections and asymmetries are in good agreement with the dispersion relation analysis except for the differential cross sections around @xmath92 where we find a slight discrepancy
. figs .
[ fig : pa90 ] and [ fig : pa85 ] compare the energy dependence of the present mami results and the recent legs results @xcite for @xmath61 at @xmath93 and @xmath80 at @xmath94 .
for the legs results the statistical error and the angle and energy dependent error have been evaluated point by point and combined to the net uncertainty bars , which are shown in figs .
[ fig : pa90 ] and [ fig : pa85 ] .
the error bars on the mami points are only statistical .
the systematic scale uncertainties are @xmath95 for the mami results and @xmath96 for the legs results .
there is an energy dependent discrepancy of up to @xmath97 between the absolute differential cross sections . for both @xmath61 and @xmath80 ,
the legs cross sections start to rise above the mami data at @xmath98 , reaching @xmath99 higher values at the highest energy legs data point at @xmath100 .
since the difference is energy dependent it will affect the absolute values of the resonant partial waves @xmath3 and @xmath4 ( see discussion below ) . in fig .
[ fig : wq ] the legs and mami angular distributions for the differential cross section and the photon asymmetry are compared at @xmath101 for @xmath0 and @xmath1 .
one sees that the differential cross sections differ not only in absolute magnitude , but also in angular shape .
the two most forward angular points at @xmath102 and @xmath103 of the legs @xmath104 differential cross section drop faster than the mami results .
the same behavior is seen for the @xmath105 cross section at @xmath106 , @xmath107 and @xmath108 .
this difference in the shape of the differential cross sections for the two data sets will play an important role in the discussion of the non born contribution for partial waves with @xmath109 . on the other hand
, the photon asymmetry agrees well for the @xmath104 channel and shows only small differences for the @xmath105 data , where the mami results are slightly above the legs data .
in @xmath110 both incident particles and the final nucleon have two spin states yielding in eight degrees of freedom . from parity conservation this
is reduced to a total number of four complex amplitudes to describe the reaction . allowing for one arbitrary phase factor , there are therefore seven independent physical quantities which need to be measured at any setting of @xmath83 and @xmath111 . different sets of amplitudes are used in the literature , e.g. the helicity amplitudes ( @xmath112 , @xmath113 , @xmath114 , @xmath115 ) introduced by jacob and wick @xcite or the pauli amplitudes ( @xmath116 , @xmath117 , @xmath118 , @xmath119 ) by chew , goldberger , low and nambu @xcite . while the observables of single pion photoproduction are more elegantly expressed by the helicity amplitudes , the pauli amplitudes are particularly suited for a decomposition into partial waves @xcite .
a complete database for pion photoproduction ( the `` complete '' experiment ) requires at least @xmath120 independent observables to specify the multipole amplitudes to all orders in @xmath121 @xcite .
such complete information is not available at present and standard multipole analyses have to rely on the differential cross section @xmath122 and the three single polarization observables @xmath75 ( photon asymmetry ) , @xmath123 ( recoil nucleon polarization ) and @xmath124 ( target asymmetry ) . for pion photoproduction from threshold up to the @xmath27 resonance region , these four observables provide sufficient conditions for a complete database , if the higher partial waves @xmath125 can be adequately represented by the born contributions .
this approach is expected to be appropriate up to the region of the higher resonances ( @xmath126 ) and even there may affect only certain multipoles ( e.g. @xmath127 ) .
such arguments were used by v. grushin @xcite to analyse the kharkov data @xcite ( @xmath122 , @xmath75 , @xmath123 and @xmath124 ) for @xmath61 and @xmath80 . in their analysis both the real and imaginary parts of the @xmath42 and @xmath43 wave amplitudes could be determined for the first time independently of the pion - nucleon phase shifts @xmath128 . the mean difference between the @xmath129 phase from the kharkov analysis and the known pion
nucleon scattering phase @xmath130 was only @xmath131 over the entire energy range @xmath132 .
the @xmath104 and @xmath105 multipole amplitudes @xmath133 ( @xmath133 stands for @xmath134 , @xmath135 ) are related to the isospin @xmath136 and @xmath13 components @xmath137 and @xmath138 by @xmath139 the multipole amplitudes @xmath140 are complex functions of the c.m .
energy w. below the two pion production threshold ( @xmath141 ) , the fermi watson theorem @xcite allows one to express the phases of the complex multipole amplitudes by the corresponding pion
nucleon scattering phase shifts @xmath142 @xmath143 where @xmath22 can be @xmath136 and @xmath13 and @xmath144 is an integer .
however , even at energies of about @xmath145 , the @xmath146 inelasticities in the @xmath147 partial wave are very small , which suggests that the fermi
watson theorem can be applied well above the two pion threshold .
there exist two basic methods to extract the multipole amplitudes from the database , the `` energy independent '' and the `` energy dependent '' approach . in the energy independent approach
, each energy is investigated independently by use of standard @xmath148 minimization techniques , the fit parameters being the real and imaginary parts of the multipole amplitudes @xmath149 .
below the two pion production threshold the fermi
watson theorem is used , which reduces the number of the necessary observables for a complete experiment by a factor of two , because only the absolute values @xmath150 of the partial wave amplitudes need to be determined from the fit . in the energy
dependent approach , the data at all energies are analyzed simultaneously .
either an energy dependent parametrization of the partial wave amplitudes must be assumed , or as in our case , the energy dependence is taken from dispersion relations .
the principal advantage of this method is that continuity is built in from the beginning and systematic errors tend to cancel out
. in some cases one or the other of these two approaches is to be preferred .
if the data are closely spaced in energy and cover both differential cross sections and polarization observables ( `` complete experiment '' ) at each energy , the energy independent approach is more advantageous .
if on the other hand the data are widely spaced and only few polarization observables are available , the energy dependent approach is the better one .
this approach is also useful if the general resonance structure is already known and the main interest is to obtain the small partial wave amplitudes .
however , the experimental data has to cover the complete energy range of the dominating resonances in order to allow a reliable separation .
for instance in the case of the @xmath151 amplitude in the @xmath2 transition , one has to cover the complete resonance region ( @xmath152 ) to get the @xmath151 multipole as precisely as possible @xcite . a general problem common to both the energy dependent and the energy independent approach is to decide at which value of the angular momentum @xmath121 the partial wave expansion should be truncated in the fit and how to approximate the higher partial waves .
the maximum value @xmath153 that the data can determine is found by fitting the angular distributions to a legendre series or , equivalently , a power series expansion in @xmath154 @xmath155 the experimentally accessible polynomial coefficients @xmath156 are quadratic or bilinear products of the electric @xmath157 and magnetic @xmath158 multipole amplitudes . in this section
we examine the sensitivity to higher partial wave contributions of the coefficients @xmath156 and @xmath159 extracted from the present data .
the difference of the cross sections for the photon polarization perpendicular and parallel to the reaction plane , @xmath160 , is particularly sensitive to higher partial waves . with partial waves up to @xmath161 the difference
@xmath162 has three polynomial coefficients @xmath163 , @xmath164 and @xmath165 with @xmath166 + b_{\sigma } & \simeq & { \rm re } \left [ ( m_{1+}-m_{1- } ) d^*_{wave } \right ] \ ; , \\
c_{\sigma } & = & | d_{wave } |^2 \ ; . \end{aligned}\ ] ] in the case that only the @xmath42 wave ( @xmath167 ) and @xmath43 waves ( @xmath168 and @xmath169 ) contribute , this difference would be proportional to @xmath163 and therefore independent of the pion angle @xmath76 .
the multipole dependence of the coefficients @xmath163 , @xmath164 and @xmath165 is described in appendix a. fig .
[ fig : sp ] shows the difference of @xmath170 for both pion reaction channels on the proton at our lowest photon energy @xmath171 , at @xmath172 and at our highest energy @xmath173 . in the @xmath87 channel
one recognizes only above the resonance a small deviation from the constant behavior .
this is due entirely to the born contribution to the @xmath164 coefficient , i.e. the interference between the real part of the dominant @xmath151 amplitude and the real part of the @xmath10 wave amplitudes , for example in terms like @xmath174 .
such contributions become extremely small at resonance since @xmath175 .
the behavior of the difference , @xmath170 , for the @xmath176 data from the legs collaboration is different .
their observed angular dependence seems to require a sizeable non born contribution from @xmath10 and @xmath177 waves .
the origin of the different angular dependence of the legs results for @xmath160 arises from the different shape of their unpolarized cross section data .
the legs photon asymmetries alone are well described by our multipole fit which takes born contributions in higher partial waves of @xmath109 into account .
in contrast the difference of @xmath160 for @xmath178 shows a strong angular dependence at all energies .
the main reason is that in charged pion photoproduction the pion pole born graph leads to significant contributions of higher partial waves .
the sensitivity of the differential cross section to higher partial waves is most pronounced at the extreme forward and backward angles .
taking partial waves up to @xmath161 , the differential cross section is @xmath179 with five coefficients of the form @xmath180 + b & \simeq & b(s_{wave},p_{wave } ) + { \rm re } \left [ ( m_{1+}-m_{1- } ) d^*_{wave } \right ] \ ; , \\ c & \simeq & c(s_{wave},p_{wave } ) + { \rm re}\left [ e_{0 + } d^*\!\!_{wave } \right ] + | d_{wave } |^2 \ ; , \\ d & \simeq & { \rm re } \left [ ( m_{1+}-m_{1- } ) d^*\!\!_{wave } \right ] \ ; , \\ e & = & | d_{wave } |^2 \ ; . \end{aligned}\ ] ] the effect of the @xmath10 waves is largest for the coefficients b and d , because of an interference term between the large @xmath151 amplitude and the @xmath10 waves .
however , the contributions of these terms can be neglected at the top of the resonance ( @xmath181 ) . as an example
this is illustrated for the combination @xmath182 & = & { \rm re } ( m_{1+}-m_{1- } ) { \rm re } e_{2- } + { \rm i m } ( m_{1+}-m_{1- } ) { \rm i m } e_{2- } \ ; , \end{aligned}\ ] ] where the first term vanishes , because @xmath183 passes through zero at the resonance position ( @xmath181 ) and the second term can be neglected , because @xmath184 is small due to the small d - wave phase .
[ fig : saidsp ] illustrates the sensitivity of the differential cross section to higher partial waves .
the solid line shows the ratio ( @xmath185 ) of the differential cross section for only @xmath42 and @xmath43wave contributions to the differential cross section with higher partial waves .
the calculated cross section ratio is shown at the pion angles @xmath186 and @xmath187 in the energy region between @xmath188 and @xmath189 .
for this calculation @xcite the said solution sm95 was used , in which partial waves up to @xmath190 were allowed to vary to fit the data set but the normal born contribution was assumed for @xmath191 . at @xmath93 , the contributions from the higher partial waves are far below @xmath192 , because they arise only from an interference term with the small @xmath42 wave @xmath134 , for example in the form @xmath193 as in eq . 15 . below and above the @xmath27 resonance , however , contributions from @xmath109 are of the order of @xmath194 of the differential cross section at @xmath195 and @xmath187 .
these contributions arise almost completely from the interference between the dominant @xmath151 amplitude and the born parts of the higher partial waves . in our analysis
there are no indications of significant non born contributions for higher partial waves ( @xmath109 ) .
an important issue for any multipole analysis is the question of how to handle the systematic errors of the experimental results for the different observables used in the database .
the ideal database should contain experimental results for both @xmath61 and @xmath80 , measured simultaneously with one setup for all photon energies and covering the full range of polar angles to minimize energy and angular dependent systematic errors .
the analysis of such data would yield a consistent separation of the isospin @xmath136 and @xmath13 parts from the partial waves of the @xmath87 and @xmath86 reaction channels .
however , the isospin separation will fail if there is a considerable inconsistency , such as a normalization error , between the @xmath80 and @xmath61 observables ( @xmath122 , @xmath75 , @xmath124 , ) .
the existing pion photoproduction data , for example the said database are far from the requirements of the ideal case .
most of the experimental data are taken at one photon energy at a time , the angular distributions are not measured simultaneously for all angles and , more importantly , the consistency between different measurements is less good than their assigned errors would suggest .
when combining data from different experiments into one database , one has the difficult task to account properly for the angular and energy dependent systematic errors .
a common procedure is to combine the experimental systematic errors in quadrature with the statistical errors or to multiply all data from one set with a systematic scale error ( @xmath196 ) by a common factor @xmath177 while adding @xmath197 to the @xmath148 @xcite .
the latter method allows for adjustments of an overall scale ( angle independent systematic errors ) , but not for systematic effects in the shape of the angular distributions which is important for the determination of the small @xmath12 amplitude . as shown in the following paragraph
, it is the shape of the differential cross section and the photon asymmetry near @xmath198 , however , which is sensitive to the @xmath12 amplitude .
combining all existing @xmath104 and @xmath105 data in one database regardless of their consistency will result in mean values for the dominant partial waves ( @xmath151 and @xmath167 ) but meaningless values for the small partial waves ( @xmath169 and @xmath199 ) .
any reliable partial wave analysis for the small amplitudes will require a careful selection of the experimental results used in the database .
in practice one has to choose observables that are sensitive to the small amplitude from one experiment , or one has to use analyzing methods which are insensitive to energy and angle dependent systematic uncertainties . fig .
[ fig : sp340 ] shows @xmath122 and @xmath75 for different multipole contributions at the @xmath27 resonance ( @xmath9 ) .
this illustrates the sensitivity to the higher partial waves ( @xmath109 ) and to the small @xmath169 amplitude .
the contribution of the higher partial waves to the @xmath200 channel is significant at all angles . for the differential cross section at @xmath201 and the photon asymmetry around @xmath202 the full calculation and the truncation to
@xmath42 and @xmath43 waves differ by as much as a factor of two .
the @xmath61 case is different and one can not distinguish between the full calculation and the result where only s and p waves are taken into account .
the main reason is that the pion pole term does not contribute in neutral pion photoproduction .
[ fig : sp340 ] shows also the sensitivity of @xmath75 and @xmath122 to the small @xmath169 amplitude for both pion channels .
the main effects of the @xmath169 multipole in the @xmath104 channel is found near @xmath203 for the photon asymmetry and near @xmath204 and @xmath187 for the differential cross section .
this sensitivity has its origin in the interference term @xmath205 , which appears in the coefficients @xmath206 and contributes to both @xmath122 and @xmath75 . without @xmath169 strength the shape of the differential cross section
would be determined by @xmath207 ; and the photon asymmetry by @xmath208 . for @xmath209
these values would change to @xmath210 and @xmath211 . setting the @xmath169 multipole to zero ( dotted line ) results in an increased differential cross section at the extreme angles @xmath195 and @xmath187 and a lower value at @xmath212 .
the photon asymmetry in contrast is enhanced at @xmath212 for @xmath213 . the same behavior for @xmath75
is found in the @xmath105 channel . in the differential cross section for @xmath105
the effect of @xmath169 is most sizeable at the backward angles . from the above discussion
it becomes obvious that a reliable extraction of the small @xmath169 amplitude will require precise photon asymmetry data around @xmath212 and precise data for the differential cross section at all angles .
we have performed three different analyses of our data in order to extract the @xmath42 and @xmath43wave amplitudes and the @xmath7 ratio .
first , a fit to the differential cross section and photon asymmetry only in the @xmath61 channel ; second , an energy independent simultaneous multipole analysis of the @xmath61 and @xmath80 data to get the isospin separation in the whole energy region ( @xmath214 ) . in the third analysis an energy dependent fixed
t dispersion analysis has been performed , which includes additional observables in order to study the stability of the different multipole solutions . since the result from the fit to the @xmath215 channel alone
has already been published @xcite , we summarize here only the key points of our analysis . the @xmath104 angular distributions for the unpolarized cross section @xmath216 , the parallel part @xmath217 ( pion detected in the plane defined by the photon polarization and the photon momentum vector ) , and perpendicular part @xmath218 can be expressed in the @xmath42 and @xmath43wave approximation by the parametrization @xmath219 where @xmath220 and @xmath221 denote the center of mass momenta of the pion and the photon , respectively , and @xmath222 indicates the parallel ( @xmath223 ) , perpendicular ( @xmath224 ) and unpolarized ( @xmath225 ) components .
the coefficients @xmath226 , @xmath227 and @xmath228 are quadratic or bilinear functions of the @xmath42 and @xmath43wave amplitudes . in particular
, @xmath229 is sensitive to the @xmath230 amplitude , because of interference with @xmath151 in the terms @xmath231 \ ; , \\ c_{\parallel } & = & 12{\rm re}[e_{1 + } ( m_{1 + } - m_{1-})^ * ] \ ; .\end{aligned}\ ] ] furthermore , the ratio @xmath232 is very close to the ratio @xmath233 as we will show below . at the @xmath27 resonance ( @xmath9 )
the ratio r can be identified with the @xmath38 value , @xmath234 this is the key point of our analysis .
the method has the advantage of being independent of absolute normalization and insensitive to many systematic errors , because @xmath38 is extracted from the ratio of the coefficients @xmath235 and @xmath236 fitted to the angular distribution of @xmath217 . analyzing the data by this method
the following result is obtained @xcite : @xmath237 without any correction for neglecting the isospin @xmath136 contributions to @xmath169 and @xmath151 , the maximum absolute systematic error for @xmath38 would be @xmath238 . to further reduce this systematic error one
can take the born contribution and estimate the size of the isospin @xmath136 contribution for @xmath169 and @xmath151 . with a conservative estimate
the absolute systematic error reduces to less than @xmath239 ( see appendix b for details ) . to obtain the isospin decomposition for the @xmath42 and @xmath43 waves in the whole energy range ( @xmath240 ) , we have performed an energy independent multipole analysis of our data . in this analysis
we fit @xmath120 parameters , the absolute values of the @xmath42 and @xmath43 waves for isospin @xmath136 and @xmath13 ( @xmath241 , @xmath242 , @xmath243 , @xmath3 , @xmath244 , @xmath4 , @xmath245 and @xmath246 ) , to the photon asymmetry and the differential cross section for the @xmath215 and @xmath200 reaction channels .
higher partial waves ( @xmath125 ) are taken into account by the born terms , including @xmath247 and @xmath248 exchange in the @xmath249-channel .
we obtain the coefficients @xmath250 , @xmath251 and @xmath252 from the angular distribution of the differential cross section and @xmath163 from the photon asymmetry .
the two pion reaction channels are decribed by @xmath120 coefficients , which are independent combinations of the isospin @xmath136 and @xmath13 multipole amplitudes of the @xmath42 and @xmath43 waves .
in addition the fermi
watson theorem is used to determine the real and imaginary parts of the @xmath42 and @xmath43wave multipole amplitudes . in fig .
[ fig : iso ] we show our fitted multipole amplitudes as @xmath253 , which is the absolute value of the amplitude up to a sign that can change with photon energy as in the case of @xmath4 .
the values of the @xmath42 and @xmath43wave amplitudes are fitted independently at each energy to the mami data @xmath61 and @xmath80 .
the results for the isospin components for the @xmath42 and @xmath43wave amplitudes are presented for 16 photon energy bins@xcite .
although systematic errors are greatly reduced because the @xmath104 and @xmath105 data are obtained with the same set up , this method is sensitive to small remaining systematic differences in the results of @xmath122 and @xmath75 for the two pion channels , because it relates 8 parameters to the 8 `` observables '' ( a , b , c and @xmath163 coefficients for the two pion channels ) .
including additional polarization observables in the database and performing an energy dependent multipole analysis overcomes these shortcomings . in a third analysis
the mami pion photoproduction data are analyzed using fixed
t dispersion relations based on lorentz invariance , isospin symmetry , unitarity and crossing symmetry @xcite .
this analysis includes the recent mami data for the differential cross section @xmath122 and photon asymmetry @xmath75 for @xmath104 and @xmath105 from the proton @xcite , both older and more recent data from bonn for the target asymmetry @xmath124 @xcite , and differential cross section data on @xmath254 production off the neutron from frascati @xcite and recently from triumf @xcite . with this method we performed both an energy dependent and an energy independent multipole fit as shown in fig .
[ fig : e2m1 ] by open circles ( energy independent ) and the solid line ( energy dependent ) .
the agreement between the @xmath255 and @xmath256 multipole results obtained from the dispersion analysis and the multipole analysis described in the previous section is very good .
both energy independent results show the same energy behavior for the two multipoles . for the real and imaginary parts of the @xmath256 amplitude the two results are in excellent agreement , the solid dots lying on top of the open circles up to @xmath257 ( see fig . [
fig : e2m1 ] ) .
the agreement is also very good in the case of the small @xmath255 amplitude . in particular , we find that the inclusion of other observables in addition to the mami pion photoproduction data decreases the fluctuations slightly for the @xmath255 multipole .
the results of the energy dependent fit ( solid line in fig . [ fig : e2m1 ] ) agree well with the @xmath151 multipole derived from the energy independent fit , but small systematic differences may be seen for the electric multipole @xmath169 at energies above the @xmath27 resonance ( @xmath258 ) .
the ratio @xmath38 at the resonance agrees quite well in both dispersion analyses , @xmath259 for the energy dependent fit and @xmath260 for the energy independent fit .
in addition we have checked the effect of changes in the database .
replacing our differential cross section by the bonn data obtained in the seventies @xcite reproduces the results for the leading multipoles but changes the smaller multipoles to some degree ( see tab .
[ tab : rem ] ) .
the bonn results for @xmath122 at very forward and backward angles force the fit to a smaller @xmath169 value near the resonance .
however , it should be kept in mind that the overall compilation of the bonn @xmath61 data by genzel et al .
@xcite results from different experimental setups , which have been re appraised and re evaluated before combining them into one data set .
the main part of the angular distribution ( @xmath261 ) for @xmath122 was measured by detecting the recoil proton in a magnetic spectrometer @xcite , i.e. by fixed angle and single energy measurements .
the differential cross section for @xmath262 was measured by detecting the two decay photons of the @xmath53 with lead glass blocks in a different experiment @xcite .
each of these experiments has a systematic error of the order of @xmath263 , and it takes much less than this systematic error to change the shape of @xmath122 ( @xmath264 ratio ) and with this the @xmath38 value . in the case of the legs data the larger cross section leads to larger values for the resonant multipoles @xmath3 and @xmath4 .
the main reason for this is that the cross section difference between the legs and mami data is energy dependent ( resonance behavior ) as shown in fig .
[ fig : pa90 ] and fig .
[ fig : pa85 ] .
inclusion of the polarization data @xmath124 , @xmath75 and @xmath123 from kharkov @xcite does not affect our fit because of the large statistical and systematic errors especially for the @xmath61 channel .
according to the fermi watson theorem the @xmath4 and
@xmath3 partial waves have the same phase @xmath265 and the ratio @xmath266 is a real quantity . as shown in fig .
[ fig : e2m1_pa ] , this ratio is strongly dependent on the photon energy and varies from @xmath267 at @xmath171 to @xmath268 at @xmath173 .
the ratio @xmath38 is defined at the @xmath27 resonance position , where @xmath181 , by @xmath269 we note that within the k matrix fomalism this ratio is free of background contributions since @xmath270 is the k matrix pole .
the extraction of the genuine @xmath27resonance parts of the magnetic dipole and electric quadrupole multipoles is model dependent , see for example the work on dynamical models @xcite and the recently proposed speed plot analysis @xcite . in order to get a reliable value for @xmath38
it is important to use a database with small angle dependent systematic errors , because the small @xmath169 amplitude depends on the shape of @xmath122 and the absolute magnitude of the photon asymmetry at @xmath212 . to obtain a consistent separation for the isospin @xmath136 and @xmath13 parts of the partial waves ,
experimental data are required with small inconsistencies in the absolute normalization between the @xmath61 and @xmath80 observables . to reduce the influence of such systematics in the multipole analysis ,
the observables @xmath75 and @xmath122 for both reaction channels @xmath87 and @xmath86 were simultaneously measured at all angles and energies with the daphne detector at mami .
the three different analyses to extract @xmath38 from the mami data agree very well with each other , and at the @xmath27 resonance we get the final result @xmath271 in addition , the discussions about the extraction of the correct @xmath38 ratio from our data , which arose in the literature after our first publication @xcite can now be summarized as follows . in the comment of the vpi group @xcite , our @xmath87 data were included in the said database and a value @xmath272 was obtained .
as pointed out in our reply @xcite the difference between the vpi result and our value @xmath273 is due to the database used in the said analysis . as has been recently confirmed by the vpi group @xcite , the exclusion of all pre@xmath274 @xmath104 differential cross section data in the said database changes @xmath38 to @xmath275 .
it is obvious that the key problem arises from the inconsistencies in the different data sets used in the analysis .
the forward and backward angles for the differential cross sections used in the said database force the vpi solution to smaller @xmath38 values
. however , this solution overpredicts the mainz photon asymmetry data at resonance , @xmath276 , compared to the mainz result , @xmath211 .
the reason is that at resonance ( @xmath9 ) the fit has only one parameter , the @xmath38 value , to describe simultaneously the shape of the differential cross section @xmath277 and the photon asymmetry at @xmath212 @xmath278 as we have stated before , the compilation of all the existing experimental data into one database can result in mean values for the dominant multipoles and meaningless values for the small multipoles . for example , if one combines the new mami data with the bonn differential cross section results , then the @xmath38 value is affected by small systematic differences in the two data sets .
the bonn differential cross sections range from @xmath279 to @xmath280 while the mami cross sections cover angles from @xmath281 to @xmath89 and are slightly below the bonn data at @xmath282 . in the combined data set the shape of the differential cross section is changed , i.e. the @xmath264 ratio gets smaller .
this is the main reason that for the combined bonn and mami data set the @xmath38 value is even below the value which one gets from the bonn data alone .
this influence of the bonn data on the extracted value of the ratio @xmath38 has been confirmed by the bnl@xcite and rpi / vpi groups@xcite . in a second comment @xcite our @xmath87 data
were analyzed by the rpi group , who obtained the result @xmath283 .
however , the inclusion of our @xmath86 data in the database lowered this value to @xmath284 @xcite , in agreement with our analysis .
we made the first simultaneous and accurate measurements of the differential cross sections and photon asymmetries for the reactions @xmath0 and @xmath1 in the @xmath27 region . using this data
we have performed a multipole analysis to obtain the isospin decomposition of the @xmath42 and @xmath43wave multipoles and extract the @xmath7 ratio over the energy range @xmath285 .
our final results at the @xmath27 resonance position ( @xmath181 ) are @xmath286 , and @xmath287 and @xmath288 for the helicity amplitudes . in the meantime
more @xmath289 data @xcite for the differential cross section and the photon asymmetry have been taken .
this new experiment covers the full range of polar angles by observing the two decay photons in the taps detector at mami .
the analysis is in progress and will produce new differential cross sections for the extreme forward and backward angles . in this way we hope also to clarify the absolute normalization problem in the differential cross section between legs and mami .
the authors wish to acknowledge the excellent support of k.h .
kaiser and h. euteneuer and the accelerator group of mami , as well as many other scientists and technicians of the institut fr kernphysik at mainz , dapnia/ sphn at saclay and infn at pavia .
we would also like to thank h. arenhvel , h. schmieden and p. wilhelm for fruitful discussions and comments .
this work was supported by the deutsche forschungsgemeinschaft ( sfb 201 ) and the uk engineering and physical sciences research council . at the @xmath27 resonance position , where the phase @xmath265 passes through @xmath212 ( @xmath295 ) we find @xmath296 , @xmath297 and negligible contributions from higher partial waves ( @xmath298 ) ( see sect .
the ratio @xmath299 of the coefficients @xmath300 and @xmath301 for @xmath0 can then be expressed by @xmath302 neglecting @xmath303 , @xmath304 and all terms with @xmath305 results in @xmath306 dividing by @xmath307 we obtain @xmath308 with @xmath309 if we further neglect @xmath310 and the isospin @xmath136 components @xmath311 and @xmath312 , then the largest correction comes from @xmath312 , which is of the order of @xmath313 of @xmath314 .
the final result depends on the ratio of @xmath315 .
if we further assume @xmath316 we obtain : \(c ) for @xmath321 @xmath322 with @xmath323 without any correction for the isospin @xmath136 contributions to @xmath169 and @xmath151 the maximum systematic error of @xmath38 would be @xmath238 absolute . to further reduce this systematic error
we can estimate the size of the isospin @xmath136 contribution by the born contribution . a conservative estimate of the remaining systematic error is less than @xmath239 absolute .
99 a. rujula , h. georgi and s.l .
glashow , phys .
* d12 * 147 ( 1975 ) . c. becchi and g. morpurgo , phys .
* 17 * 352 ( 1965 ) .
r. koniuk and n. isgur , phys . rev . *
d21 * 1868 ( 1980 ) .
gershteyn et al . , sov .
. phys . * 34 * 870 ( 1981 ) .
d. drechsel and m.m .
giannini , phys .
b143 * 329 ( 1984 ) .
s. capstick , phys . rev . *
d46 * 2864 ( 1992 ) .
a. wirzba and w. weise , phys
b188 * 6 ( 1987 ) .
k. bermuth et al . , phys . rev .
* d37 * 89 ( 1988 ) .
leinweber , proc .
`` baryons92 '' page 29 ( 1992 ) .
a. buchmann et al .
, phys . rev .
* c55 * 448 ( 1997 ) .
r. beck and h .-
krahn et al . , phys .
lett * 78 * 606 ( 1997 ) .
g. s. blanpied et al .
lett . * 79 * 4337 ( 1997 ) .
davidson and n.c .
mukhopadhyay , phys .
lett * 79 * 4509 ( 1997 ) .
g. keaton and r.l .
workman , phys .
lett * 79 * 4511 ( 1997 ) .
o. hanstein et al . , phys . lett . * b385 * 45 ( 1996 ) .
r. crawford et al . , nucl . phys .
* a603 * 303 ( 1996 ) .
d. lohmann and j. peise et al . , nucl .
instr . and meth .
* a343 * 494 ( 1994 ) .
a. schmidt , diplomarbeit , university mainz 1995 .
h. herminghaus , k.h .
kaiser and h. euteneuer , nucl .
instr . and meth .
* a138 * 1 ( 1976 ) . i. anthony et al . , nucl .
instr . and meth .
* a301 * 230 ( 1991 ) . s. hall et al . , nucl . instr . and meth .
* a368 * 698 ( 1996 ) .
f. rambo et al . , phys . rev .
* c 58 * 489 ( 1998 ) .
a. kraus et al .
lett * 79 * 1834 ( 1997 ) .
g. audit et al . , nucl .
instr . and meth .
* a301 * 473 ( 1991 ) .
a. braghieri et al . ,
instr . and meth .
* a343 * 623 ( 1994 ) .
krahn , dr . thesis , university mainz ( 1996 ) .
o. hanstein , d. drechsel and l. tiator , nucl . phys .
a632 * 561 ( 1998 ) . m. jacob and g.c .
wick , ann .
* 7 * ( 1959 ) 404 .
chew , m.i .
goldberger , f.e .
low and v. nambu , phys .
* 106 * 1345 ( 1957 ) .
raskin aand t.w .
donnelly , ann .
phys . * 191 * 78 ( 1989 ) .
d. drechsel and l. tiator , j. phys .
g * 18 * 449 ( 1992 ) .
i. barker et al . , nucl . phys . *
b95 * 347 ( 1975 ) .
w. chiang and f. tabakin , phys . rev .
* c55 * 2054 ( 1997 ) .
r. workman , phys . rev .
* c53 * 1434 ( 1996 ) .
v. grushin in a.a .
komar ( ed ) , photoproduction of pions on nucleons and nuclei , proceedings of the lebedev physics institute academy of science of the ussr , volume 186 , nova science publishers , new york and budapest 1989 .
getman et al .
, nucl . phys .
* b188 * 397 ( 1981 ) , a. belyaev et al .
, nucl . phys .
* b213 * 201 ( 1983 ) .
watson , phys . rev . * 182 * , 228 ( 1954 ) .
a. donnachie , rep .
. phys . * 36 * 695 ( 1973 ) .
r. arndt , i. strakovsky and r. workman , phys .
c53 * 430 ( 1996 ) .
g. dagostini , nucl .
* a346 * 306 ( 1994 ) + the same method is used in the legs analysis .
r. beck and h .-
krahn , phys .
lett * 79 * 4512 ( 1997 ) .
workman , nucl
davidson , private communication(1998 ) .
r. m. davidson et al . , phys . rev .
* c 59 * 1059 ( 1999 ) .
r. beck and h.p .
krahn , phys .
lett * 79 * 4510 ( 1997 ) .
m. fuchs et al . , phys . lett . * b368 * 20 ( 1996 ) .
f. hrter , dr . thesis , university mainz ( 1996 ) .
h. dutz et al . , nucl . phys . *
a601 * 319 ( 1996 ) .
d. menze et al .
, compilation of pion photoproduction data , bonn ( 1977 ) .
k. buechler et al . , nucl . phys . *
a570 * 580 ( 1994 ) .
f. carbonara et al .
, nuovo cim .
* 13a * 59 ( 1973 ) .
a. bagheri et al .
, phys . rev .
* c38 * 875 ( 1988 ) . g. fischer et al . , z. physik * 253 * 38 ( 1972 )
. h. genzel et al . , z. physik * 268 * 43 ( 1974 )
. g. fischer et al . , z. physik * 245 * 225 ( 1971 ) .
e. hilger et al . , z. physik * 268 * 19 ( 1974 ) .
r. leukel , dr . thesis , university mainz ( 1999 ) .
a.m. bernstein et al .
, phys . rev .
* c47 * 1274 ( 1993 ) . t. sato and t .- s .
lee , phys .
* c54 * 2260 ( 1996 ) .
p. wilhelm et al .
, phys . rev .
* c54 * 1423 ( 1996 ) .
wilbois et al . , phys . rev .
* c57 * 295 ( 1998 ) .
a.m. sandorfi , few body syst .
99 1 ( 1998 ) .
.imaginary parts of the multipoles @xmath4 and @xmath3 in units of @xmath324 , and the ratio @xmath325 in the energy dependent dispersion analysis .
the different rows correspond to differential cross section data from mami @xcite , legs @xcite and bonn @xcite . [
cols="^,^,^,^",options="header " , ] | tagged linearly polarized photons have been used at the mainz microtron mami for simultaneous measurements of the @xmath0 and @xmath1 reaction channels to study the @xmath2 transition .
the energy dependence of the magnetic dipole @xmath3 and electric quadrupole @xmath4 amplitudes have been extracted from these data in the photon energy range from @xmath5 to @xmath6 .
the @xmath7 ratio for the @xmath2 transition has been determined to be @xmath8 at the resonance position ( @xmath9 ) . |
carbon inclusion is unavoidable when growing gan layers by metal organic chemical vapor deposition ( mocvd ) technique due to several reasons : metalorganic species used as source materials , contaminants in the source gases and hydrocarbons from graphite susceptors . as a result ,
un - intentional carbon doping is present in gan layers as impurities and can assume different configurations in the crystal lattice . the amount of incorporated carbon depends on the growth temperature @xcite and pressure @xcite . even in the case of growth performed with molecular beam epitaxy ( mbe ) ,
gan samples are contaminated by carbon impurities as soon as they are removed from vacuum and various mitigation approaches are being developed .
@xcite on the other hand , intentional carbon doping is routinely used to obtain semi - insulating layer of gan ( gan : c ) that are critical for the fabrication of high electron mobility transistors ( hemts ) @xcite . at the same time
, the presence of a significant amount of carbon in the substrate , may lead to deep level traps acting as recombination centers in the band gap , which are detrimental to the hemts performance , potentially leading to current collapse and kink effect @xcite . therefore in order to understand the effect of carbon on the devices operation and improve their performance it is important to identify the physical origins of c - related deep level traps .
earlier theoretical calculations based on density functional theory ( dft ) within local density approximation ( lda ) @xcite suggested that c could exist as substitutional forms in gan , specifically c substituting n ( @xmath3 ) and c substituting ga ( @xmath4 ) .
it was also shown that @xmath3 acts as a shallow acceptor with @xmath10ev activation energy and c@xmath11 acts as a shallow donor with activation energy with @xmath12ev .
it was also suggested that self - compensation by @xmath4 and @xmath3 pins the fermi level in the middle of band gap and this explains the high resistivity of gan : c layer @xcite .
another possible form of a single interstitial impurity , c@xmath13 , was predicted to show amphoteric behavior , acting as a deep acceptor in @xmath14-type gan and as a deep donor in @xmath15-type gan @xcite .
furthermore , the activation energy for this acceptor level due to @xmath5 has predicted to be @xmath16ev . from the experimental standpoint , a significant number of studies to understand the trap levels , specifically the energy states in the band - gap , in carbon doped gan
have been performed using a variety of techniques . among these ,
photoionization spectroscopy ( ps ) @xcite , deep level transient spectroscopy ( dlts ) , including minority carrier transient spectroscopy ( mcts ) and photoinduced current transient spectroscopy ( picts ) @xcite , deep level optical spectroscopy ( dlos ) @xcite and cathodoluminescence ( cl ) measurement @xcite .
dlts , mcts and picts have been primarily employed to detect trap levels close to the band edges ( within @xmath171.0ev ) .
techniques based on optical methods such as dlos are used mainly to detect deeper trap levels , in an energy range close to the center of the band gap .
combinations of different types of techniques makes it possible to cover entire band gap region and potentially detect all existing trap levels . using ps technique , klein _
_ showed that one of two kinds of deep traps ( with absorption threshold at 2.85ev ) , which is responsible for the current collapse of algan / gan hemt , has a carbon origin because its concentration tracks the carbon doping level @xcite .
et al . _ using dlos measurement which is able to determine that a trap level located 1.35ev below the conduction band minimum ( cbm , @xmath18 ) was related to carbon @xcite .
_ investigated the origin of a number of deep level traps in gan using a combination of dlts , dlos and steady - state photocapacitance ( sspc ) techniques @xcite .
two of the energy levels obtained by dlts were ascribed to carbon .
one was located 0.11ev below the cbm , and its origin was assigned to @xmath4 , based on previous lda calculated results .
the other trap level appeared at @xmath19ev , where @xmath20 is the energy of the valence band maximum ( vbm ) , but its physical origin was unknown .
in addition , four more levels that were obtained by the combination of dlos and sspc were ascribed to carbon .
based on existing lda result @xcite , a trap energy level at @xmath21ev was assigned to an acceptor state of interstitial carbon ( @xmath5 ) .
two traps at @xmath22ev and @xmath23ev were c - related , but their physical forms remained unknown .
a trap at @xmath24ev was assigned to an acceptor level of @xmath3 , once again , based on lda result @xcite .
shah and coworkers performed dlts and mcts measurements @xcite and inferred that an energy level at @xmath25ev was possibly related to @xmath4 and behaved as an electron trap state .
furthermore , two energy levels responsible for trapping holes were observed and were also related to c. one was at @xmath26ev with @xmath3 assignment and the other at @xmath27ev with @xmath3-related defects or gallium vacancy .
et al . _ studied deep levels by photocurrent spectra measurements and picts . with the former method three c - related trap levels with optical threshold
near 1.31.4ev , 2.72.8ev and 3ev were observed .
the first was attributed to @xmath5 acceptor state , whereas the other two were left unassigned . more recent experimental result , carried out by honda _
_ using mcts measurements , indicated that a hole trapping state may be present at @xmath28ev and was assigned to @xmath3 based on recent dft calculation employing hybrid functionals @xcite .
they also concluded that one electron trap at @xmath29ev was associated with c , but did not specify its physical form .
from the theoretical standpoint , the majority of studies performed to understand the nature of carbon in gan have been carried out using dft within the lda approximation .
@xcite only recently , a small number of investigations @xcite have employed more sophisticated dft approaches based on hybrid functionals , with the intent to overcome the limitations of lda , and obtain a more reliable energetics for the various carbon configurations in gan . among these recent studies ,
dft calculations using heyd - scuseria - ernzerhof ( hse ) hybrid density functionals for single carbon impurities , @xmath3 and @xmath4 , were reported @xcite . unlike in previous lda results , these calculations indicates that @xmath3 may not be a shallow but a deep acceptor with @xmath30ev activation energy , which means that c can not be used as a @xmath15-type donor in gan .
furthermore , recent hse - based calculations , indicates that the acceptor level of @xmath5 is @xmath31ev as opposed to the values of @xmath32ev obtained using lda . @xcite .
these new outcomes , suggest that the assignments of experimentally observed trap levels based on earlier lda results should be re - examined using more reliable approaches , such as hse hybrid density functionals , that could provide a more accurate picture of the system energetics .
the aim of this work is twofold , first , we intend to perform a comprehensive study of the formation energies of single carbon impurity and complexes in gan using state - of - the - art dft and hse hybrid density functionals .
specifically we intend to focus on the role of complexes about which very little is known .
second , we intend to perform a systematic comparison of the numerical results with the available experimental data with two specific goals in mind : establish which energy level can be reliably assigned to a given configuration , and for which energy level additional experimental and theoretical work is needed .
we want to emphasize this last issue , since there are several experimentally observed c - related energy levels , whose physical forms are still unknown .
this manuscript focuses on the complexes that carbon forms with ga , n and their vacancies .
furthermore , this work is a companion to a second manuscript@xcite in which we present the investigation of complexes that carbon forms in gan with silicon , oxygen and hydrogen and we discuss their relative concentrations .
this paper is organized as follows . in section
[ s : methods ] , the details of the computational model are presented and the theoretical formulations of the formation energies and transition levels are provided . furthermore we describe what kind of convergence studies have been performed to determine the supercell size to be used in the case of charged defects and complexes . in section
[ s : results ] , we will outline our calculated results that will be discussed in section [ s : discussion ] . finally concluding remarks are given in section [ s : conclusion ] .
in this section the computational approach is outlined .
first the computational framework is presented .
subsequently the model used to evaluate the formation energy is given , and finally the effect of the supercell size on formation energy is discussed .
the calculations presented in this work were carried out using the projector augmented wave ( paw ) method @xcite implemented in the vasp code @xcite .
the main results were obtained using heyd - scuseria - ernzerhof ( hse ) hybrid functionals @xcite .
additionally , convergence tests were also performed using standard perdew - burke - ernzerhof ( pbe ) @xcite functionals .
the semicore ga @xmath33 electrons were included as valence , since treating ga @xmath33 electrons as core caused relatively large errors ( @xmath34ev ) in the formation energies for carbon complexes .
furthermore , in the case of hse , in order to reproduce experimental band gap value ( 3.5ev @xcite ) , the amount of exact exchange was taken to be 28% ( giving 3.45ev band gap value ) .
finally , a 425ev cutoff energy was used . during the structural optimization procedure to obtain the total energy of different configurations , the atomic positions were allowed to change until the largest force component was less than 0.05ev / .
the bulk parameters obtained from fully optimized unit cell are summarized in table [ tab : bulkpara ] . using these optimized lattice constants ,
a supercell containing total 96 atoms with orthorhombic shape was constructed and employed to study the carbon inclusions .
the brillouin zone is sampled on a mesh composed of a @xmath35 @xmath36-point grid .
convergence tests in pbe showed that the differences in formation energies between @xmath35 mesh and denser @xmath37 mesh were less than 50mev .
the same tests were also done in hse , where @xmath35 mesh and @xmath38 mesh were compared .
the differences in formation energies were less than 0.1ev . in fig .
[ f : kconv ] , the formation energies ( its definition will be provided in the next section ) for @xmath6 in ga - rich condition are presented for different @xmath36-point mesh both in pbe and hse to show the convergence .
convergence tests performed to evaluate the dependence of the calculated values on the supercell size will be discussed in section[s : supcell ] .
.calculated values for lattice constants , energy gap , formation enthalpy and static dielectric constant .
experimental results are also given.[tab : bulkpara ] [ cols="<,<,<,<,<,<,<",options="header " , ] we consider first the trap level located at @xmath39ev obtained by armstrong _
_ and that was also observed by polyakov__et al . _ _ as well as by honda__et al.__. while armstrong__et al . _ _ and polyakov__et al . _
_ did not give a clear assignment to this trap level , honda__et al . _ _
@xcite assigned it to @xmath3 ( 0/@xmath40 ) by comparing the experimental value to the calculated result obtained using hse by lyons__et al . _ _
indeed our calculated results support this assignment . in our case ,
the @xmath3 ( 0/@xmath40 ) level is located at 0.89ev above valence band maximum with 2.91ev optical activation energy ( @xmath41 ) as shown in table [ t : calcth1 ] .
furthermore , the thermal activation energy of this trap is calculated to be @xmath42=2.56ev .
consequently we argue that the trap level at @xmath43ev , originally measured by shah__et al . _ _
@xcite and that was considered to be related to @xmath3 complex or gallium vacancy , is in reality due to @xmath3 ( 0/@xmath40 ) .
we turn our attention to the trap levels originally assigned to @xmath3 ( 0/@xmath40 ) , specifically the one located at @xmath44ev measured by armstrong__et al . _ _ and @xmath45ev level by shah _
et al._. since shah__et al .
_ _ measured this energy with mcts we assume that this is a thermal activation energy . from tables [ t : calcth1 ] and [ t : calcth2 ] we can observe that the energies corresponding to ( + /0 ) of @xmath3 gives the value of 3.20ev . on the other hand ,
we notice that the trap level observed by armstrong__et al . _ _ with dlos should be treated as an optical activation energy . in this case
we observe that the ( 4+/3 + ) of @xmath5 ( 3.11ev ) as well as ( + /0 ) of @xmath8 ( 3.21ev ) are calculated optical trap levels , which have the energy very close to experimental one ( 3.28ev ) .
the next one is the trap level at @xmath46ev observed by armstrong__et al .
_ _ and polyakov__et al .
_ _ that was assigned to @xmath5 ( 0/@xmath47 ) based on a 1.13ev thermal activation energy computed by lda @xcite .
first we point out that , based on our hse calculation , the ( 0/@xmath47 ) state is energetically unfavorable while the ( 0/@xmath40 ) , that was found to be unfavorable with lda is now possible . for this ( 0/@xmath40 ) state we compute an optical activation energy of 0.82ev that corresponds to a thermal activation energy of 0.25ev
. consequently , @xmath5 ( 0/@xmath40 ) can not be the origin of the level @xmath48ev observed by armstrong__et al.__. additionally , we notice that the calculated ( + /0 ) @xmath5 level from hse results in a optical activation energy of 1.58ev , corresponding to a thermal activation energy of 0.87ev . therefore based on our results , the origin of @xmath46ev level
is not likely to be related to any of the @xmath5 states .
finally , based on hse the ( 0/@xmath40 ) state of the @xmath49 complex has a 1.31ev optical activation energy . as a result , this is likely to be the origin of the experimentally observed @xmath46ev level .
note that the ( @xmath47/@xmath50 ) state of the @xmath51 complex has @xmath52ev , but this complex is unstable in this energy region with negative binding energy .
_ observed , using dlts , a @xmath53ev level and concluded that it was c - related , but did not mention its origin .
based on our calculations the ( 0/@xmath40 ) level of @xmath5 has a 0.25ev thermal activation energy .
furthermore , the theoretical values of the activation energy for the ( + /0 ) level of @xmath9 is 0.22ev .
therefore , it is likely that the origin of the @xmath53ev level is the ( 0/@xmath40 ) state of @xmath5 and/or the ( + /0 ) state of @xmath9 . the @xmath54ev level observed with dlos by armstrong _
et al . _
was assumed to be c - related but its physical form remained unknown . from our calculated results
two levels have similar optical activation energies .
one is ( 2+/+ ) level of @xmath49 with 1.96ev and the other is ( 0/@xmath40 ) level of @xmath8 with 2.08ev .
these two states are the likely candidates to explain the physical origin of the @xmath55ev trap level .
polyakov _ et al . _ observed a optical threshold at 2.72.8ev by photocurrent spectra measurement .
the origin of this trap level is likely to be the ( 2+/+ ) level of @xmath5 , whose optical activation energy is computed to be 2.87ev .
the last two energy levels we consider are the one located at @xmath56ev measured by armstrong__et al . _
_ and at @xmath57ev measured by shah__et al.__. these two levels were assigned to @xmath4 based on lda results @xcite , for which the ( + /0 ) transition level was computed to be about 0.2ev below cbm .
however , our hse calculation shows that such ( + /0 ) level of @xmath4 does not appear within the band gap , but it is in the conduction band approximately 0.5ev above cbm .
recent hse calculation by lyons__et al . _ _
@xcite does not show such level within the band gap either .
from our results we evince that the ( + /0 ) level of @xmath58 complex has a 0.24ev thermal activation energy .
unfortunately this complex is unlikely to form when the fermi energy is around this value because its biding energy is barely positive as we can see from fig .
[ f : eb ] . based on our calculations ,
the most likely candidate is the ( + /0 ) state of @xmath9 located at 0.22ev below the cbm and/or the ( 0/@xmath40 ) state of @xmath5 at 0.25ev below the cbm .
therefore , it is probable that these levels have the same origin of the @xmath59ev level observed by honda
we have performed first - principles calculations using hse hybrid density functional in the framework of dft to investigate the characteristics of various forms of carbon inclusions in gan .
we have considered single impurities , @xmath3 , @xmath4 and @xmath5 , as well as their complexes @xmath0 , @xmath1 , @xmath2 , @xmath8 and @xmath9 . for all these configurations , different charge states have been considered and their geometries are fully optimized .
formation and binding energies of complexes have been computed and thermodynamic transition levels are obtained . among single impurities , @xmath3 behaves mainly as a deep acceptor , @xmath4 acts as a donor without inducing states in the band gap and @xmath5 shows amphoteric behavior .
both in n - rich and ga - rich conditions the 4 + charge state of @xmath5 , which assumes an octahedral interstitial position , is favorable close to the vbm ( @xmath15-type region ) .
the @xmath60 charge state of @xmath3 is favorable close to the cbm ( @xmath14-type region ) , and @xmath4 is also favorable in @xmath15-type region but only in n - rich conditions
. complexes made of combinations of single impurities , specifically @xmath0 , @xmath1 and @xmath2 , have also been considered .
@xmath0 is favorable in the upper half of the band gap ( @xmath14-type ) region , whereas @xmath1 and @xmath2 are favorable in the lower half of the band gap ( @xmath15-type ) region .
finally , complexes with vacancies are also examined , in particular @xmath8 and @xmath9 .
the former is favorable with lower formation energy comparing to the latter close the cbm , but is unstable as a complex in the @xmath14-type region with negative value of binding energy .
the latter is favorable in the @xmath15-type region comparing to the former . from the calculated formation energies we have evaluated the thermodynamic transition levels .
these are directly related to the thermal activation energies observed in experimental techniques such as dlts .
in addition , by calculating franck - condon shift , optical activation energies , which can be obtained by optical techniques such as dlos , have been evaluated from the thermal activation energies .
we compare our calculated values of activation energies with the energies of experimentally observed c - related trap levels .
the trap level observed at @xmath61ev is likely due to the ( 0/@xmath40 ) level of @xmath3 .
based on earlier investigations performed employing lda , to different energy levels : @xmath62 ev level by mcts ( thermal method ) and @xmath63ev by dlos ( optical method ) , were attributed to the ( 0/@xmath40 ) level of @xmath3 .
however , our hse calculation shows that the origin of the @xmath62ev level is the ( + /0 ) level of @xmath3 and of the level at @xmath63ev is the ( 4+/3 + ) of @xmath5 and/or the ( + /0 ) of @xmath8 . based on lda results , the trap level observed at @xmath46ev
was unanimously attributed to the ( 0/2@xmath40 ) state of @xmath5 .
however , the outcome of our hse calculation suggests its origin is the ( 0/@xmath40 ) state of the @xmath2 complex .
the trap at @xmath64ev is likely due to two configurations .
one is the ( 2+/+ ) of @xmath2 and the other is the ( 0/@xmath40 ) of @xmath8 .
looking at the trap level located at @xmath65ev we argue that its origin stems from the ( 2+/+ ) state of the @xmath5 .
multiple contributors are likely to be responsible for the trap at @xmath66ev . specifically we find that the ( 0/@xmath40 ) of @xmath5 and the ( + /0 ) of @xmath9 have energies that are close to the measured value . finally , the trap at @xmath67ev was attributed to @xmath4 based on lda results . however , based on our hse calculations , there are no gap states due to @xmath4
. therefore the origin of this trap level is still unclear , but the most likely candidate is the ( + /0 ) of @xmath9 and/or the ( 0/@xmath40 ) of @xmath5 . in this case
these levels have same origin as @xmath68ev level .
the authors would like to thank k. jones and r. tompkins of the army research laboratory and t. moustakas of boston university for many discussions and their help in understanding the experimental techniques .
the authors are grateful to a. f. wright , s. lee and n. modine from sandia national laboratory for discussing the results of our work .
the authors gratefully acknowledge financial support from the u. s. army research laboratory through the collaborative research alliance ( cra ) for multiscale multidisciplinary modeling of electronic materials ( msme ) .
this work was performed using dod hpcmp supercomputing resources and computational resources provided by the 2014 aro durip award made to e. bellotti .
47ifxundefined [ 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 \doibase http://dx.doi.org/10.1016/s0022-0248(02)01348-9 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.jcrysgro.2003.08.024 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.368814 [ * * , ( ) ] \doibase http://dx.doi.org/10.1116/1.1503782 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3285309 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.124252 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.125523 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3459968 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.117934 [ * * , ( ) ] link:\doibase 10.1103/physrevb.56.9496 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.1498879 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1518794 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1418452 [ * * , ( ) ]
\doibase http://dx.doi.org/10.1063/1.126580 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1643540 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2005379 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3488610 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1016/j.sse.2012.05.057 [ * * , ( ) ] \doibase http://dx.doi.org/10.1116/1.4820905 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2179375 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3492841 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.087404 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.035204 [ * * , ( ) ]
@noop link:\doibase 10.1103/physrevb.50.17953 [ * * , ( ) ] link:\doibase 10.1103/physrevb.54.11169 [ * * , ( ) ] link:\doibase 10.1103/physrevb.59.1758 [ * * , ( ) ] link:\doibase 10.1063/1.1564060 [ * * , ( ) ] link:\doibase 10.1063/1.2204597 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.77.3865 [ * * , ( ) ] link:\doibase 10.1103/physrevb.10.676 [ * * , ( ) ] link:\doibase 10.1103/physrevb.48.4335 [ * * , ( ) ] link:\doibase 10.1021/jp993752s [ * * , ( ) ] link:\doibase 10.1002/zaac.19402440203 [ * * , ( ) ] link:\doibase 10.1103/physrevb.7.743 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.016402 [ * * , ( ) ] link:\doibase 10.1002/pssb.201046289 [ * * , ( ) ] link:\doibase 10.1103/physrevb.51.4014 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.235104 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.84.075155 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.045112 [ * * , ( ) ] link:\doibase 10.1557/jmr.2011.360 [ * * , ( ) ] | various forms of carbon based complexes in gan are studied with first - principles calculations employing heyd - scuseria - ernzerhof hybrid functional within the framework of density functional theory .
we consider carbon complexes made of the combinations of single impurities , i.e. @xmath0 , @xmath1 and @xmath2 , where @xmath3 , @xmath4 and @xmath5 denote c substituting nitrogen , c substituting gallium and interstitial c , respectively , and of neighboring gallium / nitrogen vacancies ( @xmath6/@xmath7 ) , i.e. @xmath8 and @xmath9 .
formation energies are computed for all these configurations with different charge states after full geometry optimizations . from our calculated formation energies ,
thermodynamic transition levels are evaluated , which are related to the thermal activation energies observed in experimental techniques such as deep level transient spectroscopy .
furthermore , the lattice relaxation energies ( franck - condon shift ) are computed to obtain optical activation energies , which are observed in experimental techniques such as deep level optical spectroscopy .
we compare our calculated values of activation energies with the energies of experimentally observed c - related trap levels and identify the physical origins of these traps , which were unknown before . |
identification of an intriguing collective dynamical state , namely the chimera state , in an ensemble of coupled identical nonlinear oscillators with nonlocal coupling @xcite has initiated intense research activities in the recent literature @xcite .
a chimera state represents a spatially inhomogeneous state characterized by coexisting coherent and incoherent domains in an ensemble of identical oscillators .
experimentally , chimera has also been revealed in populations of coupled chemical oscillators @xcite , in electro - optical systems @xcite and in metronomes @xcite .
real world examples mimicking chimera states can be found in power grids @xcite , in unihemispheric sleep of animals @xcite , in multiple time scales of sleep dynamics @xcite , etc .
different types of chimera states such as amplitude mediated chimera @xcite , intensity induced chimera @xcite , and so on have also been identified @xcite .
recently , it has been shown that a symmetry breaking coupling in the stuart - landau oscillators leads to the manifestation of chimera death @xcite .
the collective state with inhomogeneous flipping between the steady states ( oscillation death ) of an ensemble was called chimera death . very recently , noise induced coherence - resonance chimeras in a network of excitable elements was reported @xcite .
coherence - resonance chimeras are associated with alternating switching of the location of coherent and incoherent domains . in this manuscript
, we unravel a novel dynamical regime emerging from the inhomogeneous synchronized states .
in particular , we consider an ensemble of identical nonlinear oscillators coupled via a common dynamic environment with nonlocal coupling .
we show that the ensemble of oscillators splits into coexisting coherent and incoherent domains for appropriate strength of the nonlocal coupling .
the nearby oscillators in the coherent domains exhibit in - phase synchronized oscillations while the adjacent oscillators in the incoherent domains exhibit out - of - phase oscillations .
it is to be noted that the term coherent / incoherent domains here represents the homogenous / inhomogenous nature of the dynamics of the attractors in the corresponding domain and does not refer to the nature of the attractors , that is whether they are coherent attractors with a fixed center of rotation or incoherent attractors with more than one center of rotation .
further , we find that some of the nearby coherent domains exhibit in - phase oscillations in one of the domains and anti - phase oscillations in the other resembling the phase - flip bifurcation / transition @xcite . specifically , the out - of - phase synchronized nearby coherent domains are interspersed by an incoherent domain ( where the phases of the adjacent oscillators flip between 0 and @xmath0 ) at the phase - flip transition , which we call a phase - flip chimera .
it is a dynamically active emerging behavior in contrast to the chimera death @xcite , where nearby oscillators populate the same branch of the inhomogeneous steady state in the coherent domain while the nearby oscillators populate different branches of the inhomogeneous steady state in the incoherent domain .
further , we find that the conventional chimera is preceded by the phase - flip chimera in addition to the other collective dynamical regimes such as coherent and synchronized states in an ensemble of the paradigmatic rssler oscillators both in the periodic and chaotic regimes .
we have used the measures , namely the strength of incoherence and the probability distribution of the correlation coefficient to characterize the phase - flip chimeras and the observed dynamical transitions .
in addition , we have also employed the framework of the master - stability function ( msf ) to demarcate the synchronized and desynchronized parameter space which agrees very well with the simulation results .
it is to be pointed out that the synchronized regime is a multistability regime coexisting with conventional chimeras and phase - flip chimeras depending on the distribution of the initial conditions .
the plan of the paper is as follows .
we discuss the emergence of phase - flip chimera in an ensemble of rssler oscillators in the periodic regime in sec .
ii and in the chaotic regime in sec .
we provide certain quantification measures such as the strength of incoherence , probability distribution of the correlation coefficient and the framework of the master stability functon to characterize the different collective dynamical behavior in the ensemble of oscillators in sec .
iv . discussion on the global dynamical behavior in terms of two - parameter phase diagrams is provided in sec .
v. emergence of phase - flip chimera with global coupling among the agents in the common environment will be discussed in sec .
vi and conclusion will be provided in sec .
vii . in appendix
a , we consider the effect of relaxation time while the effect of coupling to other variables is discussed in appendix b.
to elucidate the above results , we consider an ensemble of identical rssler oscillators with a common dynamic environmental coupling represented as @xmath1 [ ross ] @xmath2 , where @xmath3 , @xmath4 , @xmath5 and @xmath6 are the system parameters , @xmath7 is the number of oscillators in the ensemble .
the oscillators in the ensemble are coupled indirectly via nonlocally coupled dynamic agents @xmath8 in the common environment .
@xmath9 is the strength with which the agent @xmath10 from the common environment interacts with the @xmath11th oscillator in the ensemble .
@xmath12 is the strength of the nonlocal coupling . @xmath13
$ ] is the number of nearest neighbors on each side of any oscillator in the ring with a coupling radius @xmath14 .
the environment / medium plays a crucial role in facilitating the complex collective dynamics such as decoherence , dissipation and relaxation in quantum systems @xcite , in coordinated rhythms in biological systems @xcite , and in quorum sensing @xcite .
the dynamics of the individual agent given by @xmath15 in eq .
( [ ross]d ) is related to the interactions of molecules between the cells and their environment @xcite . in the following
, we will demonstrate the existence of phase - flip chimera in an ensemble of rssler oscillators in both periodic and chaotic regimes .
( g ) to clearly show the phase - flip chimera characterized by coexisting coherent and incoherent domains , where the nearby oscillators exhibit in - phase oscillations in the coherent domain while the adjacent oscillators exhibit out - of - phase oscillations in the incoherent domain . ] individual rssler oscillators in the ensemble exhibit periodic oscillations for the parameters @xmath16 , @xmath17 and @xmath18 .
we have used random initial conditions uniformly distributed between @xmath19 to @xmath20 .
we have fixed @xmath21 , @xmath22 , and the coupling radius @xmath23 as @xmath24 ( @xmath21 is chosen only for clarity in the sense that the jumping phases can be distinctly seen .
one can indeed choose any value of @xmath7 for our analysis ) .
snapshots of the instantaneous phases @xmath25 , @xmath26 , and the spatiotemporal evolution of the oscillators are depicted in fig .
[ pre_fig1 ] for different values of the strength of the nonlocal coupling @xmath12 .
the oscillators evolve in asynchrony for @xmath27 ( see figs . [
pre_fig1](a ) and [ pre_fig1](b ) ) , while their time averaged frequencies are entrained ( see the inset of fig .
[ pre_fig1](a ) ) .
being identical rssler oscillators in the periodic regime , the frequencies of all the oscillators are always entrained ( see the insets of fig . [ pre_fig1 ] ) in the entire parameter regimes we have traced .
the nonlocal coupling leads to the splitting of the ensemble into coexisting coherent and incoherent domains as depicted in figs .
[ pre_fig1](c ) and [ pre_fig1](d ) for @xmath28 , thereby confirming the existence of chimera state
. further increase in the strength of the nonlocal coupling results in the synchronous evolution of the ensemble of rssler oscillators .
phase and complete synchronous evolution of the oscillators are clearly evident from figs .
[ pre_fig1](e ) and [ pre_fig1](f ) , respectively , for @xmath29 .
phase - flip chimera emerges after the complete synchronized state , which is illustrated in figs .
[ pre_fig1](g ) and [ pre_fig1](h ) for @xmath30 .
phase - flip chimeras with two adjacent out - of - phase synchronized spatially coherent domains interspersed by a spatially incoherent domain are clearly evident from the spatiotemporal plot in fig .
[ pre_fig1](h ) , while their frequencies remain entrained .
however , for the ease of visualization of the phase - flip chimera characterized by coexisting out - of - phase synchronized coherent domains ( where the nearby oscillators exhibit in - phase synchronized oscillations among them in each of the synchronized domains ) along with an asynchronous incoherent domain ( where nearby oscillators exhibit out - of - phase synchronized oscillations ) interspersing the coherent domains at the phase - flip transition are enlarged and depicted in fig .
[ pre_fig2 ] .
it is to be noted that the chimera states investigated so far have emerged as an intermediate state in the transition from completely incoherent to completely coherent states , in general . in contrast , the phase - flip chimera emerges after the completely synchronized state .
note that these are not steady states as in the case of chimera death states @xcite , but evolve dynamically , and they are also not two - cluster states as the oscillator index can not be reordered .
for the parameters @xmath31 , @xmath32 and @xmath33 , the uncoupled rssler oscillators evolve in ( [ ross ] ) chaotically .
snapshots of the instantaneous phases and the spatiotemporal evolution of the oscillators are depicted in fig .
[ pre_fig3 ] for different values of coupling , @xmath12 .
the time averaged frequencies of all the oscillators are shown in the insets of fig . [ pre_fig3 ] .
one can observe that frequencies of the oscillators are entrained to the same frequency as indicated by a straight line of @xmath8 .
for lower values of @xmath12 , the coupled oscillators evolve in asynchrony ( not shown here ) , whereas the ensemble of oscillators splits into coexisting coherent and incoherent domains , confirming the existence of chimera , as shown in figs .
[ pre_fig3](a ) and [ pre_fig3](b ) for @xmath29 . increasing @xmath12
further , the ensemble of oscillators evolve in complete synchrony ( see figs . [
pre_fig3](c ) and [ pre_fig3](d ) for @xmath34 ) .
the synchronized oscillators become desynchronized for further larger @xmath12 , the dynamics of which is illustrated in the snapshot of the instantaneous phases and the spatiotemporal plots in figs . [
pre_fig3](e ) and [ pre_fig3](f ) , respectively , for @xmath35 .
phase - flip chimera emerges from the desynchronized state upon increasing the strength of the nonlocal coupling further as depicted in figs .
[ pre_fig3](g ) and [ pre_fig3](h ) for @xmath36 . as a function of the strength of the nonlocal coupling @xmath12 .
( a ) desynchronized state for @xmath37 , ( b ) chimera state for @xmath29 , ( c ) synchronized state for @xmath34 , and ( d ) desynchronized state for @xmath35 . in fig ( a ) the value of @xmath38 , while in ( c)-(d ) it is fixed as 10 . ]
the dynamics of the individual rssler oscillators changes and plays a crucial role in the emergence of the observed collective dynamical behaviors of the ensemble of rssler oscillators in fig .
[ pre_fig3 ] as a function of @xmath12 .
the rssler oscillators always exhibit periodic oscillations throughout the entire dynamical transition regimes of the ensemble discussed in fig .
[ pre_fig1 ] , whereas in the case of chaotic oscillations of the individual oscillators the dynamical nature of the oscillators changes as will be discussed below . in the absence of the couplings @xmath38 and @xmath39 ,
the uncoupled individual rssler oscillators exhibits chaotic oscillations as pointed above . a couple of uncoupled representative rssler oscillators ( namely @xmath40 and @xmath41 ) exhibiting chaotic oscillations are depicted in fig . [ pre_fig4](a ) .
the filled circles and triangles in fig .
[ pre_fig4 ] are the poincar points .
the value of @xmath9 in the other figures [ pre_fig4](b)-(d ) is fixed as @xmath22 as in fig .
[ pre_fig3 ] .
the ensemble of rssler oscillators displays chimera state for @xmath29 ( see fig . [ pre_fig3](a ) ) .
a representative oscillator from each of the coherent and incoherent domains is displayed in fig .
[ pre_fig4](b ) .
the synchronized oscillators in the coherent domain are entrained to periodic oscillations as indicated by the blue ( light grey ) line in fig .
[ pre_fig4](b ) , while the asynchronous oscillators in the incoherent domain exhibit chaotic oscillations as indicated by the black ( dark grey ) line in fig . [ pre_fig4](b ) .
the ensemble of oscillators are synchronized for the strength of the nonlocal coupling @xmath34 ( see fig .
[ pre_fig3](c ) ) , where the dynamics of the ensemble of oscillators become periodic as indicated by the representative oscillators in fig .
[ pre_fig4](c ) .
the ensemble of rssler oscillators are desynchronized as already shown in fig .
[ pre_fig3](e ) for @xmath35 rendering the oscillators to exhibit chaotic oscillations as displayed in fig .
[ pre_fig4](d ) and the dynamics of the individual rssler oscillators remain chaotic for further larger values of @xmath12 .
( g ) and [ pre_fig3](g ) .
top row represents the rssler oscillators in the periodic regime and bottom row corresponds to the rssler oscillators exhibiting chaotic oscillations . ] attractors and the corresponding time series plots of the representative oscillators from the out - of - phase synchronized coherent domains of the phase - flip chimera are depicted in fig .
[ pre_fig5 ] .
the top row represents the rssler oscillators in the periodic regime and the bottom row corresponds to the rssler oscillators exhibiting chaotic oscillations .
attractors exhibting both in - phase and anti - phase oscillations for the periodic case and chaotic case along with their poincar points are shown in figs .
[ pre_fig5](a ) and [ pre_fig5](c ) , respectively .
time series plots in figs .
[ pre_fig5](b ) and [ pre_fig5](d ) clearly displays the out - of - phase oscillations of the two adjacent coherent domains of the phase - flip chimera .
the notion of the strength of incoherence @xmath42
@xcite , was recently introduced by gopal et al to characterize and to distinguish various collective dynamical states , defined as @xmath43 where @xmath44 is the heaviside step function , and @xmath45 is a predefined threshold .
normally , @xmath45 is chosen as a certain percentage value of the difference between the upper / lower bounds , @xmath46 , of the allowed values of @xmath47 .
@xmath48 is the number of bins of equal size @xmath49 .
the local standard deviation @xmath50 is introduced as @xmath51 ^ 2 } \hspace{0.1cm}\big>_{t } , \hspace{0.1cm}m=1,2, ...
m,\ ] ] where @xmath52 , @xmath53 , @xmath54 is the dimension of the individual unit in the ensemble , @xmath26 , @xmath55 , and @xmath56 denotes the time average . when @xmath50 is less than @xmath45 , @xmath57 , otherwise @xmath58 ( @xmath59 in the present case is chosen as @xmath60 ) .
the local standard deviation @xmath50 has some finite value in the incoherent domain @xmath61 , which is always greater than @xmath45 and hence @xmath62 , thereby resulting in unit value for the strength of incoherence @xmath42 in the incoherent domain . on the other hand , the standard deviation @xmath50 is always zero in the coherent domain and hence @xmath63 , thereby resulting in the null value of @xmath42 .
since the chimera states are characterized by coexisting coherent and incoherent domains , the strength of incoherence @xmath42 will have intermediate values between zero and one , @xmath64 .
as a function of the strength of the nonlocal coupling @xmath12 in ( a ) periodic case , ( b ) chaotic case . the probability distribution pr@xmath65 of the correlation coefficient ( c ) conventional chimera in the periodic case ( d ) phase - flip chimera in the periodic case , ( e ) conventional chimera in the chaotic case and ( f ) phase - flip chimera in the chaotic case . ]
the strength of incoherence is shown in figs .
[ pre_fig6](a ) and [ pre_fig6](b ) as a function of the strength of the nonlocal coupling @xmath12 characterizing the dynamical transition discussed in figs .
[ pre_fig1 ] and [ pre_fig3 ] , respectively .
unit value of @xmath42 in the range of @xmath66 in fig .
[ pre_fig6](a ) corroborates the asynchronous evolution of the rssler oscillators with periodic oscillations .
intermediate value of @xmath42 between zero and unity confirms the existence of conventional chimera in the range of @xmath67 .
null value of @xmath42 in the range of @xmath68 attributes to the synchronous evolution of the ensemble of rssler oscillators .
fluctuations in the value of @xmath42 close to unity elucidates the existence of phase - flip chimera state .
as the phase - flip chimera is characterized by out - of - phase synchronized coherent domains interspersed by an incoherent domain , most of the bins during this state have a mixture of in - phase and anti - phase oscillations , while a few bins may have completely coherent ( either in - phase or anti - phase ) oscillations leading to fluctuating values of @xmath42 close to unity .
now , we will discuss the dynamical transtion observed in fig .
[ pre_fig3 ] for the chaotic oscillations of the individual rssler oscillators in terms of the strength of incoherence @xmath42 displayed in fig .
[ pre_fig6](b ) .
asynchronous evolution of the oscillators are indicated by the unit value of @xmath42 in the range of @xmath69 , whereas the intermediate value of @xmath64 confirms the existence of conventional chimera in the range of @xmath70 .
narrow range of synchronized state is confirmed by @xmath71 in the range of @xmath72 .
the oscillators get desynchronized in the range of @xmath73 as indicated by a unit value of @xmath42 .
fluctuations in @xmath42 close to unity beyond @xmath74 corroborates the existence of phase - flip chimera .
we have also estimated the probability distribution of the correlation coefficient defined as @xmath75 where i=1,2 .. ,n , to characterize the conventional chimera and phase - flip chimera . in eq .
( [ prob ] ) , @xmath76 represents the time average and @xmath77 is the time shift .
the correlation coefficient is estimated by using each oscillator in the ensemble as a reference oscillator and averaging it over the number of oscillators @xmath7 .
the probability distribution of the correlation coefficient is depicted in figs .
[ pre_fig6](c)-[pre_fig6](f ) for different values of @xmath12 in both the periodic ( figs .
[ pre_fig6](c ) and [ pre_fig6](d ) ) and chaotic ( figs .
[ pre_fig6](e ) and [ pre_fig6](f ) ) regimes .
correlation coefficient for completely synchronized oscillators acquire unit value and for antisynchronous state acquires @xmath19 , whereas for the desynchronous state it is characterized by the intermediate values between @xmath78 . since the conventional chimera is characterized by the coexistence of synchronized and asynchronous domains , the probability distribution of the correlation coefficient is large near unit value and small at other values of correlation coefficient as seen in figs . [ pre_fig6](c ) and [ pre_fig6](e ) . on the other hand , the phase - flip chimera is characterized by only two inhomogeneous states ,
namely in - phase and anti - phase synchronized states , the correlation coefficient of the phase - flip chimera acquires only @xmath20 and @xmath19 as its values .
consequently , the probability distribution of the correlation coefficient at phase - flip chimera has only two values at @xmath20 and @xmath19 as evident from figs .
[ pre_fig6](d ) and [ pre_fig6](f ) . and the coupling radius @xmath23 .
( a ) periodic state ( b ) chaotic state .
the parameter spaces marked as @xmath79 , @xmath80 , @xmath81 , @xmath82 , and @xmath83 correspond to the desynchronized state , conventional chimera , phase - flip chimera , coherent state and complete synchronized state , respectively .
the dotted lines correspond to the stability curves estimated from the eigenvalues of the variational equation ( [ ross2 ] ) . ]
a two - parameter phase diagram as a function of the strength of the nonlocal coupling @xmath84 and the coupling radius @xmath85 is depicted in fig .
[ pre_fig7](a ) to gain a global perspective of collective dynamics emerging from the ensemble of rssler oscillators exhibiting periodic oscillations in a dynamic environment with nonlocal coupling .
we have used the above quantification measures discused in sec .
[ quanti ] to demarcate the different dynamical regimes of the ensemble of rssler oscillators .
the dynamical regimes delineated in fig .
[ pre_fig1 ] are indicated as @xmath79 , @xmath80 , @xmath86 and @xmath81 , which correspond to the desynchronized states , chimera states , synchronized states and phase - flip chimeras , respectively .
in addition to the above regimes , we have also found the coherent states marked as @xmath82 , where all the oscillators evolve in coherence in the range of the coupling radius @xmath87 and @xmath88 . using the well known master stability function ( msf ) formalism @xcite , the entire parameter space in fig .
[ pre_fig7 ] can be demarcated into desynchronized state ( @xmath79 ) and synchronous state ( @xmath86 ) .
the stability of the synchronized manifold @xmath89 and @xmath90 ; @xmath91 ) is determined by the variational equations @xmath92 [ ross1 ] where @xmath93 and @xmath94 are the solution of the uncoupled equation ( [ ross ] ) .
@xmath95 , where @xmath96 , and @xmath97 are the deviation of @xmath98 from the synchronized solution @xmath99 .
@xmath100 is the eigenvector of the coupling matrix @xmath101 and @xmath102 are the eigenvalues , where @xmath103 with @xmath104 and @xmath105 for @xmath106 and @xmath107 . here
@xmath108 and @xmath109 for @xmath110 .
the eigenvalues are given by @xmath111 the eigenvalue @xmath112 corresponds to the perturbation parallel to the synchronization manifold , while the other @xmath113 eigenvalues correspond to the perturbation transverse to the synchronization manifold .
the transverse eigenmodes should be damped out to have a stable synchronization manifold .
the stability of the synchronization manifold depends only on the largest eigenvalue @xmath114 and the corresponding variational equation becomes @xmath115 now , the value of the largest lyapunov exponent of the variational equation ( [ ross2 ] ) for each value of the parameters in the two - parameter phase diagram ( fig .
[ pre_fig7 ] ) is used to demarcate the two - parameter phase diagram into desynchronized state ( @xmath79 ) and synchronous state ( @xmath86 ) as indicated by the dotted lines in fig .
[ pre_fig7](a ) . before the dotted lines the largest lyapunov exponent acquires positive values while for the parameters above the dotted lines it acquires negative values attributing to the stability of the synchronization manifold .
figure [ pre_fig7 ] elucidates that the simulation results are in agreement with the results obtained using the semi - analytic approach , namely the master stability function formalism in demarcating the synchronized and desynchronized regimes .
it is to be noted that the synchronous state is a multistable state with coexisting chimera states , coherent states and phase - flip chimeras depending on the distribution of the initial conditions ( for a given choice of @xmath116 and @xmath23 or @xmath116 and @xmath9 ) .
the multistability nature of the synchronized parameter space coexisting along with the phase - flip chimera will be discussed in fig .
[ pre_fig8 ] . ) and complete synchronized states ( green , @xmath117 ) in ( a ) periodic and ( b ) chaotic cases of the coupled rssler oscillators . ] a similar two - parameter phase diagram in the range of @xmath118 and @xmath119 $ ] of the ensemble of rssler oscillators exhibiting chaotic oscillations is shown in fig .
[ pre_fig7](b ) . the main difference between the emergent dynamics from the ensemble of rssler oscillators with periodic oscillations ( see fig . [ pre_fig7](a ) ) and that with chaotic oscillations ( see fig . [ pre_fig7](b ) )
is that in the later case the phase - flip chimera ( marked as @xmath81 ) is immediately preceded by the asynchronous oscillations , indicated by @xmath79 , of the ensemble .
that is , the dynamical transition occurs in the sequence of desynchronized state ( @xmath79 ) , chimera state ( @xmath80 ) , coherent state ( @xmath82 ) or completely synchronized state ( @xmath86 ) depending on the value of @xmath23 , again desynchronized state ( @xmath79 ) and finally phase - flip chimera ( @xmath81 ) as a function of @xmath12 , which is evident from the two - parameter phase diagram in fig .
[ pre_fig7](b ) .
the synchronized and desynchronized parameter space is also demarcated by solving the variational equation eq .
( [ ross2 ] ) as indicated by the dotted lines in fig .
[ pre_fig7 ] . here
, the parameter space enclosed between the dotted lines are characterized by the negative values of the largest lyapunov exponent of the variational equations attributing to the stable synchronized state .
it is to be noted that in both the two - parameter phase diagrams , the discussed transition occurs for @xmath120 , below which the ensemble of oscillators remain in asynchronous state elucidating that the phase - flip chimera may not emerge in the nearest neighbor coupling between the agents .
it also reveals that an appropriate coupling radius is necessary for the emergence of phase - flip chimera . in order to demonstrate the stability of the phase - flip chimera
, we have depicted the basin of attraction of the ensemble , eq .
( [ ross ] ) , in fig .
[ pre_fig8 ] as a function of @xmath121 and @xmath122 , which correspond to the initial conditions of @xmath123 and @xmath124 variables .
the initial conditions for the variables @xmath125 and @xmath126 are uniform random numbers distributed between @xmath19 and @xmath127 , whereas that of @xmath123 and @xmath124 variables are the uniform random numbers distributed between @xmath128 and @xmath129 , respectively .
it is clear that both the phase - flip chimeras ( indicated by circles ) and completely synchronized states ( indicated by triangles ) coexist for a wide choice of initial conditions corroborating the stability and robustness of the phase - flip chimera . and @xmath12 .
the parameter space marked as @xmath130 and @xmath83 corresponds to the desynchronized state , conventional chimera , phase - flip chimera , and complete synchronized state , respectively .
the dotted lines corresponds to the stability curves estimated from the eigenvalues of the variational equation ( [ ross2 ] ) . ] to explore the effect of the environmental coupling on the observed dynamical transition in detail , we have plotted the two - parameter phase diagram as a function of the strength of the nonlocal coupling @xmath12 and the strength @xmath9 with which the local agents interact with their respective oscillators in figs .
[ pre_fig9](a ) and [ pre_fig9](b ) for periodic and chaotic oscillations of the individual uncoupled rssler oscillators , respectively .
the parameter space leading to desynchronous state , conventional chimera , phase - flip chimera and complete synchronous state are marked as @xmath130 and @xmath83 , respectivey , in both the figures .
asynchronous and completely synchronous parameter space are also demaracted ( indicated by dotted lines ) using the value of the largest lyapunov exponent of the variational equation ( [ ross2 ] ) . in the periodic regime of the rssler oscillators ,
the coupled oscillators remain asynchronous in the entire range of @xmath12 for small values of @xmath9 as seen in fig .
[ pre_fig9](a ) . in the range of @xmath131 , desynchronized state
is followed by conventional chimera and the phase - flip chimera as @xmath12 is increased .
increasing @xmath9 further , the conventional chimera and phase - flip chimera regimes are seperated by complete synchronized regime upto to @xmath132 .
transition from asynchronous to complete synchronized state occurs for @xmath133 as a function of @xmath12 .
thus the emergence of phase - flip chimera is spread over a wide range of @xmath134 and @xmath12 . for the chaotic oscillations of the individual rssler oscillators ,
we have observed similar transitions as in fig .
[ pre_fig9](a ) as a function of @xmath9 and @xmath12 except for the fact that the range of phase - flip chimera extends to much wider range of @xmath9 and @xmath12 ( see fig .
[ pre_fig9](b ) ) while the desynchronized state preceeds the phase - flip chimera .
the existence of phase - flip bifurcation / transition in two coupled oscillators induced by the environmental coupling in eq.([ross ] ) was shown in @xcite .
we find that the environmental coupling among the agents and the oscillators induces phase - flip transition / bifurcation while the nonlocal coupling among the agents in the common environment facilitates the onset of incoherent domain at the phase - flip transition . in the absence ( or for low values of the strength ) of the environmental coupling
one can not observe the phase - flip chimera as is evident from the two - phase diagram in fig .
[ pre_fig9 ] .
we have also found that the emergence of the phase - flip chimera depends on the relaxation time of the environmental coupling in eq .
1(d ) and phase - flip chimera emerges for the parameter @xmath135 , which determines the relaxation time of the external agent @xmath10 ( see appendix a for more details ) . and ( b ) the space - time evolution of the ensemble of rssler oscillators in the chaotic regime with global coupling between the agents for the value of the strength of the global coupling @xmath36 . ]
it is also to be noted that we have identified the emergence of the phase - flip chimera in the case of globally coupled agents .
snapshots of the instantaneous phases @xmath136 and the space - time evolution of the ensemble of rssler oscillators in the chaotic regime but with global coupling between the agents for the value of the strength of the global coupling @xmath36 is shown in fig .
[ pre_fig10 ] .
the value of the other parameters and the distribution of the initial conditions are the same as in fig .
[ pre_fig3 ] .
it is clear from the figure that the phase - flip chimera characterized by two adjacent out - of - phase synchronized coherent domains , where nearby oscillators are exhibiting in - phase synchronized oscillations , interspersed by an incoherent domain , where the nearby oscillators exhibit out - of - phase oscillations ( see the inset of fig .
[ pre_fig10 ] ) , exists even with the global coupling between the agents in the common environment .
we have also confirmed the emergence of phase - flip chimera from the ensemble for the periodic oscillations of the individual rssler oscillators with the global coupling between the agents .
it is known that in an ensemble of globally coupled oscillators one can reorder the spatial index of the oscillators such that the spatial inhomogenity with coexisting in - phase and out - of - phase synchronized oscillators in the phase - flip chimera state ( see fig . [ pre_fig2 ] ) can be recasted as a two - cluster state @xcite .
however , we would like to point out that once the spatial ordering of the globally coupled oscillators are fixed by indexing them from @xmath127 to @xmath7 , then the distribution of initial conditions as stated above among the oscillators results in spatially inhomogeneous states with coherent out - of - phase synchronized domains interspersed by incoherent domain comprised of nearby oscillators exhibiting in - phase and anti - phase oscillations .
this confirms the emergence of phase - flip chimera even with global coupling among the agents .
indeed , several recent investigations reported the existence of chimera states in an ensemble of globally coupled oscillators @xcite .
finally , we have also checked whether phase - flip chimeras occur when the coupling is additionally given to @xmath123 or @xmath124 variables that is to eq .
( [ ross]a ) or eq .
( [ ross]b ) . in these cases
only transition from desynchronized state to synchronized state occurs ( see appendix b for more details ) .
in summary , we have identified an interesting type of a collective dynamical regime , called phase - flip chimeras , in an ensemble of identical rssler oscillators coupled indirectly via the agents from a common dynamic environment , where the agents are coupled nonlocally with a coupling radius @xmath23 .
such interactions are found in the diffusion of biomolecues between the cells and their environment @xcite .
the phase - flip chimera is characterized by two out - of - phase synchronized spatially coherent domains interspersed by a spatially incoherent domain comprised of nearby oscillators exhibiting out - of - phase oscillations .
the oscillators in each of the coherent domains exhibit phase synchronized oscillations among themselves , whereas the two adjacent coherent domains in the phase - flip chimera exhibit out - of - phase synchronized oscillations .
the robustness of the phase - flip chimera is also confirmed by depicting its occurrence in a wide range of parameters using the two - parameter phase diagram .
further , it is also shown that the phase - flip chimera is preceded by the conventional chimera and it emerges only after the completely synchronized state emerges . in the chaotic regime the phase - flip chimera is immediately preceded by asynchronous state , which in turn emerges as a desynchronized state from the synchronous evolution of the ensemble of oscillators .
we have used the strength of incoherence , probability distribution of the correlation coeffecient and the master stability function to characterize the observed dynamical transition of the ensemble of the rssler oscillators .
it is also confirmed that the phase - flip chimera emerges even in the global coupling among the agents , whereas with the nearest neighbor coupling such a state does not emerge .
the work of vkc forms part of a research project sponsored by insa young scientist project under grant no .
sp / ysp/96/2014
. dvs is supported by the serb - dst fast track scheme for young scientist under grant no . st / ftp / ps-119/2013 .
ml is supported by a nasi platinum jubilee senior scientist fellowship .
the dependence of the dynamics of the oscillators on the relaxation time is depicted in fig . [ time ] .
dynamics of the agent @xmath137 ( as an example ) is shown in fig .
[ time](a ) for @xmath39 , and in fig .
[ time](b ) for @xmath30 .
the dynamics of randomly chosen oscillators are shown in figs .
[ time](c ) and [ time](d ) for different values of the relaxation parameter @xmath6 occurring in eq .
( [ ross4 ] ) . is shown in ( a ) for @xmath39 , and ( b ) for @xmath30 . in both the figures @xmath22 .
the dynamics of random oscillators are shown in ( c ) and ( d ) for different values of the relaxation parameter @xmath6 . ] as the value of the parameter @xmath6 is increased the relaxation time ( pointed out by filled circles in the insets ) decreases both in the absence ( fig .
[ time](a ) ) and in the presence ( fig .
[ time](b ) ) of the coupling between the agents @xmath8 in eq .
( [ ross4 ] ) .
it is to be noted that the relaxation time is very less in the presence of the coupling between the oscillators ( see fig .
[ time](b ) ) compared to the uncoupled oscillators ( see fig . [ time](a ) ) .
further , the value of the parameter @xmath6 which determines the relaxation time dictates the collective behavior of the ensemble of the oscillators .
phase - flip chimeras are observed for @xmath135 . for @xmath138
the phase - flip chimera is already shown in figs . 1 and 3 .
for @xmath139 , the out - of - phase oscillations of the adjacent out - of - phase synchronized coherent domains of the phase - flip chimera are represented by a couple of representative oscillators in fig .
[ time](c ) . on the other hand , for @xmath140
, we observe only synchronized state for any choice of initial conditions and coupling strength as shown by the representative oscillators in fig .
[ time](d ) .
thus we find that the emergence of the phase - flip chimera depends on the relaxation time of the environmental coupling in eq .
( [ ross4 ] ) and that it emerges for the parameter @xmath141 , which determines the relaxation time of the @xmath8 s . of the ensemble of rssler oscillators coupled via @xmath124-variable displaying ( a ) desynchronized state for @xmath142 , ( b ) completely synchronized state for @xmath27 , ( c ) desynchronized state for @xmath142 and ( d ) completely synchronized state for @xmath143 .
left column for periodic oscillations and right column for chaotic oscillations of the rssler oscillators .
insets in fig.(b ) and ( d ) depict the snapshots of @xmath144 . ]
we have also examined the typical transition from desynchronized state to synchronized states by coupling the agent @xmath10 to the other variables @xmath145 or @xmath146 in the same fashion as in eq .
( [ ross4 ] ) .
snapshots of the instantaneous phases @xmath136 of the ensemble of rssler oscillators coupled via the @xmath124-variable are shown in fig .
[ otherv ] .
the left column corresponds to the case of periodic oscillations when the oscillators are uncoupled , while the right column corresponds to the case of the chaotic oscillations of the uncoupled oscillators .
desynchronized state is observed for @xmath142 ( see fig . [ otherv](a ) ) for the case of periodic oscillations . upon increasing the coupling strength the oscillators
are completely synchronized as shown in fig .
[ otherv](b ) for @xmath27 and remains synchronized for any large coupling strength . for the case of chaotic oscillations of the uncoupled rssler oscillators
, the oscillators are desynchronized for low values of the coupling strength as shown in fig .
[ otherv](c ) for @xmath142 and remain completely synchronized above a threshold value of @xmath12 as illustrated in fig .
[ otherv](d ) for @xmath143 . hence
, only the existence of completely synchronized state from the desynchronized state is found and we are not able to find phase - flip chimera for any value of @xmath9 and @xmath12 .
similar scenario arises for @xmath123 coupling as well ( which we do not present here ) .
d. battogtokh and y. kuramoto , nonlinear . phenom .
sys . * 5*,380 ( 2002 ) ; d. m. abrams and s. h. strogatz phys .
lett . * 93 * , 174102 ( 2004 ) ; d. m. abrams , r. mirollo , s. h. strogatz and d. a. wiley , phys . rev . lett .
* 101 * , 084103 ( 2008 ) ; g. c. sethia , a. sen and f. m. atay , phys .
* 100 * , 144102 ( 2008 ) ; m. j. panaggio and d. m. abrams , nonlinearity 28 , r67 ( 2015 ) . c. .r .
laing and c. c. chow , neural computation * 13 * , 1473 ( 2000 ) ; c. r. laing , phys . rev .
e 81 , 066221 ( 2010 ) ; g. bordyugov , a. pikovsky , and m. rosenblum , phys . rev .
e * 82 * , 035205 ( 2010 ) ; s. olmi , a. politi , and a. torcini , europhys .
lett.*92*,60007 ( 2010 ) .
i. omelchenko , y. maistrenko , p. hvel , and e. schll , phys .
* 106 * , 234102 ( 2011 ) ; i. omelchenko , b. riemenschneider , p. hvel , y. maistrenko , and e. schll , phys .
e * 85 * , 026212 ( 2012 ) ; i. omelchenko , o. e. omelchenko , p. hvel , and e. schll , phys . rev .
* 110 * , 224101 ( 2013 ) .
s. r. ujjwal , and r. ramaswamy , phys .
e * 88 * , 032902 ( 2013 ) ; d. pazo , and e. montbri , phys .
x , * 4 * , 011009 , ( 2014 ) ; l. schmidt , k. schnleber , k. krischer , and v. garciamorales , chaos * 24 * , 013102 ( 2014 ) . j. m. cruz , j. escalona , p. parmananda , r. karnatak , a. prasad , and r. ramaswamy , phys .
e * 81 * , 046213 ( 2010 ) ; a. prasad , j. kurths , s. k.dana and r. ramaswamy , phys .
e * 74 * , 035204(2006);a .
sharma , m. d. shrimali , a. prasad , r. ramaswamy and u. feudel , phys .
e * 84 * , 016226 ( 2011 ) .
h. ammann , r. gray , i. shvarchuck , and n. christensen , phys . rev .
lett . * 80 * , 4111 ( 1998 ) ; w. h. zurek , s. habib , and j. p. paz , phys . rev . lett . * 70 * , 1187 ( 1993 ) ; s. a. gurvitz , l. fedichkin , d. mozyrsky , and g. p. berman , phys . rev
. lett . * 91 * , 066801 ( 2003 ) ; s. braig and k. flensberg , phys .
b * 68 * , 205324 ( 2003 ) .
a. kuznetsov , m. krn , and n. kopell , siam j. appl .
math.*65 * , 392 ( 2004 ) ; r. wang and l. chen , j. biol .
rhythms*20 * , 257 ( 2005 ) ; d. gonze , s. bernard , c. waltermann , a. kramer , and h. herzel , biophys . j.*89 * , 120 ( 2005 ) .
e. ullner , a. zaikin , e. i. volkov , and j. garcia - ojalvo , phys . rev
99 * , 148103 ( 2007 ) ; e. ullner , a. koseska , j. kurths , e. volkov , h. kantz , and j. garcia - ojalvo , phys . rev .
e. * 78 * , 031904 ( 2008 ) .
k. okuda , physica d * * 6**3 , 424 ( 1993 ) ; v. hakim and w .- j .
rappel , phys .
a * * 4**6 , r7347 ( 1992 ) ; n. nakagawa and y. kuramoto , prog .
. phys . * * 8**9 , 313 ( 1993 ) ; y. kuramoto , prog .
phys.**9**4 , 321 ( 1995 ) .
d. h. zanette and a. s. mikhailov , phys .
e * * 5**7 , 276 ( 1998 ) ; d. h. zanette and a. s. mikhailov , phys .
e * * 6**2 , r7571(r ) ( 2000);w .
wang , i. z. kiss , and j. l. hudson , chaos * * 1**0 , 248 ( 2000 ) ; m. wickramasinghe and i. z. kiss , plos one * 8 * , e80586 ( 2013 ) . | we report the emergence of a collective dynamical state , namely phase - flip chimera , from an ensemble of identical nonlinear oscillators that are coupled indirectly via the dynamical variables from a common environment , which in turn are nonlocally coupled .
the phase - flip chimera is characterized by the coexistence of two adjacent out - of - phase synchronized coherent domains interspersed by an incoherent domain , in which the nearby oscillators are in out - of - phase synchronized states .
attractors of the coherent domains are either from the same or different basins of attractions depending on whether they are periodic or chaotic .
conventional chimera precedes the phase - flip chimera in general .
further , the phase - flip chimera emerges after the completely synchronized evolution of the ensemble in contrast to conventional chimeras which emerge as an intermediate state between completely incoherent and coherent states .
we have also characterized the observed dynamical transitions using the strength of incoherence , probability distribution of correlation coefficient and the framework of master stability function . |
beta decay is an important process at the intersection of nuclear physics , astrophysics , and particle physics . the rapid neutron - capture process ( @xmath0 process ) proceeds through neutron - rich nuclei , the beta - decay rates for which determine final abundance distributions .
the significance of the reactor neutrino anomaly for exotic new neutrino physics depends on forbidden beta - decay rates in neutron - rich fission products @xcite . in both these cases ,
the important rates are difficult or impossible to measure ; we need to be able to calculate them instead .
the random - phase approximation ( rpa ) and its generalization , the quasiparticle random - phase approximation ( qrpa ) , nowadays typically used in conjunction with skyrme energy - density functionals ( edfs ) , are established tools for treating nuclear excitations .
the matrix version of the charge - changing ( or pn " ) skyrme qrpa has been applied with some success in spherical nuclei . when the rotational symmetry of the mean field is broken , however , the dimension of the mean - field two - quasiparticle basis increases by orders of magnitude and the qrpa matrix becomes too large to fit in the main memory of a typical computer without aggressive truncation .
even then , supercomputing is needed to solve the equations .
we have constructed a deformed matrix skyrme pnqrpa program @xcite from the code reported in ref .
@xcite but can not use it in reasonable amounts of computing time .
the finite amplitude method ( fam ) is a much more efficient scheme for finding the linear response .
@xcite first proposed the method and ref .
@xcite quickly applied it to obtain the rpa response in spherical and deformed nuclei .
ref . @xcite generalized the approach to the qrpa , and ref .
@xcite applied the generalization to monopole transitions . in this article
, we further extend the fam to charge - changing qrpa transitions of arbitrary intrinsic angular momentum projection @xmath1 in deformed nuclei .
we call the resulting approach the skyrme proton - neutron finite amplitude method ( pnfam ) .
to illustrate the method , we examine the effects of skyrme s tensor terms on beta - decay rates .
minato and bai @xcite observed that a tensor interaction can reduce beta decay half - lives of magic and semi - magic nuclei considerably , bringing them into closer accord with experiment .
if a similar reduction takes place in deformed nuclei , it might make it impossible to include an isoscalar pairing interaction without underpredicting half - lives . on the other hand , it might instead allow a better - behaved isoscalar pairing interaction , one that depends less on mass than those in use today .
after a preliminary pnfam analysis of the effects of tensor interaction in both semi - magic and deformed nuclei , we assess the situation here .
this work will serve as a stepping stone towards @xmath0-process studies in the rare - earth region , evaluation of neutrino - capture rates , and a more data - rich determination of the time - reversal ( t ) odd parts of energy - density functionals .
the rest of the article is organized as follows : section [ sec : theory ] lays out the form of the our skyrme functionals and discusses the application of the fam to beta decay , section [ sec : method ] presents our implementation and consistency checks , and section [ sec : results ] uses the pnfam to study the tensor interaction in a small set of open - shell and deformed nuclei .
section [ sec : conclusions ] is a conclusion .
in the particle - hole channel we use the standard general skyrme edf , the details of which may be found in many places , e.g. in refs . @xcite . in the notation of ref .
@xcite , the edf takes the form @xmath2 where @xmath3 \rho_{tt_3}^2 + c_t^{\delta\rho } \rho_{tt_3}\nabla^2\rho_{tt_3 } \\ & + c_t^\tau \rho_{tt_3}\tau_{tt_3 } + c_t^{j } \mathbb{j}^2_{tt_3 } + c_t^{\rho\nabla j}\rho_{tt_3}\bm\nabla\cdot{\ensuremath{\mathbf{j}}}_{tt_3 } \end{aligned}\ ] ] is bilinear in local time - even densities , and @xmath4{\ensuremath{\mathbf{s}}}_{tt_3}^2 + c_t^{\delta s}{\ensuremath{\mathbf{s}}}_{tt_3 } \cdot \bm\nabla^2 { \ensuremath{\mathbf{s}}}_{tt_3 } + c_t^{j } { \ensuremath{\mathbf{j}}}_{tt_3}^2 \\ & + c_t^t { \ensuremath{\mathbf{s}}}_{tt_3 } \cdot { \ensuremath{\mathbf{t}}}_{tt_3 } + c_t^{s\nabla j } { \ensuremath{\mathbf{s}}}_{tt_3 } \cdot \bm\nabla \times { \ensuremath{\mathbf{j}}}_{tt_3 } \\ &
+ c_t^f { \ensuremath{\mathbf{s}}}_{tt_3 } \cdot { \ensuremath{\mathbf{f}}}_{tt_3 } + c_t^{\nabla s } \left(\bm\nabla \cdot { \ensuremath{\mathbf{s}}}_{tt_3}\right)^2 \end{aligned}\ ] ] is bilinear in time - odd local densities .
only the coupling constants @xmath5 $ ] and @xmath6 $ ] are allowed to be density - dependent themselves , viz .
: @xmath7 & = c_{t,0}^\rho + c_{t,\rho}^\rho \rho_{00}^{\sigma_\rho}\\ c_t^s[\rho_{00 } ] & = c_{t,0}^s + c_{t,\rho}^s \rho_{00}^{\sigma_s } \ , , \end{aligned}\ ] ] and even they depend only on the total density @xmath8 our implementation of the pnfam , through a code we call , is self - consistent and so must be preceded by a hartree - fock - bogoliubov ( hfb ) calculation , for which we use the popular code hfbtho @xcite .
because the pnfam treats only charge - changing transitions and hfbtho does not allow proton - neutron mixing , our results depend only on charge - changing densities [ those with isospin indices @xmath9 ; the usual coulomb and kinetic contributions to the total energy in eq .
are not necessary . in the particle - particle ( pairing ) channel
, we use a density - dependent interaction of the form @xmath10 where @xmath11 @xmath12 is the saturation density of nuclear matter and @xmath13 $ ] controls the density - dependence .
this form is similar to that allowed by hfbtho .
the @xmath14 pairing term , however , has no effect in the hfb calculation as long as explicit proton - neutron mixing is forbidden .
the pairing strength @xmath15 is thus unconstrained by the mean field and becomes a free parameter in our subsequent pnfam calculation . on the other hand the @xmath16 pairing , though important in the hfb calculation , has no dynamical effect on gamow - teller transitions . it does play a role for other multipoles , however , and we set its strength @xmath17 to the average of the hfb proton and neutron pairing strengths ( which hfbtho allows to be different ) , that is @xmath18 .
although the coupling constants of eqs . and can be derived from the parameters that specify a skyrme interaction " ( the @xmath19 and @xmath20 parameters ) @xcite , there need be no underlying interaction and the couplings of the time - odd part of the functional need not be connected with those of the time - even part .
most skyrme edfs are fitted to ground - state properties of spherical or axially - symmetric even - even nuclei , which are independent of the time - odd functional . even if properties of odd nuclei are included , time - odd densities and currents
appear not to contribute very much @xcite . as a result ,
up to relations that follow from gauge invariance , the time - odd couplings in the edf picture are undetermined by such fits . in recent parameterizations , e.g. the unedf @xcite and sv @xcite series of functionals , that fact is made explicit :
the time - odd coupling constants are either neglected completely or are constrained solely by gauge invariance .
some time - odd couplings can be profitably fit to the energies and strengths of gamow - teller resonances ; see , e.g. , ref . @xcite or ref.@xcite . here
we will sometimes use the simple prescriptions of ref .
the ability to treat charge - changing resonances in deformed nuclei via the pnfam should soon open the door to a better determination of the t - odd functional .
when tensor terms are included , there is an additional subtlety in the time - even channel that determines mean - field properties in even - even nuclei .
the term involving the spin - current tensor @xmath21 can actually be decomposed into three tensor components @xcite : @xmath22 most groups ( e.g. the authors of ref .
@xcite ) have thus far fit the tensor terms solely in spherical nuclei . there , the spin - orbit density @xmath23 is the only non - vanishing component of eq .
@xcite . when spherical symmetry is broken , however , the contribution of the pseudotensor @xmath24 is non - zero in the mean field and introduces another ( undetermined ) coupling constant . in our calculations with hfbtho ,
the coupling constant multiplying @xmath25 is restricted to obey the relation @xmath26 .
as detailed in ref .
@xcite , this is only one of several options in deformed nuclei .
the choice means that we can not require both that the functional be gauge - invariant and that it have a non - vanishing @xmath27 .
@xcite derives a form of the fam that corresponds to the like - particle qrpa .
the formulation is general enough , however , to cover the charge - changing case as well . in the following
we discuss the special features of the fam that follow from charge changing , i.e. from choosing an external field that transforms neutrons into protons ( e.g. for @xmath28 decay ) .
charge - changing transitions are generated by a weak external field @xmath29 , with ( complex ) angular frequency @xmath30 , of the form @xmath31 where @xmath32 is a small real parameter and @xmath33 is a one - body operator that could depend on @xmath30 but in our application does not . transformed to the quasiparticle basis , it has the form @xmath34 where the ellipses refer to terms of the form @xmath35 that do not contribute to the linear response .
the summation runs over every pair of quasiparticle states in the basis , avoiding double - counting .
a one - body @xmath28 transition operator ( fermi , gamow - teller , or forbidden ) can be written in a single - particle basis as @xmath36 where the index @xmath37 runs over proton states and the index @xmath38 over neutron states .
the @xmath39 are the single - particle matrix elements of the transition operator . here , unlike in the charge - conserving fam , @xmath33 is non - hermitian . without proton - neutron mixing in the static hfb solution , the bogoliubov transformation yields @xmath40 and @xmath41 where @xmath42 and @xmath43 label the proton and neutron quasiparticle states , and u and v are the usual bogoliubov transformation matrices @xcite .
the weak external field @xmath33 induces weak time - dependent oscillations in the quasiparticle annihilation operators ( e.g. for neutrons ) , @xmath44 these oscillations in turn lead to oscillations in the charge - changing density matrix elements @xmath45 and @xmath46 , the charge - changing pairing tensors @xmath47 and @xmath48 , and the resulting energy functional @xmath49 $ ] .
the single - particle hamiltonian @xmath50 and pairing potential @xmath51 likewise acquire time - dependent pieces through the relations @xmath52 where @xmath53 is a proton index and @xmath54 a neutron index , or vice versa .
the oscillations in all quantities occur with the same frequency @xmath30 , and the time - dependence , which is contained only in exponentials like those in eq.([eq : f1 ] ) , can be factored out and removed .
the fam then amounts to solving the small - amplitude limit of the time - dependent hfb equation ( with the time dependence factored out ) .
the reason the procedure is so efficient is the numerical computation of the derivatives of @xmath50 and @xmath51 in the direction of the perturbation , i.e. with respect to @xmath32 . in our pnfam the differentiation is somewhat easier than in the like - particle case because the charge - changing parts of @xmath55 and @xmath56 vanish at the hfb minimum ( since our hfb does nt mix protons with neutrons . )
that restriction , together with the linear dependence of @xmath50 and @xmath51 on the charge - changing densities for all published skyrme functionals , means that the pnfam numerical derivatives are independent of the parameter @xmath32 .
in fact , our code does nt reference @xmath32 at all and we do not need to worry , as did the authors of ref.@xcite , about choosing @xmath32 small enough so that terms of @xmath57 are negligible , but large enough to avoid round - off errors .
when all is said and done , the pnfam equations for the linear response become @xmath58 where the @xmath59 s are the hfb quasiparticle energies and @xmath60 and @xmath61 are the pieces of the frequency - dependent hfb hamiltonian matrix , expressible in terms of @xmath50 and @xmath51 , that multiply the quasiparticle pair - creation and annihilation operators , as in eq .
( [ eq : f2 ] ) @xcite .
one can go on from eqs . to derive the traditional matrix - qrpa equations by expanding @xmath62 and @xmath63 in @xmath64 and @xmath65 ( on which @xmath66 depends implicitly via the densities ) and taking the limit of vanishing external field .
but the point of the fam is to solve the nonlinear system of equations ( [ eq : famequations ] ) instead of constructing the traditional qrpa @xmath67 and @xmath68 matrices . after solving eqs .
( [ eq : famequations ] ) for the amplitudes @xmath69 and @xmath70 , one can compute the strength distribution for the operator @xmath33 : @xmath71 where @xmath72 the last form , combined with eq
. for complex frequency @xmath73 , leads to @xmath74 which shows that the fam transition strength a distance @xmath75 above the real axis is just the qrpa strength function smeared with a lorentzian of width @xmath75 .
the last three equations imply the symmetry @xmath76 to evaluate non - unique forbidden decay rates we will need to take into account the interference between distinct transition operators , called @xmath33 and @xmath77 here for simplicity .
such terms have the form @xmath78 we compute them by using , e.g. , the operator @xmath33 to generate the response and then calculating the effect on the quantity represented by @xmath77 : @xmath79\,.\ ] ] in deformed nuclei , all the results above are in the intrinsic frame , where angular momentum is not conserved .
the symmetry must be restored , at least approximately , and the crudest way to do so is by treating the intrinsic state like the particle in the particle - rotor model @xcite . in that picture
every intrinsic state corresponds to the lowest state in a rotational band , and has a rotational energy @xmath80 where @xmath81 is the moment of inertia of the nucleus .
we use the hfb version of the belyaev formula @xcite ( see the appendix for details ) to approximate @xmath81 : @xmath82 here @xmath83 and @xmath84 label quasiparticle states of the same particle type ( proton or neutron ) .
the energy shifts are typically only tens of kev , but their effects in beta - decay rates are magnified by the phase - space integrals and can be non - negligible ( see sec .
[ sec : results ] ) . in this work
we consider both allowed and first - forbidden beta decay .
expressions for the relevant impulse - approximation operators were worked out some time ago , e.g. in refs.@xcite . in this section
we restrict ourselves to allowed decay ; the more complicated expressions for the first - forbidden decay can be found in appendix [ app : forbidden - formulas ] .
the total allowed decay rate is proportional to the sum of individual transition strengths @xmath85 to all energetically allowed states @xmath86 in the daughter nucleus , weighted by phase - space integrals : @xmath87 where the constant @xmath88 s comes from superallowed decay @xcite .
the phase - space integral , containing the details of final - state lepton kinematics , is @xcite @xmath89 where @xmath90 is the charge of the daughter nucleus , @xmath91 , @xmath92 is the electron energy in units of electron mass , @xmath93 is the electron momentum , and @xmath94 is one of the ( generalized ) fermi functions @xcite @xmath95^{2 } 4^{k_e } ( 2pr)^{2(\gamma_{k_e}-k_e ) } \\
\quad \times \exp(\pi y ) \frac{\left|\gamma\left(\gamma_{k_e}+iy\right)\right|^{2 } } { \left[\,\gamma\left(2\gamma_{k_e}+1\right)\,\right]^{2 } } \ , .
\end{gathered}\ ] ] here @xmath96 is related to the orbital angular momentum of the emitted electron ( see , e.g. , ref .
@xcite for the definition ) , @xmath97 , @xmath98 , and @xmath99 is the nuclear radius .
( the primakoff - rosen approximation to this expression @xcite is often used for computing allowed decay but we retain the more general form , which also applies to forbidden beta decay . ) the coulomb function @xmath100 is @xmath101 to use these expressions we need the energies in the final nucleus with respect to the ground state of the initial nucleus .
we take our ground - state - to - ground - state @xmath102 value from the approximation in ref.@xcite , @xmath103 where @xmath104 and @xmath105 are the proton and neutron fermi energies from the hfb solution , @xmath106 mev is the mass difference between the neutron and hydrogen atom , and the ground - state energy is taken to be the sum of the lowest proton and neutron quasiparticle energies : @xmath107 one virtue of the approximation in eqs .
( [ eq : qvalue ] ) and ( [ eq : egs1 ] ) is that the independent - quasiparticle approximation to the ground - state energy cancels out in calculations of beta - decay lifetimes @xcite .
we use the approximation for the ground - state energy only when studying strength distributions . how do we actually evaluate lifetimes from the pnfam response ?
. implies that the transition strength of the operator @xmath33 between the qrpa state with energy @xmath108 and the initial ground state is the residue of the function @xmath109 at that energy , @xmath110 \,,\ ] ] and eq .
that the cross terms contributing to forbidden decay rates are @xmath111 \,.\ ] ] the connection to residues allows us to represent beta - decay rates as contour integrals of the pnfam response in the complex - frequency plane .
the representation is complicated a little by the fact that the phase - space integrals are not analytic functions .
but we can replace the phase - space integrals with other functions that are analytic , at least inside the contour , and that coincide with the phase - space integrals at the poles of the strength function that contribute to the integral .
a high - order polynomial of the form @xmath112 fitted to the phase - space integral on the real axis , serves our purpose . while we do not know the exact locations of the poles of the strength function , we do know they lie on the positive real axis ( and that mirrored , unphysical poles lie on the negative real axis ) . with the polynomials
, we can cast the equations for beta - decay rates in a form that captures the contributions of all the individual excited states in a contour that encloses them .
the gamow - teller part of the rate takes the form @xmath113 \\ = & -\frac{\ln 2}{\kappa } \sum_n \operatorname{res } [ f_\textrm{poly } s(\sigma\tau_-),\omega_n ] \\
= & -\frac{\ln 2}{\kappa } \frac{1}{2\pi i } \oint_c { \ensuremath{\mathrm{d}}}\omega \ : f_\textrm{poly } ( \omega ) s(\sigma\tau_-;\omega ) , \end{split}\ ] ] where the contour @xmath114 encloses the same poles @xmath38 that are initially summed over .
a practical choice for the contour is a circle @xmath115 crossing the real axis at the origin and at the maximum energy @xmath116 figure [ fig : contour ] displays such a contour schematically .
a circular contour allows the use of the symmetry in eq . to halve the number of pnfam computations .
the analog of eq . for first - forbidden beta decay is lengthy and presented in the appendix [ app : forbidden - formulas ] .
as mentioned , our method begins with the use of the code hfbtho@xcite to carry out an axially - deformed hfb calculation . in our tests 16 harmonic oscillator shells
are enough to allow the low - energy strength functions to converge , and we adopt that number for all half - life calculations .
our contour integration requires a reasonably accurate polynomial approximation to the fermi integrals in eq . .
figure [ fig : polyfit ] illustrates the quality of our fit to the allowed - decay fermi integral . in practice , a 10th - order polynomial of the form is more than sufficient .
another requirement is that the integrands are smooth enough to allow numerical quadrature .
figure [ fig : contourint ] demonstrates that that is the case , displaying a typical integrand as a function of the curve parameter @xmath19 in eq . .
the integrand is indeed smooth enough to treat with conventional quadrature ; we use the compound simpson s @xmath117 rule .
( eq . ) for the beta decay of @xmath118ba .
the solid line is calculated with the exact fermi function @xmath119 ( eq . ) and the points correspond to the polynomial approximation . ] allowed decay rate of @xmath120ba , with sko and the tensor interaction .
the integrand behaves well enough to allow simple quadrature .
the origin in the complex plane corresponds to @xmath121 ( see text).,width=317 ] to test the pnfam solver itself , we compare in fig . [
fig : matrix - comparison ] the pnfam gamow - teller transition strength function in the deformed isotope @xmath122ne with that produced by the traditional matrix - qrpa code used in ref .
the matrix code uses the vanderbilt hfb solver @xcite as its starting point .
the slight differences between the two strength functions are due to similarly slight differences in the hfb solutions , which in turn stem from different single - particle bases and truncation schemes .
ne with the same function from the matrix qrpa ( lines ) , smeared with a lorentzian .
we use the skyrme functional skm * without including @xmath123 terms or pairing in the qrpa . ]
finally , we turn to our prescription for the nuclear moment of inertia .
the approximation in eq .
[ eq : rotcorrection ] appears to yield systematically higher values than does experiment , indicating that almost none of the nuclei we examine below are as rigid as the straightforward extension of the beliaev formula predicts .
( we can implement a better approximation that takes into account rpa correlations the thouless - valatin prescription @xcite once a like - particle fam for general @xmath1 exists . ) to assess the sensitivity of the gamow - teller half - life to the rotational energy correction , we use the sko functional detailed in the next section to calculate half - lives in a few test nuclei ( fig .
[ fig : rotcorr - sensitivity ] ) .
although the strong dependence of the phase - space integral on the energy released in the decay can make a correction of the order of ten percent to the half - life , that error is still at most comparable to the error in the calculated @xmath102 value .
the accuracy of the generalized beliaev moment of inertia is therefore good enough for use with present - day energy functionals .
recent work @xcite on semi - magic nuclei in the spherical qrpa indicates that tensor terms in skyrme edfs have significant effects on beta - decay rates . here
we explore the issue in open shell nuclei , both spherical and deformed .
we choose a set of isotopes for which both beta - decay rates and the allowed contribution to those rates have been measured . to make contact with ref .
@xcite we use the same underlying sko functional , with the same additional tensor piece ( i.e. the interaction parameters @xmath124 , @xmath125 , in the notation of ref .
we also adopt the ref .
@xcite procedure of breaking self - consistency by omitting the central @xmath126 terms from the hfb calculation while including them in the qrpa .
unlike ref .
@xcite , however , we include the rotational energy correction and we approximate the ground - state energy by the sum of the lowest proton and neutron quasiparticle energies .
these differences in procedure have small effects on the @xmath102 value and half - life ( via the phase space available to emitted leptons ) .
the last difference with ref .
@xcite : we use the quenched value @xmath127 rather than 1.27 for the axial - vector coupling constant .
we fit the constants in the isovector pairing interaction eq . to a three - point interpolation of measured separation energies .
the sko functional with this pairing interaction reproduces @xmath102 values well , both with and without tensor terms .
[ fig : halflives ] shows the ratios of computed and experimental partial gamow - teller half - lives for our set of nuclei .
the tensor interaction systematically reduces the half - lives , as in the magic and semi - magic nuclei examined by ref . @xcite . in our spherical nuclei , with or without open shells ,
the agreement with experiment improves dramatically . in the deformed isotopes , however , the half - lives with sko tend to be quite low even without the tensor terms , which actually make the half - lives too short .
the situation is thus more complicated than it seems when restricted to spherical systems .
figure [ fig : strfunction ] shows the effects of the tensor interaction in more detail , in four isotopes .
the new terms pull gamow - teller strength down in energy in each case , and smear the resonances .
the movement of strength to lower energies explains the decrease in half - life ; the lower - energy strength means more phase space for leptons and an increased rate .
how many of the features in figs .
[ fig : halflives ] and [ fig : strfunction ] are due to the violation of self - consistency between the hfb and qrpa calculations ?
how many are due to the limited set of nuclei that we examine ? to the restriction to allowed decay ? to the simple addition of a tensor interaction without any attempt to refit data ?
we ca nt address these questions fully here , but can make a start .
we now investigate a slightly larger set of nuclei ( that overlaps our original set ) with a fully self - consistent calculation that includes first - forbidden contributions to the rate .
we choose as a starting point the functional sko@xmath128 @xcite , which reproduces experimental @xmath102 values as well as sko and does a good job on beta - decay rates in semi - magic isotopes @xcite .
we adjust the time - odd part of the functional , setting @xmath129 to avoid instability ( observed , e.g. , in @xcite ) and @xmath130 to reproduce the gamow - teller resonance energy in @xmath131pb .
we leave the other coupling constants untouched .
when we include the tensor interaction we use the values implied by the skyrme @xmath19 and @xmath20 parameters ; the relations between these parameters and the @xmath114 s are given , e.g. , in ref .
all this is the same prescription for the time - odd terms that was found practical ( without tensor terms ) in ref .
@xcite .
[ fig : halflives - skop ] shows some of the results . without the tensor interaction it is possible to roughly reproduce the half - lives through an appropriate strength for the isoscalar pairing interaction in eq .
[ eq : vpp ] ( about 60% of isovector interaction strength ) ; in the analysis leading to figs .
[ fig : halflives ] and [ fig : strfunction ] we did not include isoscalar pairing , which has been the most convenient remedy for many of the qrpa s deficiencies .
ref.@xcite suggests that the tensor interaction can obviate strong isoscalar pairing .
to begin to test this idea , we add the same tensor terms we used with the sko functional .
there is no particular justification for this choice other than its successes with sko and the lack of any work on tensor interactions in conjunction with sko@xmath128 . yet , as figure [ fig : halflives - skop ] shows , these tensor terms lower the half - lives in very much the same way as isoscalar pairing . and first - forbidden contributions included .
the bare sko@xmath128half - lives ( solid blue line ) are systematically longer than experiment and can be reduced by introducing either isoscalar pairing ( dashed green line ) , a full tensor interaction ( dashed red line ) , or only the time - odd components associated with such an interaction ( dotted line).,width=321 ] as mentioned above , however , the simple addition of a tensor interaction spoils the functional s ability to reproduce data .
we have compensated for the problem in the time - odd channel by readjusting @xmath132 , but have done nothing to repair the time - even channel , the original parameters of which were obtained through careful fits to energies , radii , etc .
we therefore look at what happens when we leave the time - even part of sko@xmath128 alone , adding tensor terms to the time - odd part only . to make the changes truly minimal
, we allow the tensor interaction to alter only the two time - odd coupling constants @xmath133 and @xmath134 that receive no contribution from any other piece of a typical density - dependent interaction .
we again set @xmath135 , and refit @xmath132 ( now to 181 @xmath136 ) to reproduce the @xmath131pb resonance .
[ fig : halflives - skop ] shows that even these modifications , which ( again ) do not alter other predictions , mimic much of the reduction in half - lives produced by isoscalar pairing .
these results suggest that the time - odd piece of the functional is much richer than previously suspected .
the time is ripe for a much more careful analysis of all these terms .
our pnfam will allow data from charge - exchange reactions in deformed nuclei to be included in fits .
a like - particle version would increase the range of usable data further .
together with modern optimization techniques , the efficient calculation of linear response should make for a vast improvement in our ability to describe beta decay and predict it in important @xmath0-process isotopes where measurement is not possible .
we have adapted the finite amplitude method for the computation of beta - decay strength functions and rates in axially - deformed even - even nuclei with modern skyrme - like energy - density functionals .
while formally equivalent to the traditional matrix qrpa , the fam is far more robust and just as useful as long as the full set of qrpa energies and transition - matrix elements is not needed . to demonstrate the pnfam s power ,
we have taken a first look at the effect of skyrme s tensor terms on allowed and first - forbidden beta decay in open - shell isotopes .
we find that the tensor interaction lowers half - lives in deformed nuclei much like it does in the spherical nuclei studied in ref .
working with the functional sko@xmath128 , we are able to roughly reproduce measured rates in a range of nuclei without strong isoscalar pairing and without spoiling the predictions of the functional in even - even systems .
it is clearly time to explore time - odd functionals systematically , and we intend to do so soon . finally , beta decay is only one possible application of the pnfam .
neutrino scattering , hadronic charge exchange , and double - beta decay are three others that come quickly to mind .
we thank dr .
markus kortelainen , dr . nobuo hinohara and prof .
witold nazarewicz for useful discussions , and dr .
ewing lusk for help with load - balancing in our computations .
support for this work was provided through the scientific discovery through advanced computing ( scidac ) program funded by u.s .
department of energy , office of science , advanced scientific computing research and nuclear physics , under award number de - sc0008641 , er41896 and by the u.s .
department of energy topical collaboration for neutrinos and nucleosynthesis in hot and dense matter , under award number de - sc0004142 .
z. z. acknowledges the support of the tubitak - turkey , fellowship no:2219 .
we used resources at the national energy research scientific computing center , which is supported by the office of science of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 .
the beliaev formula @xcite for the moment of inertia in the bcs approximation is easy to extend to the hfb approximation .
writing the wave function of the rotating state to the first order in the rotational speed @xmath30 @xmath137 yields for the moment of inertia @xmath138 and carrying out the contractions , we obtain the expression @xmath139 the beliaev formula @xcite @xmath140 applies in the special case that the quasiparticle transformation is diagonal .
when forbidden operators contribute non - negligibly to beta decay , the transition strength @xmath85 in must be replaced by a more general integrated shape function @xmath141 where @xmath146 . here
mev@xmath148 is the nucleon mass , and @xmath149 fm is the ( reduced ) electron compton wavelength .
all operators and resulting quantities are normalized to the electron mass so that the quantity @xmath150 in eq .
is dimensionless @xcite .
the factors @xmath151 arise from the transformation from intrinsic to laboratory reference frames @xcite : @xmath152 where @xmath153 the shape factors for the non - unique forbidden decay are worked out e.g. in ref . @xcite . expressing the squared matrix elements and the interference terms in terms of residues and replacing the kinematic parts of integrands by polynomial expressions that closely approximate them along a portion of the real axis (
see main text ) one can write first - forbidden shape functions in the form @xmath154 where the @xmath155 are linear combinations of functions @xmath156 and @xmath157 is defined in eqs . and .
for @xmath158 we have and finally , for @xmath168 have @xmath169 we have used the shorthand @xmath170 where @xmath83 is the fine - structure constant and @xmath99 is the nuclear radius .
the polynomials @xmath171 are fitted to the various integrated kinematical factors so that @xmath172 , with @xmath173 in the interval @xmath174 $ ] ( within our contour ) . here
@xmath175 where the function @xmath176 is @xmath177 35ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty http://arxiv.org/abs/1309.4146 [ ( ) ] link:\doibase 10.1103/physrevc.87.064302 [ * * , ( ) ] link:\doibase 10.1103/physrevc.84.014332 [ * * , ( ) ] link:\doibase 10.1103/physrevc.76.024318 [ * * , ( ) ] link:\doibase 10.1103/physrevc.80.044301 [ * * , ( ) ] link:\doibase 10.1103/physrevc.84.014314 [ * * , ( ) ] link:\doibase 10.1103/physrevc.84.041305 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.122501 [ * * , ( ) ] link:\doibase 10.1103/physrevc.65.054322 [ * * , ( ) ] link:\doibase 10.1103/physrevc.69.014316 [ * * , ( ) ] link:\doibase 10.1016/j.cpc.2005.01.001 [ * * , ( ) ] link:\doibase 10.1016/j.cpc.2013.01.013 [ * * , ( ) ] link:\doibase 10.1103/physrevc.81.024316 [ * * , ( ) ] link:\doibase 10.1103/physrevc.82.024313 [ * * , ( ) ] link:\doibase 10.1103/physrevc.85.024304 [ * * , ( ) ] http://arxiv.org/abs/1312.1746 [ `` , '' ] ( ) , link:\doibase 10.1103/physrevc.79.034310 [ * * , ( ) ] link:\doibase 10.1103/physrevc.86.031306 [ * * , ( ) ] link:\doibase 10.1103/physrevc.76.014312 [ * * , ( ) ] link:\doibase 10.1103/physrevc.80.064302 [ * * , ( ) ] @noop _ _ , ed .
( , ) @noop _ _ ( , ) @noop _ _ ( , ) @noop * * , ( ) link:\doibase 10.1103/physrevc.71.055501 [ * * , ( ) ] link:\doibase 10.1088/0034 - 4885/22/1/305 [ * * , ( ) ] link:\doibase 10.1103/physrevc.60.014302 [ * * , ( ) ] link:\doibase 10.1103/physrevc.67.064314 [ * * , ( ) ] link:\doibase 10.1016/0029 - 5582(62)90741 - 1 [ * * , ( ) ] link:\doibase 10.1103/physrevc.79.041301 [ * * , ( ) ] \doibase doi:10.1016/j.physletb.2009.03.077 [ * * , ( ) ] link:\doibase 10.1103/physrevc.83.054316 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _ , vol . | we describe and apply a version of the finite amplitude method for obtaining the charge - changing nuclear response in the quasiparticle random phase approximation .
the method is suitable for calculating strength functions and beta - decay rates , both allowed and forbidden , in axially - deformed open - shell nuclei .
we demonstrate the speed and versatility of the code through a preliminary examination of the effects of tensor terms in skyrme functionals on beta decay in a set of spherical and deformed open - shell nuclei . like the isoscalar pairing interaction , the tensor terms systematically increase allowed beta - decay rates .
this finding generalizes previous work in semimagic nuclei and points to the need for a comprehensive study of time - odd terms in nuclear density functionals . |
the single - particle shell model with strong central and spin - orbit potentials is a cornerstone of nuclear structure physics since about sixty years .
it was introduced by haxel , jensen and sue @xcite and goeppert - mayer @xcite in order to explain the experimentally observed occurrence of magic numbers for nuclei close to the valley of @xmath0-stability . in this picture , nucleons move as independent particles in a spherical mean - field .
the central potential usually has a shape similar to the nuclear density distribution and the spin - orbit potential is peaked at the nuclear surface .
the large spin - orbit potential explains the observed energy splittings between two spin - orbit partners with total angular momentum @xmath1 and parity @xmath2 for a given orbital angular momentum @xmath3 , see the left part of figure [ fig:01 ] .
this essential feature of the model generates the particular shell structure with the appearance of major shell closures at nucleon numbers 2 , 8 , 20 , 28 , 50 , 82 , 126 , and subshell closures at , e.g. , 14 , 40 , 114 .
the single - particle wave functions of the mean - field calculation can serve as a basis in order to construct fully antisymmetrized many - body wave functions . using a suitably chosen subset in the many - body space with an appropriate residual interaction , the excitation spectrum of a nucleus
can be described with high accuracy by diagonalizing the corresponding many - body hamiltonian , see , e.g. , ref .
@xcite .
experiments with exotic nuclei suggest a change of the well - known shell structure outside the valley of stability and the appearance of new ( sub-)shell closures often attributed to changes in the spin - orbit splitting or halo effects .
it is very important to understand the evolution of the single - particle energies with a change in isospin asymmetry and to identify the origin of this effect .
a change in the size of shell gaps can have important consequences in r - process nucleosynthesis @xcite . in recent years , contributions to the nucleon - nucleon potential that change the usual shell structure when
the isospin asymmetry of the system increases have been discussed extensively in the literature .
there are mainly two contributions to the nucleon - nucleon potential that have been considered : 1 . a central interaction of majorana - type @xmath4 ( spin - isospin interaction ) and 2 .
a non - central isospin - tensor interaction @xmath5 with the tensor operator @xmath6 , where the absolute strength of the potentials depends on the particular radial functions @xmath7 and @xmath8 with @xmath9 .
the monopole part of these interactions is found to be strongly attractive between protons and neutrons when the spins of the two nucleons are anti - aligned , i.e. , for @xmath10 and @xmath11 ( or @xmath12 and @xmath13 ) with @xmath14 , @xmath15 ( in case of the central spin - isospin interaction one has @xmath16 ) .
hence , the filling of one level with protons ( neutrons ) will lead to a strong shift of the corresponding neutron ( proton ) states changing the spin - orbit splittings and , depending on the involved levels , the size of the shell gaps . of particular importance
is the behaviour of the levels that define the conventional shell gaps .
it is interesting to note that for the ( sub-)shell closures at 14 , 20 , 28 , 50 , 82 , 114 and 126 at least one level defining the gap belongs a pair of so - called pseudospin partners .
these states are marked by solid circles in the central level scheme in figure [ fig:01 ] .
the total and orbital angular quantum numbers of the partner states are related to the so - called pseudo orbital angular momentum @xmath17 by @xmath18 , @xmath19 and @xmath20 , @xmath21 , respectively .
both levels have the same parity @xmath22 .
in addition , the principal quantum number @xmath23 of the state with @xmath18 is larger than the principal quantum number of the state with @xmath20 by one unit .
for example , the doublets @xmath24 with @xmath25 , @xmath26 and @xmath27 with @xmath25 , @xmath28 are pseudospin partners .
the indices @xmath29 and @xmath30 denote the spin - orbit and pseudospin representation of the states , respectively . many partner levels of pseudospin pairs are experimentally found to be almost degenerate in energy suggesting the existence of a particular symmetry in the mean - field hamiltonian for the single - particle states .
this feature has been noticed already more than forty years ago and the concept of pseudospin and pseudo - orbital angular momentum was introduced @xcite .
a breaking of the pseudospin symmetry in stable and exotic nuclei is potentially important for a change of the level structure and the appearance of new magic numbers .
it is important to identify the contributions to the nuclear interaction that are responsible for a breaking or the restoration of the pseudospin symmetry . to this end , a quantitative measure of the pseudospin splitting
is needed that can be used to compare and improve theoretical models .
bohr , hamamoto and mottelson @xcite discussed a transformation of the form @xmath31 that connects the angular momentum part of the pseudospin partner levels ( see eq.([eq : sigmadotr ] ) ) .
however , it can not account for the change of the principal quantum number @xmath23 .
hence , a dynamical symmetry involving the momentum operator @xmath32 is required . also it was observed that the su(3 ) and pseudo - su(3 ) states of the shell model are related by transformations that are elements of the orthosymplectic superalgebra osp(1/2 ) @xcite .
the many - particle pseudo - su(3 ) theory is successfully used in the description of heavy deformed nuclei in the context of the interacting boson model ( ibm ) applying supersymmetry , see , e.g. , ref .
eventually , the relevant momentum helicity operator @xmath33 was identified that accomplishes the pseudospin transformation and it was noticed that the dirac equation of free or in - medium nucleons ( in the dirac - brueckner approach ) is invariant under this transformation @xcite . the essential step in explaining the pseudospin symmetry was the application of a relativistic description . in this approach , e.g. , in the form of relativistic mean - field ( rmf ) models , the single - particle states are described by dirac spinor wave functions with large upper and small lower components .
ginocchio observed that the second - order differential equation for the lower component is approximately invariant with respect to the pseudospin symmetry transformation in rmf models with large scalar and vector potentials of nearly equal magnitude .
in addition , the radial wave functions of this lower component are almost identical for two pseudospin partner states @xcite .
the pseudospin symmetry was studied in various models for nuclear structure without and with deformation and in nucleon - scattering @xcite .
a symmetry breaking pseudospin - orbit potential can be derived from the wave equation of the lower component . in principle
, this potential should be a measure for the strength of the symmetry breaking .
it should appear as a small perturbation in comparison with the dominating central potential in a mean - field description of nuclei .
however , this potential exhibits a pole at finite radii and its position depends strongly on the level energy itself .
therefore , one can not really estimate the strength of the splitting and test how the choice of the various contributions to the interaction influences the symmetry restoration or breaking as observed experimentally .
the problems related to this peculiarity have been discussed in the literature , see , e.g. , ref .
@xcite .
besides the single - particle levels that constitute pseudospin pairs there are several states that are ( pseudo-)unpaired in a nucleus .
there are all the @xmath34 states corresponding to @xmath35 irrespective of the principal quantum number @xmath23 and the states for given pseudospin @xmath36 with ( pseudo ) principal quantum number @xmath37 ( @xmath38 ) .
e.g. , for the pseudoangular momentum @xmath26 the @xmath39 level has no pseudospin partner whereas for @xmath40 the pairs @xmath41 appear .
see the right part of figure [ fig:01 ] .
the latter feature is a typical observation for a system of levels respecting exact supersymmetry .
it is therefore tempting to apply supersymmetric quantum mechanics to the nuclear single - particle shell model and to go beyond the harmonic oscillator case @xcite .
as will be shown below , the approach employing supersymmetry allows to extract a pseudospin breaking potential that is regular in contrast to the standard relativistic description . from the condition of vanishing perturbation potential
, general potentials can be derived that exhibit exact pseudospin symmetry . in order to compare the standard description of pseudospin symmetry with the supersymmetric approach
, the relativistic hamiltonian will be used as a starting point in the theoretical considerations below . in principle , however , the supersymmetric approach can also be applied to non - relativistic hamiltonians .
the paper is organized as follows . in section
[ sec : rmf ] the basic equations for describing single - particle states in relativistic models are given and the notation for quantum numbers , wave functions , potentials and hamiltonians is established .
the conventional description of pseudospin symmetry in the relativistic approach is discussed in section [ sec : ps ] and a simple example is introduced that will accompany the following theoretical presentation . in section [ sec : susy ] , first the concept of supersymmetric quantum mechanics is presented as far as necessary to solve the problem considered here .
more detailed accounts can be found , e.g. , in refs .
then the supermomenta and supersymmetric partner hamiltonians are defined .
the form of the reduced supermomenta at small radii leads to a particular choice of energy shifts that relate the pseudospin partner hamiltonians of the relativistic upper component wave functions to the partner hamiltonians in the supersymmetric approach .
then , the connection between the wave functions in the standard relativistic and the supersymmetric description is given .
a regular symmetry breaking potential is derived that quantifies the pseudospin splitting .
it can be used in the future to compare and improve nuclear structure calculations . in the final subsection ,
general potentials in relativistic hamiltonians with exact pseudospin symmetry are discussed .
the last section contains conclusions and an outlook .
the motion of the nucleons in rmf models is governed by strong scalar and vector potentials ( or self - energies ) originating from the minimal coupling of scalar and vector mesons to the nucleons .
in addition , a tensor potential can appear if vector mesons are coupled non - minimally to the meson field tensors , i.e. by a derivative coupling .
in general , there are both isoscalar and isovector contributions to the nucleon self - energies .
the details of the particular rmf model are not important in the present discussion .
see , e.g. , references @xcite for a more extensive presentation . here , it suffices to consider only the calculation of the single - particle states for given scalar , vector and tensor potentials in the dirac equation . in the following spherical symmetry will be assumed . in this case , single - particle levels are characterized by the principal quantum number @xmath42 , the orbital angular momentum @xmath43 , the parity @xmath44 , the total angular momentum @xmath45 , and its projection @xmath46 . in the relativistic description
it is convenient to combine the quantum numbers @xmath3 and @xmath47 by introducing the quantity @xmath48 where @xmath49 and @xmath50 see table [ tab : kjl ] for the relation between the various quantum numbers .
levels with different @xmath51 for given @xmath23 and @xmath52 are degenerate in energy .
a single - particle state @xmath53 with energy @xmath54 ( including the rest mass @xmath55 ) of a nucleon is obtained by solving the time - independent dirac equation @xmath56 where @xmath57 - i \beta \vec{\alpha } \cdot \vec{t}(r)\ ] ] is the relativistic hamiltonian with momentum operator @xmath32 and dirac matrices @xmath58 and @xmath0 in standard notation .
the system of units is chosen such that @xmath59 .
the scalar , vector and tensor potentials are denoted by @xmath60 , @xmath61 and @xmath62 . here , the sign convention with a positive scalar field @xmath63 is used . for spherical symmetry
it is customary to write the dirac spinor as @xmath64 with radial wave functions @xmath65 and @xmath66 in the upper and lower components , respectively , that are normalized according to @xmath67 = 1 \ : .\ ] ] the spin and angular dependence is contained in the spinor spherical harmonics @xmath68 where the orbital angular momentum @xmath69 of the spherical harmonic @xmath70 is coupled with the nucleon spin @xmath71 of the spinor @xmath72 to the total angular momentum @xmath73 .
the upper and lower component in ( [ eq : psi_dirac ] ) have an opposite sign for the quantum number @xmath52 in the spinor spherical harmonics and consequently opposite parity with a difference of one in the orbital angular momentum . by applying the hamiltonian ( [ eq : h_dirac ] ) to the dirac spinor ( [ eq : psi_dirac ] ) and using the relations @xmath74 and @xmath75 the set of coupled first - order differential equations @xmath76 for the radial wave functions is obtained .
a simple decoupling leads to the independent wave equations @xmath77 with schrdinger - like hamiltonians @xmath78 \\ \nonumber & & +
v - s + m\end{aligned}\ ] ] and @xmath79 \\ \nonumber & & + v - s + m \ : , \end{aligned}\ ] ] respectively , that explicitly depend on the quantum number @xmath52 .
a prime denotes the derivative with respect to the radius @xmath80 .
the quantities @xmath81 and @xmath82 depend on the single - particles energy @xmath83 appearing as the eigenvalue in equations ( [ eq : hf ] ) and ( [ eq : hg ] ) .
the hamiltonians @xmath84 and @xmath85 have the same spectrum , even though the potentials have a different form .
@xmath86 can be considered as twice the position depending effective nucleon mass @xmath87 .
it is always positive in standard rmf models .
the dependence on the energy @xmath54 is a relativistic effect .
since @xmath54 contains the rest mass @xmath55 of the nucleon , the variation of @xmath86 with a change in the nucleon binding energy is very small . for vanishing scalar and vector fields @xmath60 and @xmath61 at large radii @xmath80
, @xmath88 reduces to the non - relativistic energy @xmath89 of the bound nucleon .
the hamiltonian ( [ eq : hf ] ) for the upper component wave function @xmath90 can be written in the familiar non - relativistic form @xmath91 with central and spin - orbit potentials @xmath92\end{aligned}\ ] ] since @xmath93 and @xmath94 .
the spin - orbit potential contains contributions from the tensor potential @xmath95 and the derivative of the sum @xmath96 whereas the main contribution to the central potential arises from the difference @xmath97 . with the typical strong vector and scalar potentials of almost similar magnitude inside nuclei , the usual depth of the central potential and the strong surface - peaked spin - orbit potential
are nicely explained in the relativistic approach .
a tensor potential is usually omitted in traditional rmf calculations but it has a distinct effect on the spin - orbit splitting that can be relevant for the size of the shell gaps .
the basic idea in the relativistic description of pseudospin symmetry is a comparison of the hamiltonians @xmath85 , @xmath98 and of the radial wave functions @xmath99 , @xmath100 for the two pseudospin partner levels of the doublet @xmath101 .
it is easy to see that for every state with @xmath102 the pseudospin partner state has the quantum number @xmath103 .
the corresponding orbital angular momenta are @xmath104 and @xmath105 , respectively .
similarly , the total angular momenta are given by @xmath106 and @xmath107 , respectively .
hence , the relation between @xmath17 and @xmath52 is given by @xmath108 the single - particle states in the relativistic model with @xmath109 , i.e. the @xmath110 states , have no pseudospin partner . in this case
see table [ tab : kjl ] for a summary of the relations for the quantum numbers .
a simple example shows the main features in the relativistic description of pseudospin partner states . here , the lowest pair of levels with @xmath26 , i.e.
, @xmath112 will be considered .
the shape of scalar and vector potentials in rmf models is often well approximated by a woods - saxon form , i.e. , @xmath113 and @xmath114 with @xmath115\}$ ] . assuming a radius of @xmath116 fm , a diffuseness of @xmath117 fm and absolute magnitudes of @xmath118 mev and @xmath119 mev , respectively , ( corresponding to the the mean - field potentials of neutrons in a nucleus like @xmath120ca )
one finds single - particles energies of @xmath121 mev and @xmath122 mev with the @xmath123 level lying only @xmath124 mev above the @xmath125 level .
the corresponding wave functions of the upper and of the lower component in the dirac spinor are depicted in the top and bottom panel of figure [ fig:02 ] .
the wave functions @xmath126 and @xmath127 are very different where the former has an additional node because there exists a stronger bound @xmath39 state at @xmath128 mev without node ( except at @xmath129 ) . in contrast , the wave functions @xmath130 and @xmath131 resemble each other closely , a sign for the almost perfect pseudospin degeneracy of the two states . due to the different orbital angular momenta and corresponding centrifugal potentials in @xmath84 and @xmath132
there is a clear difference in the radial dependence of the upper component wave functions at small @xmath80 .
but for the lower component radial wave function the @xmath80 dependence is very similar .
in fact , the orbital angular momenta in the hamiltonians @xmath85 and @xmath133 of a pseudospin pair @xmath134_{so}$ ] for @xmath135 are identical since @xmath136 . the centrifugal potential @xmath137 in ( [ eq : hg2 ] ) is invariant with respect to the replacement @xmath138 .
the difference between the potentials @xmath139 could be considered as a measure for the pseudospin splitting because for @xmath140 the hamiltonians @xmath85 and @xmath133 are identical and yield the same energy eigenvalues
. however , there is a problem in this approach since the quantity @xmath88 has a zero at finite @xmath80 inside the nucleus with usual scalar and vector potentials in rmf models .
this is most easily seen in figure [ fig:03 ] where the potential difference @xmath141 and the non - relativistic energy @xmath142 are plotted .
they are identical at a certain radius @xmath80 and the quantity ( [ eq : defb ] ) is zero .
consequently , the difference potential ( [ eq : delta1 ] ) , shown as a red solid line in figure [ fig:03 ] , has a pole .
its position is strongly energy dependent .
thus , @xmath143 can not be considered small as compared to the potential @xmath97 in the differential equation and it is therefore not a good measure for the symmetry breaking . for vanishing tensor potential @xmath144 ,
the pseudospin - orbit potential ( [ eq : delta1 ] ) is zero for @xmath145 or @xmath146 with a constant @xmath147 . as a consequence of this condition ,
the main contribution @xmath97 to the central potential ( [ eq : vc ] ) is just an overall shift of the energy scale for exact pseudospin symmetry .
but then there will be no sufficiently strong central potential to bind the states as required in a nucleus .
a case where exact pseudospin symmetry is found is the relativistic harmonic oscillator with @xmath148 and @xmath144 @xcite .
the effective mass @xmath149 is a constant , @xmath150 but @xmath151 and @xmath152 .
hence , the identification of @xmath153 as the relevant quantity to measure the pseudospin splitting is doubtful
. it would be convenient to have an independent measure of the symmetry breaking that has a regular behaviour for all radii and that vanishes for exact pseudospin symmetry .
the problems related to the conventional relativistic explanation of pseudospin symmetry call for an alternative approach that allows to derive a regular symmetry breaking potential which is a small perturbation without a pole and useful for a quantitative comparison of different models .
the main idea of the supersymmetric approach to describe the pseudospin symmetry in nuclei is similar to the standard relativistic description . in both cases
one does not compare the original model hamiltonians of the pseudospin partner systems but closely related hamilton operators with the same spectra .
in fact , it will be shown below that a comparison of the supersymmetric partner hamiltonians provides a way to define a proper pseudospin splitting potential .
these partner potentials are derived from the hamiltonians @xmath84 and @xmath132 for the upper component wave functions of the pseudospin partners @xmath52 and @xmath103 , respectively . to this end
, the hamiltonian ( [ eq : hf2 ] ) is written as @xmath154 with a hamiltonian @xmath155 of a supersymmetric pair and an energy shift @xmath156 to be determined below .
a corresponding relation holds for the pseudospin partner hamiltonian @xmath132 .
it is well known , that every second - order hamiltonian can be factorized in a product of two hermitian conjugate first - order operators @xcite . in the present case ,
operators @xmath157 and @xmath158^{\dagger}$ ] are introduced such that @xmath159 then , the hermitian operators @xmath160 are formed that are so - called supercharges with respect to the involution @xmath161 because @xmath162 . in the next step ,
the supersymmetric hamiltonian @xmath163^{2 } = \left[q_{2}(\kappa)\right]^{2 } = \left ( \begin{array}{cc } h_{1}(\kappa ) & 0 \\ 0 & h_{2}(\kappa ) \end{array } \right)\ ] ] is obtained with the supersymmetric partner hamiltonians ( [ eq : h1def ] ) and @xmath164 on the diagonal .
obviously , @xmath165 commutes with @xmath166 and @xmath167 , i.e. , @xmath168 = \left [ h_{s}(\kappa ) , q_{2}(\kappa ) \right ] = 0\ ] ] and the supercharges @xmath169 ( @xmath170 ) are generators of the symmetry transformation .
the supercharges ( [ eq : supercharge ] ) and the hamiltonian ( [ eq : hs ] ) with the commutators ( [ eq : alg1 ] ) and the anticommutator @xmath171 are the most simple example of a supersymmetric algebra . since @xmath165 is the square of the hermitian operators @xmath169 , all eigenvalues @xmath172 of the eigenvalue equation @xmath173 with the two - component wave function @xmath174 are non - negative .
it is easily seen by applying @xmath169 to equation ( [ eq : hsev ] ) , that @xmath155 and @xmath175 have the same spectrum of positive energies @xmath176 where the operators @xmath157 and @xmath177 connect the components of the wave function ( [ eq : psi_susy ] ) by the transformations @xmath178 if there is an eigenstate @xmath179 ( @xmath180 ) with energy @xmath181 , the supersymmetry is called exact because @xmath182 and this ground state obeys the symmetry of the hamiltonian @xmath165 . in this case ,
@xmath183 or @xmath184 i.e. , the hamiltonian @xmath155 or @xmath175 has an additional state at zero energy that is not appearing for its supersymmetric partner hamiltonian . in this paper ,
the usual convention is chosen such that @xmath155 has a ground state at zero energy if the supersymmetry is exact .
if there is no state @xmath179 at energy @xmath181 , the supersymmetry is called broken since @xmath185 for all @xmath186 and the partner hamiltonians @xmath155 and @xmath175 have identical spectra . the main task in the supersymmetric description of pseudospin symmetry is the construction of the operators @xmath157 and @xmath177 .
the particular form of the hamiltonian ( [ eq : hf2 ] ) suggests to use the ansatz @xmath187
\qquad b_{\kappa}^{+ } = \left [ q_{\kappa } ( r ) - \frac{d}{dr } \right ] \frac{1}{\sqrt{a_{\kappa}(r)}}\ ] ] with ( the principal quantum number @xmath23 is suppressed in the following if not explicitly needed ) @xmath188 and a quantity @xmath189 that will be called supermomentum by dimensional reasons .
the traditional superpotential is given by @xmath190 however , it has the dimension energy@xmath191 .
the supersymmetric partner hamiltonians are found as @xmath192\ ] ] and @xmath193\ ] ] with the symmetry @xmath194 if the supersymmetry is exact , the supermomentum @xmath195 can be found from the ground state wave function @xmath196 of the hamiltonian @xmath155 as @xmath197 because of equation ( [ eq : bkm0 ] ) , i.e. , it is given by the negative logarithmic derivative of the groundstate wave function .
comparing equation ( [ eq : h1 ] ) via ( [ eq : hfh1 ] ) with equation ( [ eq : hf ] ) leads to the defining equation for the supermomentum @xmath189 in general .
however , it is useful to split off the dependence on @xmath52 and the tensor potential @xmath95 first and to introduce the reduced supermomentum @xmath198 with the results @xmath199\end{aligned}\ ] ] and @xmath200 \ : .\end{aligned}\ ] ] these two hamiltonians show the symmetry @xmath201 however , it is irrelevant for the present considerations .
subsequently , the defining equation for the reduced supermomentum @xmath202 with @xmath203 is obtained .
note that @xmath204 depends on the energy shift @xmath156 and @xmath205 on the eigenvalue @xmath206 .
equation ( [ eq : ric ] ) is the central relation that connects the reduced supermomentum @xmath207 with the relativistic potentials @xmath60 , @xmath61 and @xmath208 . in general , the reduced supermomentum @xmath207 is not uniquely determined by equation ( [ eq : ric ] ) and suitable boundary conditions have to be specified . in the present case , it will be required that @xmath209 for @xmath129 .
it is convenient , e.g. , in numerical applications , to transform the riccati equation ( [ eq : ric ] ) , an inhomogeneous nonlinear first - order differential equation , into a linear second - order differential equation to determine the reduced supermomentum @xmath207 for given scalar , vector and tensor potentials @xmath60 , @xmath61 and @xmath208 . to this end
, a new function @xmath210 is introduced that is related via @xmath211 to the reduced supermomentum @xmath207 . with
@xmath212 one immediately obtains the homogeneous differential equation @xmath213 that can be solved with standard techniques .
using equation ( [ eq : ric ] ) the supersymmetric partner hamiltonian ( [ eq : h2p ] ) assumes the form @xmath214 - b_{\kappa } \ : .\end{aligned}\ ] ] it will be useful for the comparison with the hamiltonian @xmath215 of the pseudospin partner system .
the asymptotic behaviour of the reduced supermomentum for @xmath216 and @xmath217 is easily derived from equation ( [ eq : ric ] ) assuming regular @xmath207 and regular potentials @xmath60 , @xmath61 and @xmath208 .
the reduced superpotential has to vanish for @xmath216 otherwise a diverging potential would appear in the hamiltonian @xmath84 . for small radii
the reduced supermomentum is given by @xmath218 i.e. , it increases almost linearly with the radius . at large radii
it becomes a constant @xmath219 } + o(r^{-1})\ ] ] for potentials @xmath60 , @xmath61 and @xmath208 that approach zero at large radii @xmath80 . the energy shift @xmath156 with the condition
@xmath220 has still to be determined .
the asymptotic form ( [ eq : pkinf ] ) translates to a dependence @xmath221 } \ : r \right\}\ ] ] of the auxiliary function @xmath222 in equation ( [ eq : ykd ] ) for @xmath217 .
starting at large radii and integrating inwards a stable solution for @xmath223 is found .
of course , if the relativistic potentials do not approach zero at large radii , a different asymptotic behaviour of the reduced supermomentum will be obtained .
considering the relation ( [ eq : pp ] ) , the full supermomentum @xmath189 approaches the same constant as @xmath207 for large @xmath80 .
in contrast , the asymptotic behaviour of @xmath189 for @xmath216 is determined by the angular momentum term @xmath224 . for @xmath225 ,
one finds @xmath226 and @xmath227 for @xmath228 .
the standard superpotential ( [ eq : wk ] ) shows the same asymptotic limits as the full supermomentum . from general considerations of supersymmetric quantum mechanics
it is known , that the type of supersymmetry is determined by the asymptotics of the superpotential . if there is a change of sign in @xmath229 when comparing the limits @xmath230 with @xmath217 , exact supersymmetry follows and a single non - degenerate state at zero energy exists .
conversely , if the superpotential @xmath229 does not change the sign in the above limits the supersymmetry is broken and all eigenstates are doubly degenerate with positive energy . with the foregoing considerations ,
the supersymmetric description can be applied to the system of pseudospin levels appearing in nuclei . for given pseudo orbital angular momentum @xmath231
there is an unpaired state with total angular momentum @xmath232 , orbital angular momentum @xmath233 and energy @xmath234 , cf . the left part of figure [ fig:04 ] . choosing the energy shift as @xmath235 the ground state energy of the hamiltonian @xmath236
is placed at zero energy and the case of exact supersymmetry is obtained .
the supersymmetric partner hamiltonian @xmath175 has the same energy eigenvalues as @xmath155 except for the zero - energy level .
the orbital angular momentum in @xmath175 is @xmath237 as seen in the centrifugal potential in equation ( [ eq : h2p ] ) .
the pseudospin partner levels with @xmath238 and @xmath239 are eigenstates of the hamiltonian @xmath132 with @xmath240 , see the right part of figure [ fig:04 ] .
the lowest state has a larger energy @xmath241 than the ground state of @xmath84 . for this system the case of broken supersymmetry applies where @xmath242 and the supersymmetric partner hamiltonian @xmath243 have identical eigenvalues . the orbital angular momentum in @xmath215 is again given by @xmath244 due to @xmath245
the energy shift @xmath246 is now chosen such that the asymptotic form of the reduced supermomentum @xmath247 is the same as for the reduced supermomentum @xmath207 in the limit @xmath216 .
this condition leads with equation ( [ eq : pr0 ] ) to the relation @xmath248\ ] ] since @xmath249 . for exact pseudospin symmetry , @xmath250 and
consequently @xmath251 .
even if the pseudospin symmetry is broken , @xmath252 and @xmath253 are almost identical and the difference can be neglected as a small relativistic correction . without the contribution of the tensor potential @xmath254 ,
the energy shift @xmath246 is determined by @xmath255 \ : .\ ] ] in the top panel of figure [ fig:05 ] the reduced supermomenta @xmath207 and @xmath247 of the pseudospin pair levels with @xmath256 and @xmath257 are shown for the example in section [ sec : ps ] . with the present choice of the energy shifts @xmath258 mev and @xmath259 mev , the reduced supermomenta display the same approximately linear behaviour for @xmath216 as required .
in fact , they are almost identical for radii below 5 fm . at large @xmath80
they become constant and approach different values for @xmath217 as predicted by equation ( [ eq : pkinf ] ) with @xmath260 because @xmath261 .
the full supermomenta @xmath262 and @xmath263 are shown in the bottom panel of figure [ fig:05 ] .
they contain the angular momentum contribution and , hence , they diverge for @xmath216 .
the change of sign for @xmath262 indicates the case of exact supersymmetry with a single state at zero energy for @xmath264 .
in contrast , there is no change of sign for @xmath263 corresponding to the case of broken supersymmetry .
the supersymmetric description of pseudospin degenerate states allows to establish relations between the wave functions in the dirac spinor ( [ eq : psi_dirac ] ) and in the supersymmetric state vector ( [ eq : psi_susy ] ) . according to equation ( [ eq : cplg ] )
the lower component wave function of the spinor is given by @xmath265 for the pseudospin partner state @xmath266 with @xmath225 and similarly for the state @xmath267 with @xmath268 and @xmath103 .
using equation ( [ eq : bpsi ] ) the supersymmetric partner @xmath269 a_{n\kappa } } } \left ( q_{n\kappa } - \frac{\kappa}{r } - t + \frac{d}{dr}\right ) \psi_{1}(n\kappa , r)\ ] ] of the upper component wave function @xmath270 is obtained where the eigenenergy @xmath172 in the supersymmetric formulation is identified with the energy @xmath271 . with equation ( [ eq : gk2 ] )
one finds @xmath272\ ] ] and @xmath273\ ] ] for the two pseudospin partner states .
these functions are normalized as @xmath274 and correspondingly for the second pair . in the case of exact pseudospin symmetry , the wave functions @xmath275 and @xmath276 are identical provided that the upper component wave functions @xmath90 and @xmath277 in the dirac spinors are normalized to the same value .
this is beautifully seen in figure [ fig:06 ] for the example of section [ sec : ps ] with a much better agreement of the two pseudospin partner wave functions than in the standard approach , cf .
figure [ fig:02 ] .
in contrast to the wave functions @xmath278 and @xmath279 , there is no zero in the functions @xmath275 and @xmath276 because they are the lowest states of the supersymmetric partner hamiltonians @xmath175 and @xmath215 with the same orbital angular momentum @xmath280 .
the factors @xmath281 and @xmath282 in equations ( [ eq : psi2 ] ) and ( [ eq : psi2p ] ) are very small at small radii @xmath80 explaining the observation of the close similarity of the wave functions @xmath278 and @xmath279 in the usual comparison . at larger radii
, however , the contribution from the upper component wave functions @xmath90 and @xmath277 , respectively , becomes more important .
the hamiltonians @xmath283 and @xmath243 of the supersymmetric partners for the pseudospin partner levels are almost identical .
the difference is given by the potential @xmath284 \\ \nonumber & = & \frac{2}{\sqrt{a } } \frac{d}{dr } \frac{q_{\kappa}-q_{\kappa^{\prime}}}{\sqrt{a } } + 2 t \frac{2\kappa+1}{ar}\end{aligned}\ ] ] with @xmath285 and @xmath286 where equation ( [ eq : h2pp ] ) has been used and the difference between @xmath252 and @xmath253 has been neglected . the quantity @xmath287 is a regular function for all radii @xmath80 with the same contribution from the tensor potential @xmath208 as the previously defined pseudospin - orbit potential @xmath153 in ( [ eq : delta1 ] ) .
however , the pseudospin breaking potential ( [ eq : delta2 ] ) is a regular function for all @xmath80 without a pole as observed for @xmath153 . without tensor potential ,
the pseudospin symmetry breaking potential depends on the derivative containing the reduced supermomenta of the pseudospin partner systems instead on the derivative of the difference @xmath97 as in the traditional approach ( [ eq : delta1 ] ) . from figure [ fig:03 ]
it is evident that @xmath287 with @xmath26 for the example in section [ sec : ps ] is really small as compared to @xmath141 and can be considered as a perturbation . in a relativistic model with tensor potential @xmath208 , both the non - relativistic spin - orbit potential ( [ eq : vso ] ) and the pseudospin symmetry breaking potential ( [ eq : delta2 ] )
are directly affected by @xmath208 with the same functional form , only the prefactors @xmath288 and @xmath289 are different .
the dependence of @xmath287 on the scalar and vector potentials @xmath60 and @xmath61 is , however , more subtle through the riccati equation for the reduced supermomenta . in general , there are three potentials , @xmath60 , @xmath61 and @xmath208 , in the relativistic hamiltonian ( [ eq : h_dirac ] ) . with one constraint @xmath290
to achieve exact pseudospin symmetry , only two independent functions of these potentials can be specified freely ( subject to some conditions ) . since the difference potential ( [ eq : delta2 ] )
is a function of @xmath86 ( containing the sum @xmath291 ) and @xmath208 , it is reasonable to use these quantities as independent variables and to determine the reduced supermomenta , @xmath207 and @xmath247 and finally the scalar and vector potentials , @xmath60 and @xmath61 .
this approach will be followed in the considerations below . as an alternative ,
the potentials in the relativistic hamiltonian could be derived from two given supermomenta @xmath207 and @xmath247 . in order to simplify the notation ,
the quantities @xmath292 are introduced replacing the energy shifts @xmath156 and @xmath246 . from the condition @xmath293 one obtains the general solution for the difference @xmath294 with the function @xmath295 where @xmath296 is an integration constant .
writing @xmath297 the difference of the defining equations ( [ eq : ric ] ) for @xmath207 and @xmath247 leads to the solution @xmath298 = \frac{1}{2 } \left [ % a_{\kappa } \omega % \frac{e(\kappa)- e(\kappa^{\prime})}{2\kappa+1 } \frac{a \omega r}{1+cr } - \frac{d}{dr } \ln ( 1 + c r ) \right]\ ] ] for the auxiliary momentum @xmath299 .
the sum of the two defining equations gives in combination with the definition of @xmath86 the scalar and vector potentials @xmath300 + m \\ \label{eq : resv } v(r ) & = & \frac{1}{2 } \left [ e + e_{0 } - a(r ) + u(r ) \right]\end{aligned}\ ] ] with the auxiliary potential @xmath301 \end{aligned}\ ] ] that is invariant under the replacement @xmath302 .
for the determination of the potentials @xmath60 and
@xmath61 in the relativistic hamiltonian that exhibits exact pseudospin symmetry , the given functions @xmath86 and @xmath208 have to be regular , but not every choice is admissible . at least , one has to require that @xmath303 for @xmath304 and that @xmath208 is selected such that the integral in equation ( [ eq : cdef ] ) is finite for all @xmath80 . in addition , there is the constraint that @xmath305 has to be zero at @xmath129 because @xmath306 by construction of the supersymmetric partner hamiltonians and that @xmath307 , otherwise the momentum @xmath299 is not regular for all radii .
it is worthwhile to note that the potential ( [ eq : defw ] ) depends on the quantum number @xmath52 if @xmath308 and that exact pseudospin symmetry is found only in the particular doublet of states with given @xmath309 and @xmath310 if the relativistic potentials are assumed to be identical for all states . it is instructing to discuss the resulting potentials for a few particular choices of @xmath86 and @xmath208 .
there are two major cases depending on the value of @xmath86 at @xmath129 .
1 . @xmath311 but @xmath312 if @xmath304 . in this case , which might not be realized in a physical system , the integration constant @xmath296 in equation ( [ eq : cdef ] ) can have any non - negative value as long as @xmath307 for all @xmath80 .
if @xmath305 is not identically zero , the reduced supermomenta of the pseudospin partner system with different @xmath17 are different and the potential @xmath313 explicitly depends on @xmath52 .
this can occur even for vanishing tensor potential if @xmath314 .
@xmath315 , the case usually encountered in physical systems .
it follows that @xmath316 and that @xmath308 only if there is a tensor interaction .
see the discussion in the paragraph above . in most relativistic hamiltonians without a tensor interaction and everywhere finite effective mass of the nucleon there is still a large variation of potentials that exhibit exact pseudospin symmetry . for @xmath317 and then @xmath318 for all radii @xmath80 , the reduced supermomenta of the pseudospin partners are identical with @xmath319 and the auxiliary potential becomes @xmath320 = \frac{a(r)}{4 } \omega^{2 } r^{2 } \ : , \ ] ] i.e. , a modified harmonic oscillator potential with position depending effective mass . with the scalar and vector potentials ( [ eq : ress ] ) and ( [ eq : resv ] ) , respectively , the central and and spin - orbit potentials ( [ eq : vc ] ) and ( [ eq : vso ] ) are found as @xmath321 in the non - relativistic hamiltonian ( [ eq : hfnr ] ) .
the standard harmonic oscillator is obtained for constant @xmath322 , @xmath317 and @xmath323 such that @xmath324
. then the scalar and vector potentials are @xmath325 leading to the non - relativistic central and spin - orbit potentials @xmath326 in general , the choice of the position depending effective mass @xmath327 has an important effect on the potentials that appear in the non - relativistic hamiltonian .
however , a reasonable effective mass should approach the free nucleon mass @xmath55 for large radii @xmath80 .
then the potentials will approach the harmonic oscillator form in the limit @xmath217 .
the occurrence of almost degenerate pseudospin partner levels in atomic nuclei is usually attributed to a symmetry of the underlying relativistic hamiltonian as it appears , e.g. , in rmf models .
the lower - component radial wave functions in the dirac spinors are found to be very similar . from the differential equation of these wave functions a pseudospin - orbit potential is extracted that breaks the pseudospin symmetry .
however , it is not a regular function and it can not be considered as a small perturbation .
in addition , this potential does not vanish for the usual harmonic oscillator potential which exhibits exact pseudospin symmetry . in the supersymmetric approach to pseudospin symmetry ,
the supersymmetric partner hamiltonians of the upper - component wave functions are compared .
it leads to a regular symmetry breaking potential that is small in comparison with the dominating central potential in the mean - field hamiltonian .
the corresponding partner wave functions are identical in the case of exact pseudospin symmetry . employing both exact and broken supersymmetry , all levels in a sequence of states for given pseudo orbital angular momentum @xmath36 , including the lowest unpaired state , fit into the scheme .
this lowest level defines the reduced supermomentum and consequently the potentials in the hamiltonians for both pseudospin partner systems in the case of exact pseudospin symmetry .
all higher lying pairs are degenerate in energy . from the condition of vanishing symmetry breaking potential , general relativistic potentials that exhibit exact pseudospin symmetry
can be derived .
modified harmonic oscillator potentials with position depending effective mass are the most simple class .
the present description relies on the differential equations for the radial wave functions and the corresponding first - order differential operators in the factorization of the hamiltonian for given quantum number @xmath52 .
it is also possible to generalize the approach to a form independent of a particular @xmath52 . in this case
the operators @xmath328 \frac{1}{\sqrt{a } } \qquad b^{- } = \frac{1}{\sqrt{a } } \left [ q - \vec{\sigma } \cdot \left ( \vec{t } + i \vec{p } \right ) \right]\ ] ] in the generators ( [ eq : supercharge ] ) of the symmetry transformation act on the full wave functions in coordinate space . these and more formal aspects of the supersymmetric approach can be studied in the future . the symmetry breaking potential depends on derivatives of the supermomenta for the two pseudospin partner states and the tensor potential .
the supermomenta itself are determined via a riccati equation by the scalar , vector and tensor potentials in the relativistic hamiltonian .
a simple example shows that the reduced supermomenta of pseudospin partner states are very similar rising approximately linear at small radii as in the case of the harmonic oscillator .
it seems that all saturating potentials with a flat bottom will show an approximate pseudospin symmetry .
this question can be explored in more detail by expanding the reduced supermomenta in a power series in terms of the radius @xmath80 .
calculating the supermomenta for actual selfconsistent rmf models will help to quantify the various sizes of the observed pseudospin symmetry breaking .
the study of the symmetry breaking potential will eventually lead to improved rmf parametrizations since it can help to identify the relevant modifications of the model that are required to obtain a better description of the experimentally observed breaking and restoration of the pseudospin symmetry in nuclei
. it will be interesting to see how the isospin dependence of the effective interaction changes the degree of pseudospin symmetry breaking .
of special importance is the tensor interaction ( neglected in most rmf models ) that affects both the spin - orbit and the pseudospin - orbit splitting .
the method that is developed in this paper is not restricted to a relativistic description of the single - particle states . in principle , it can also be applied to the usual schrdinger equation for the single - particle wave functions as it appears , e.g. , in the skyrme hartree - fock method , because it relies only on the schrdinger - equivalent differential equation of the upper component in the dirac spinor . instead of the scalar ,
vector and tensor potentials in the relativistic hamiltonian , there are also at least three independent quantities that characterize the mean - field in the non - relativistic hamiltonian : the position depending effective mass , the central and the spin - orbit potential .
the relation between these functions and the reduced supermomentum will be different as compared to the relativistic description , however , the same principles of the supersymmetric method can be applied .
the author is grateful to o. sorlin and d. lacroix for many stimulating discussions and a critical reading of the manuscript .
pseudospin symmetry was found as a particular limit in the application of supersymmetry to the dirac equation in refs .
[ lev04,lev05 ] by directly factorizing the dirac hamiltonian .
00 o. haxel , j. h. d. jensen and h. e. sue , naturwissenschaften 35 ( 1948 ) 376 .
h. e. sue , o. haxel and j. h. d. jensen , naturwissenschaften 36 ( 1949 ) 153 j. h. d. jensen , h. e. sue and o. haxel naturwissenschaften 36 ( 1949 ) 155 .
o. haxel , j. h. d. jensen and h. e. suess , phys .
rev . 75 ( 1949 ) 1766 .
m. goeppert mayer , phys .
rev . 75 ( 1949 ) 1969 . e. caurier , g. martnez - pinedo , f. novacki , a. poves and a. p. zuker , rev .
77 ( 2005 ) 427 .
d. guillemaud - mueller et al . , nucl .
phys . a 426 ( 1984 ) 37 .
a. huck et al .
, phys . rev .
c 31 ( 1986 ) 2226 .
h. keller et al . , z. phys . a 348 ( 1994 ) 61 .
t. motobayashi et al . , phys .
b 346 ( 1995 ) 9 . h. scheit et al . , phys .
77 ( 1996 ) 3967 .
t. glasmacher et al .
b 395 ( 1997 ) 163 . h. simon et al . , phys .
83 ( 1999 ) 496 .
a. ozawa , t. kobayashi , t. suzuki , k. yoshida and i. tanihata , phys .
84 ( 2000 ) 5493 h. grawe and m. lewitowicz , nucl .
a 693 ( 2001 ) 116 j. i. prisciandaro et al .
b 510 ( 2001 ) 17 d. sohler et al . , phys . rev .
c 66 ( 2002 ) 054302 .
r. v. f. janssens et al .
lett b 546 ( 2002 ) 55 .
z. dlouh , d. baiborodin , j. mrzek and g. thiamov , nucl .
. a 722 , ( 2003 ) c36 j. p. schiffer , et al .
92 ( 2004 ) 162501 .
dinca et al . , phys .
c 71 ( 2005 ) 041302 .
b. fornal et al . , phys .
c 72 ( 2005 ) 044315 .
h. grawe et al . , eur .
j. a 25 ( 2005 ) 357 .
l. gaudefroy et al . , phys .
97 ( 2006 ) 092501 ; 99 ( 2007 ) 099202 .
m. rejmund et al . , phys .
c 76 ( 2007 ) 021304 .
a. gade and t. glasmacher , prog .
. phys . 60
( 2008 ) 161 b. chen , j. dobaczewski , k .-
kratz , k. langanke , b. pfeiffer , f .- k .
thielemann and p. vogel , phys .
b 355 ( 1995 ) 37 .
j. m. pearson , r. c. nayak and s. goriely , phys .
b 387 ( 1996 ) 455 . k. -l .
kratz , b. pfeiffer and f .- k .
thielemann , nucl .
phys . a 630 ( 1998 ) 352 .
kratz , k. farouqi and b. pfeiffer , prog .
nucl . phys .
59 ( 2007 ) 147 .
g. martnez - pinedo et al .
59 ( 2007 ) 199 .
j. dobaczewski , i. hamamoto , w. nazarewicz and j. a. sheikh , phys .
72 ( 1994 ) 981 .
g. a. lalazissis , d. vretenar , w. pschl and p. ring , phys .
b 418 ( 1998 ) 7 .
t. otsuka , r. fujimoto , y. utsuno , b. a. brown , m. honma and t. mizutaki , phys .
87 ( 2001 ) 082502 ; 91 ( 2003 ) 179202 .
j. m. pearson , phys .
b 513 ( 2001 ) 319 .
p. ring , g. a. lalazissis , d. vretenar , nucl .
phys . a 701 ( 2002 ) 503 .
t. otsuka , prog .
suppl . 146
( 2002 ) 6 .
t. otsuka nucl .
a 722 ( 2003 ) c347 . t. otsuka , nucl .
phys . a 734 ( 2004 ) 365 .
t. otsuka , t. suzuki , r. fujimoto , h. grawe and y. akaishi , phys .
95 ( 2005 ) 232502 .
t. otsuka , t. matsuo and d. abe , phys .
97 ( 2006 ) 162501 .
t. otsuka and d. abe , prog .
59 ( 2007 ) 425 .
j. dobaczewski , n. michel , w. nazarewicz , m. poszajczak and j. rotureau , prog .
59 ( 2007 ) 432 .
g. col , h. sagawa , s. fracasso and p. f. bortignon , phys .
b 646 ( 2007 ) 227 . k. t. hecht and a. adler , nucl .
phys . a 137 ( 1969 ) 129 .
a. arima , m. harvey and k. shimizu , phys .
30b ( 1969 ) 517 .
a. bohr , i. hamamoto and b. r. mottelson , phys .
26 ( 1982 ) 267 .
o. castaos , m. moshinsky and c. quesne , phys . lett .
b 277 ( 1992 ) 238 . a. b. balantekin , o. castaos and m. moshinsky , phys .
b 284 ( 1992 ) 1 .
f. iachello , rev .
65 ( 1993 ) 569 .
a. l. blokhin , c. bahri and j. p. draayer , phys .
74 ( 1995 ) 4149 .
a. l. blokhin , c. bahri and j. p. draayer , j. phys .
a : math . gen . 29 ( 1996 ) 2039 .
j. n. ginocchio , phys .
78 ( 1997 ) 436 . j. n. ginocchio and a. leviatan , phys .
b 425 ( 1998 ) 1 .
j. n. ginocchio , nucl .
phys . a 654 , ( 1999 ) 663c .
j. n. ginocchio , phys .
rep . 315 ( 1999 ) 231 .
j. n. ginocchio and a. leviatan , phys .
87 ( 2001 ) 072502 .
a. leviatan and j. n. ginocchio , phys .
b 518 ( 2001 ) ? ? ? ? .
j. n. ginocchio , phys .
c 66 ( 2002 ) 064312 . j. n. ginocchio , phys . rep . 414 ( 2005 ) 165 .
j. meng , k. sugawara - tanabe , s. yamaji , p. ring and a. arima , phys .
c 58 ( 1998 ) r628 .
j. meng , k. sugawara - tanabe , s. yamaji and a. arima , phys .
c 59 ( 1999 ) 154 .
k. sugawara - tanabe , s. yamaji and a. arima , phys .
c 62 ( 2000 ) 054307 .
s. marcos , l. n. savushkin , m. lopez - quelle and p. ring , phys .
c 62 ( 2000 ) 054309 .
p. alberto , m. fiolhais , m. malheiro , a. delfino and m. chiapparini , phys .
86 ( 2001 ) 5015 .
p. alberto .
m. fiolhais , m. malheiro , a. delfino and m. chiapparini , phys .
c 65 ( 2002 ) 034307 .
k. sugawara - tanabe , s. yamaji and a. arima , phys .
c 65 ( 2002 ) 054313 . s .-
zhou , j. meng and p. ring , phys .
91 ( 2003 ) 262501 .
j. n. ginocchio , a. leviatan , j. meng and s .- g .
zhou , phys .
c 69 ( 2004 ) 034303 .
p. alberto , r. lisboa , m. malheiro and a. s. de castro , phys .
c 71 ( 2005 ) 034313 . c. berkdemir , nucl .
a 770 ( 2006 ) 32 .
q. xu and s .- j .
zhu , nucl phys .
a 768 ( 2006 ) 161 .
w. h. long , h. sagawa , j. meng and n. van giai , phys .
b 639 ( 2006 ) 242 .
a. s. de castro , p. alberto , r. lisboa and m. malheiro , phys .
c 73 ( 2006 ) 054309 . h. leeb and s. wilmsen , phys .
c 62 ( 2000 ) 024602 . j. n. ginocchio , phys . rev .
c 65 ( 2001 ) 054002 . h. leeb and s. a. sofianos , phys .
c 69 ( 2004 ) 054608 . j .- y .
guo , r .- d .
wang and x .- z .
fang , phys .
c 72 ( 2005 ) 054319 .
guo and x .- z .
fang , phys .
c 74 ( 2006 ) 024320 .
a. lahiri , p. k. roy and b. bagchi , int .
phys . a 5 ( 1990 ) 1383 ; 4579 .
f. cooper , a. khare and u. sukhatme , phys .
251 ( 1995 ) 267 .
h. kalka and g. soff : _ supersymmetrie _ , b. g. teubner ( stuttgart , 1997 ) .
f. cooper , a. khare and u. sukhatme : _ supersymmetry in quantum mechanics _ , world scientific publishing ( singapore , 2001 ) .
j. walecka , ann .
83 ( 1974 ) 491 .
b. d. serot and j. d. walecka , adv .
( 1986 ) 1 .
reinhard , rep .
52 ( 1989 ) 439 . p. ring , prog .
nucl . phys . 37
( 1996 ) 37 .
a. typel and h. h. wolter , nucl .
a 656 ( 1999 ) 331 .
f. hofmann , c. m. keil and h. lenske , phys .
c 64 ( 2001 ) 034314 .
t. brvenich , d. g. madland , j. a. maruhn and p .-
reinhard , phys .
c 65 ( 2002 ) 044308 .
m. bender , p .- h .
heenen and p .-
reinhard , rev .
phys . 75 ( 2003 ) 121 .
r. d. mota , v. d. granados , a. queijeiro , j. garca and l. guzmn , j. phys . a : math .
36 ( 2003 ) 4849 .
r. lisboa , m. malheiro , a. s. de castro , p. alberto and m. fiolhais , phys .
c 69 ( 2004 ) 024319 .
j. n. ginocchio , phys .
95 ( 2005 ) 252501 .
guo , x .- z .
fang and f .- x .
phys . 757 (
2005 ) 411 .
e. schrdinger , proc .
irish acad . a 46 ( 1940 ) 9 ; 46 ( 1940 ) 183 ; 47 ( 1941 ) 53 .
l. infeld , phys .
59 ( 1941 ) 737 .
l. infeld and t. e. hull , rev .
( 1951 ) 21 .
a. leviatan , phys .
92 ( 2004 ) 202501 .
a. leviatan , int .
e 14 ( 2005 ) 111 .
.[tab : kjl]relation between the various quantum numbers in the relativistic description of single - particle states .
pseudospin partners with the same pseudoangular momentum @xmath17 are placed in the same column . [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] | the evolution of single - particle energies with varying isospin asymmetry in the shell model is an important issue when predicting changes in the shell structure for exotic nuclei . in many cases
pseudospin partner levels , that are almost degenerate in energy for stable nuclei , are relevant in extracting the size of the shell gaps .
a breaking of the pseudospin symmetry can affect the size of these gaps and change the magic numbers accordingly .
the strength of the pseudospin splitting is expected to depend in particular on isovector - dependent and tensor contributions to the effective nuclear interaction .
a description employing supersymmetric quantum mechanics allows to derive a pseudospin symmetry breaking potential that is regular in contrast to the pseudospin - orbit potential in the conventional relativistic treatment .
the derived perturbation potential provides a measure to quantify the symmetry breaking and it can be employed to improve mean - field calculations in order to better reproduce the experimentally observed shell evolution .
general potentials with exact pseudospin symmetry are obtained that can be used in relativistic mean - field hamiltonians .
pseudospin symmetry , supersymmetry , shell model , relativistic mean field model , single - particle states 11.30.pb , 21.10.pc , 21.30.fe , 21.60.cs |
ultra - relativistic heavy ion collisions provide a unique environment to study the properties of strongly interacting matter at high temperature and high energy density @xcite .
leptons and photons are penetrating probes of the hot , dense medium since they are not affected by the strong interaction and therefore they can probe the whole evolution of the collision . in the low invariant mass range of produced lepton pairs ( @xmath0 gev/@xmath1 ) , we can study vector meson in - medium properties through their dilepton decays , where modifications of mass and width of the spectral functions observed may relate to the possibility of chiral symmetry restoration @xcite .
the dilepton spectra in the intermediate mass range ( @xmath2 gev/@xmath1 ) are directly related to thermal radiation of the quark - gluon plasma ( qgp ) @xcite .
however , contributions from other sources have to be obtained experimentally .
such contributions include background pairs from correlated open heavy flavor decays ( @xmath3 or @xmath4 ) .
in addition , photons in the low transverse momentum range @xmath5 gev/@xmath6 are used to study thermal radiation from qgp and hadronic gas . in this article
, i will summarize the results from na60 , phenix , star , and alice on thermal dileptons and photons presented at _
quark matter 2014_.
at this conference , the star collaboration presented their 200 gev @xmath7 reference measurement using high statistics 2012 data @xcite .
the cocktail simulation with expected hadronic contribution is consistent with the data .
a similar conclusion holds for @xmath7 collisions at lhc energies @xcite .
in addition , in d+au and p+pb collisions , hadronic cocktails are also found to be consistent with the measurements , indicating that there is no medium radiation observed there @xcite . at the sps ,
the low mass dilepton enhancements in the ceres @xmath9 data @xcite and in the na60 @xmath10 data @xcite indicate substantial medium effects on the @xmath11-meson spectral function .
the precise na60 measurement of the low mass enhancement provides a decisive discrimination between the dropping - mass scenario @xcite and the massively broadened spectral function @xcite .
the latter one was found to be able to describe the data consistently . at rhic ,
the phenix experiment observed a significant enhancement in the @xmath12 data above the expectation from hadronic sources for @xmath13 gev/@xmath1 and @xmath14 gev/@xmath6 in 200 gev au+au collisions @xcite .
models @xcite that successfully describe the sps dilepton data , fail to describe the phenix data .
after the time - of - flight detector upgrade , star reported the dielectron spectra in 200 gev au+au collisions at qm2011 and the low mass excess was not as significant as what phenix observed in 0 - 80% and 0 - 10% collisions @xcite .
further comparisons point to the fact that the discrepancy between star and phenix comes from 0 - 20% central collisions only @xcite .
spectral function and qgp thermal radiation , respectively @xcite .
the sum of the `` cocktail '' , `` in - medium '' , and `` qgp '' is shown as the solid curve on top of the data points.,scaledwidth=80.0% ] at quark matter 2012 , phenix reported the dielectron results from the hadron blind detector from 20 - 92% au+au collisions at 200 gev , which are consistent with the data in the corresponding centralities in the previous publication @xcite .
star reported the dielectron spectra from 19.6 , 39 , and 62.4 gev au+au collisions @xcite .
see also ref .
@xcite . at this conference ,
star reported the dielectron spectrum from 27 gev au+au collisions @xcite .
the 200 gev au+au results were finalized @xcite .
in addition , the @xmath15 differential measurements were also shown in au+au collisions at 19.6 , 27 , 39 , and 62.4 gev @xcite .
a broadened spectral function @xcite , which describes sps dilepton data , consistently accounts for the star low mass excess in au+au collisions at 19.6 , 27 , 39 , 62.4 , and 200 gev , as shown in fig .
[ fig:1 ] .
it also describes the @xmath15 dependence of low mass excess up to 2 gev/@xmath6 for all these collision energies @xcite .
in addition , the excess dielectron mass spectrum in the mass region 0.3 - 0.76 gev/@xmath16 in 200 gev au+au collisions is found to follow @xmath17 dependence , where @xmath18 is the number of participant nucleons in a collision @xcite .
dilepton spectra in the intermediate mass range are directly related to the thermal radiation of the qgp @xcite .
thanks to the fact , that the charm cross section at 17.3 gev is small and with a vertex detector to reject the charm background , na60 presented the unprecedented excess dimuon mass spectrum in in+in collisions , as shown in fig .
[ fig:2 ] . from its intermediate mass region ,
one can obtain the temperature of the emitting source .
in addition , more differential results such as the @xmath15 , azimuthal angle , and polar angle dependences of the excess dimuons in different mass regions were obtained @xcite . based on those precise measurements
, it was identified that the excess dimuons in the intermediate mass region comes from the early emission from the hot , dense medium . at rhic , with the data sets currently available , it is difficult to measure correlated charm contribution or qgp thermal radiation in the intermediate mass region . at rhic energies , there is so far no clear statement about thermal radiation in the intermediate mass region .
the recent star measurements in 200 gev au+au indicate that the trend there could be slightly different in 0 - 80% and 0 - 10% collisions @xcite . however , the results are not precise enough to draw a conclusion .
the recent detector upgrade with the heavy flavor tracker at star , completed in 2014 , will provide precise charm cross section measurements @xcite .
this will help to understand heavy quark dynamics in the medium and constrain model inputs to calculate dilepton mass spectra from correlated charm contribution .
however the measurements of @xmath19 correlations will still be challenging if not impossible .
an independent approach is planned with the muon telescope detector upgrade ( mtd ) @xcite , which was completed in 2014 .
the @xmath20 correlations will measure the contribution from heavy flavor correlations to the dielectron or dimuon continuum with the 200 gev au+au data taken in 2014 @xcite .
this will allow to access the thermal radiation in the intermediate mass region .
in addition , the beam energy scan phase ii at rhic in 2018 - 2019 will enable to collect sufficient data in au+au collisions from 7.7 to 19.6 gev @xcite
. the charm cross section is significantly reduced at lower energies .
one would like to see whether at some collision energies the transition from low to intermediate dilepton mass region would be smooth , which would imply chiral symmetry restoration .
photons at @xmath21 gev/@xmath6 are used to study thermal radiation from qgp and hadronic medium . for @xmath21 gev/@xmath6 , phenix measured direct photon yields from dielectron measurements and found an excess in 0 - 20% au+au over @xmath7 at @xmath22 gev , exponential in @xmath15 with the inverse slope parameter 221 mev @xcite . at lhc ,
an excess of direct photon yield in 0 - 40% pb+pb collisions above @xmath7 was reported at @xmath23 2.76 tev , exponential in @xmath15 with the inverse slope parameter 304 mev @xcite .
if indeed the excess is from the qgp phase , the measurements at rhic and lhc would indicate that the initial temperature of the qgp evolution is as high as 300 - 600 mev @xcite .
on the other hand , @xmath24 of direct photons has been found to be substantial in the range @xmath21 gev/@xmath6 in central 0 - 20% au+au collisions at @xmath22 gev @xcite .
model calculations @xcite for qgp thermal photons in this kinematic region significantly under - predict the observed @xmath24 , while if a significant contribution from the hadronic sources at later stages is added , the excess of the spectra and the observed @xmath24 at @xmath21 gev/@xmath6 are described reasonably well @xcite . in addition , alice reported that significant @xmath24 is observed for direct photons , though the systematic uncertainties are large @xcite . in 0 - 80% au+au from star ,
0 - 92% au+au from phenix , and @xmath7 from phenix at 200 gev @xcite .
, scaledwidth=70.0% ] of direct photons and pions in au+au collisions at 200 gev @xcite .
, scaledwidth=95.0% ] at this conference , star presented their first direct photon yields from dielectron measurements and found an excess in 0 - 80% au+au over @xmath7 at @xmath22 gev for @xmath21 gev/@xmath6 @xcite . for @xmath25 gev/@xmath6 ,
the yields are consistent with the number of binary scaled @xmath7 measurements .
the lack of the @xmath26 meson measurements at @xmath27 gev/@xmath6 leads to large uncertainties in the hadronic cocktail components .
this leads to large systematic uncertainties ( greater than 50% ) in the direct virtual photon yields for @xmath28 gev/@xmath6 . within these uncertainties ,
the star measurements are consistent with the published phenix results @xcite , as shown in fig .
[ fig:3 ] . using external conversion technique , phenix measured the direct photon yields down to @xmath29 of 0.4 gev/@xmath6 .
excess yields in au+au over @xmath7 are observed at @xmath22 gev , which are exponential in @xmath15 with slope parameters independent of centrality @xcite . it is found that the excess follows @xmath30 dependence .
in addition , phenix observed that direct photons have a significant triangular flow @xmath31 , similar to that of pions , as shown in fig .
[ fig:4 ] .
in addition , alice reported that they were doing critical assessments for the measurements of @xmath15 spectrum and @xmath24 of direct photons in pb+pb collisions at 2.76 tev .
conclusions should be drawn after assessments are completed @xcite .
the excess @xmath15 spectrum of thermal photon in a+a collisions depends on the measurements of the direct photon spectra in a+a and @xmath7 collisions .
the future improvements thus come from those from a+a and @xmath7 .
in particular , for a+a collisions , the measurement using internal conversion method would benefit from the precise @xmath26 measurement at @xmath32 2 gev/@xmath6 , while that using external conversion would benefit from more precise @xmath33 and @xmath26 measurements . in @xmath7 collisions ,
the measurement of direct photon at @xmath32 2 gev/@xmath6 has large uncertainties at rhic and does not exist at lhc .
future precise measurement is critical to get a solid physics picture of the excess @xmath15 yield .
in addition , it has been proposed that dilepton @xmath24 measurements as a function of @xmath15 in different mass regions would allow to probe the properties of medium from a hadron - gas - dominated to a qgp - dominated scenario @xcite . at the previous and this quark matter , star reported the dielectron @xmath24 measurements from 200 gev au+au collisions . within uncertainties ,
the data are compatible with known hadronic sources without qgp or hadron gas thermal radiation @xcite .
much more data are needed for the dielectron @xmath24 measurement to provide additional sensitivity to study the thermal radiation from the different phases in addition to the @xmath20 correlation measuring the charm correlation contribution . in the future
, dimuon might provide an alternative approach considering the trigger capabilities @xcite .
in qcd vacuum , the quark condensate has non - zero expectation value , which leads to spontaneous chiral symmetry breaking .
this generates 99% of the visible mass in the universe .
in addition , the mass degeneracy between chiral partners , like @xmath34 and @xmath35 mesons , is lifted .
when the temperature and density is high enough , theoretical calculations predict that the chiral symmetry will be restored .
the mass difference between @xmath11 and @xmath36 will disappear . experimentally , there are many efforts since decades to search for the signature of the chiral symmetry restoration .
since it is very difficult and challenging to measure @xmath36 , the experimental efforts goes to the study of the modification of @xmath11 spectral function .
an observed broadened and structureless spectral function would imply chiral symmetry restoration .
ratios as a function of energy in au+au collisions together with the excess yield predictions from a model @xcite .
( bottom panel ) : the integral dielectron excess yields within star acceptance in the mass region 0.3 - 0.7 gev/@xmath16 normalized by mid - rapidity pion yields in au+au collisions @xcite . data
are measurements from 19.6 to 200 gev .
the rising line represents model expectations and the grey band is for the extrapolations assuming the normalized excess yields do not change down to 7.7 gev .
, title="fig:",scaledwidth=60.0% ] ratios as a function of energy in au+au collisions together with the excess yield predictions from a model @xcite .
( bottom panel ) : the integral dielectron excess yields within star acceptance in the mass region 0.3 - 0.7 gev/@xmath16 normalized by mid - rapidity pion yields in au+au collisions @xcite .
data are measurements from 19.6 to 200 gev .
the rising line represents model expectations and the grey band is for the extrapolations assuming the normalized excess yields do not change down to 7.7 gev .
, title="fig:",scaledwidth=60.0% ] it is found that a broadened @xmath11 spectral function @xcite , describes sps dilepton data , consistently accounts for the star low mass excess at 19.6 , 27 , 39 , 62.4 , and 200 gev .
furthermore , it is found that coupling to baryons plays an essential role in the modification of @xmath11 spectral function in the hot , dense medium . the total baryon density , derived from the @xmath37 ratios , does not change significantly from 17.3 to 200 gev , as shown in the top panel of fig .
[ fig:5 ] .
therefore , the measurements from 17.3 to 200 gev probe the temperature and system evolution dependence .
figure [ fig:5 ] ( bottom panel ) shows the integral dielectron excess yields within star acceptance in the mass region 0.3 - 0.7 gev/@xmath16 normalized by mid - rapidity pion yields .
no significant energy dependence of the normalized excess yields from 19.6 to 200 gev is observed @xcite . in order to be sensitive to the total baryon density effect ,
one need to have measurements at lower energies . at 7.7 gev
, it is found that the total baryon density increases by a factor of two , as indicated in the top panel of fig .
[ fig:5 ] .
current data at 7.7 and 11.5 gev at rhic are not sufficient for dilepton analysis .
the future measurements from the beam energy scan phase ii at rhic , will map out the dependence of modified @xmath11 spectral function on the total baryon density from 7.7 to 19.6 gev au+au collisions @xcite . the grey band shown in the bottom panel of fig .
[ fig:5 ] represents the precision projection if the normalized excess yield does not change .
if it increases when the energy decreases as predicted by models , the uncertainties down to 7.7 gev might be smaller .
in addition , there exists a proposal of a new na@xmath38 experiment at sps to measure the excess dimuon mass spectra precisely at 6 - 20 gev @xcite .
these , together with the future measurements from alice at lhc @xcite , hades at sis18 @xcite , and cbm at sis100 @xcite will provide a unique opportunity to study chiral symmetry restoration .
this edition of the quark matter conferences has seen a stream of results unprecedented in their quality and in their quantity , being presented by the experimental collaborations working at rhic and at the lhc .
thermal photon and dilepton measurements from a broad beam energy range enable us to study the fundamental properties of qgp , of chiral symmetry restoration , and will provide stringent tests of the dynamical evolution scenarios of relativistic nuclear collisions .
+ * acknowledgment * the work of of lr is supported in part by the u. s. department of energy under contract no .
de - ac02 - 98ch10886 and under early career research program funding award no .
fwp@xmath392013-bnl - po143 .
g. agakichiev _
et al . _ , .
r. arnaldi _ et al .
_ , ; h.j .
specht ( na60 ) , aip .
conf . proc .
1322 ( 2010 ) 1 - 10 .
brown and m. rho , phys . rep . * 269 * , 333 ( 1996 ) .
r. rapp and j. wambach , eur .
j. a**6 * * , 415 ( 1999 ) ; h. van hees and r. rapp , ; t. renk and j. ruppert , .
et al . _ , .
r. rapp , j. wambach , and h. van hees , arxiv:0901.3289 .
o. linnyk _
. l. adamczyk _
_ , ; l. ruan _ et al .
_ , . j. zhao _
et al . _ , . c. gale and l. ruan , .
e. atomssa _
et al . _ , ; i. tserruya _
et al . _ , .
x. dong _ et al .
_ , ; f. geurts _ et al .
et al . _ , .
p. huck ( star ) , these proceedings .
l. adamczyk _
et al . _ , . http://rnc.lbl.gov/hft/docs/hft@xmath40final@[email protected] ; s. kleinfelder _ et al . _ , . http://drupal.star.bnl.gov/star/system/files/mtd@[email protected] ; z. xu , bnl ldrd project 07 - 007 ; l. ruan _ et al . _ , ; y. sun et al . , ; y. wang et al . , .
_ , arxiv:1207.7043 .
star public note : https://drupal.star.bnl.gov/star/starnotes/public/sn0598 .
et al . _ , .
et al . _ , .
et al . _ , .
r. chatterjee and d.k .
srivastava , .
h. van hees , c. gale , and r. rapp , .
d. lohner _
et al . _ , j. phys .
446 ( 2013 ) 012028 .
s. mizuno ( phenix ) , these proceedings ; a. adare _ et al .
_ , arxiv : 1405.3940 . f. bock ( alice ) , these proceedings .
r. chatterjee _ et al .
. l. adamczyk _
_ , arxiv:1402.1791 .
g. usai , these proceedings .
t. galatyuk , these proceedings .
claudia hohne , these proceedings . | i summarize and discuss some of the experimental results on the low and intermediate mass dileptons and direct photons presented at _
quark matter 2014_. # 1#2#3#4#1 * # 2 * , # 3 ( # 4 ) vector meson in - medium modification , chiral symmetry restoration , quark - gluon plasma thermal radiation , dileptons , thermal photons |
the genomes of a number of organisms are already known and more are being completed at a rate of perhaps one a month .
once a genome @xcite has been sequenced the amino - acid sequences of _ all _ its proteins are known .
the complete set of proteins of an organism is generally referred to as its proteome .
here we use genome data for _ e. coli _ to systematically estimate the change in the interactions of all the proteins of this bacterium when the salt concentration is varied .
_ e. coli _ can grow in environments with a very wide range of salt concentrations @xcite , and so its proteins must function _ in vivo _ over a wide range of salt concentrations .
the potassium ion concentration inside the cell can vary from approximately @xmath1 to @xmath2 mm @xcite ; potassium is the predominant cation in living cells .
clearly , the proteins must remain soluble over this range , and they should bind to the other proteins which they are required to bind to in order to function , but they should not interact strongly with other proteins .
the study of a proteome is often called proteomics . here , as the physical properties of proteins are studied ( as opposed to their chemical properties such as catalytic function )
we are doing what may be called physical proteomics .
there has been extensive theoretical work on the salt dependence of the interactions in individual proteins , particularly for the protein lysozyme @xcite .
see refs . for corresponding experimental work
. however , as far as the author is aware , this is the first attempt to characterise the interactions of _ all _ the proteins of an organism .
we will consider the proteins separately , i.e. , as single component solutions .
of course inside a bacterium the proteins exist as a mixture of thousands of components .
future work will consider multi - component mixtures of the proteins .
we have chosen _
e. coli _ as it is a bacterium , and therefore a relatively simple organism , and as it has been extensively studied .
however , the distribution of charges on the proteins of almost all organisms is very similar and so our results apply to almost all organisms , including _ h.
sapiens_. the only exceptions are some extremophiles @xcite . in the next section we use genome data to estimate the charges on the proteins of _
e. coli_. this data is used in the third section where we calculate the variation in their second virial coefficients as the salt concentration is varied .
the last section is a conclusion .
_ e. coli k-12 _ has a proteome of 4358 proteins .
the amino acid sequences of all of them are known from the sequencing of its genome @xcite .
_ k-12 _ is the name of a strain of _
e. coli_. runcong and mitaku have analysed the charge distributions of a number of organisms using a simple approximate method of estimating the charge on a protein at neutral ph from its amino - acid sequence @xcite .
we will follow their analysis but use a slightly different approximation for the charge on a protein with a given amino - acid sequence . of the 20 amino acids , 5 have pk values such
that they should be at least partially charged at neutral ph @xcite .
these are two highly acidic amino acids , aspartic acid and glutamic acid , two highly basic amino acids , lysine and arginine and one somewhat basic amino acid , histidine .
aspartic and glutamic acids have pk s far below 7 and lysine and arginine have pk s far above 7 and so we assume that all 4 of these amino acids are fully charged at neutral ph .
aspartic and glutamic acids then each contribute @xmath3 to the charge on a protein , and lysine and arginine each contribute @xmath4 .
histidine has a pk of around 6 - 6.5 ( this will depend on the environment of the amino acid ) .
the equation for the fraction @xmath5 of a basic group such as histidine that is charged at a given ph is @xmath6 where pk is the pk value for the basic group .
this equation is just the henderson - hasselbalch equation @xcite rearranged .
taking pk@xmath7 @xcite and at ph=7 we have that the fraction of histidines charged is @xmath8 . as this is small
we assume for simplicity that all the histidine amino acids are uncharged .
thus , with these assumptions for the charges on these 5 amino acids , our estimate for the net charge @xmath9 on a protein is simply given by @xmath10 where @xmath11 , @xmath12 , @xmath13 and @xmath14 are the the protein s total numbers of lysines , arginines , aspartic acids and glutamic acids , respectively .
the subscripts @xmath15 , @xmath16 etc .
correspond to the standard single letter codes for the amino acids @xcite .
the charge @xmath9 is in units of @xmath17 where @xmath17 is the elementary charge .
note that runcong and mitaku @xcite assume that the histidine amino acids contribute @xmath4 to the charge , that is the only difference between our analysis and that of runcong and mitaku .
as the histidine amino acid is quite a rare amino acid , approximately 1 in 50 amino acids is a histidine , the difference between the results we obtain and those of runcong and mitaku @xcite is not large but our charges are shifted to more negative values . using runcong and mitaku s approximation the mean charge on a protein is @xmath18 units more positive than the mean charge we find here . as a check on our algorithm , we can compare the prediction of equation ( [ qdef ] ) for chicken lysozyme to that of a titration experiment to determine the charge .
equation ( [ qdef ] ) predicts that chicken lysozyme @xcite has a net charge of 8 at neutral ph .
titration experiments on lysozyme give a titratable charge of close to @xmath19 at ph=7 @xcite .
using equation ( [ qdef ] ) we can obtain estimates for the charges of all 4358 proteins of _ e. coli _ @xcite .
the results are shown in fig .
[ qcoli ] , where we have plotted the number of proteins @xmath20 as a function of net charge @xmath9 .
the distribution is centered almost at a net charge @xmath21 , and for not - too - large @xmath22 the distribution is roughly symmetric and gaussian .
the mean charge is @xmath23 .
given the approximate nature of our equation for the charge on a protein , equation ( [ qdef ] ) , the data is probably consistent with a mean charge of 0 .
the approximation scheme of mitaku and runcong @xcite yields a mean charge of @xmath24 .
also , although when @xmath22 is not too large the distribution can be seen to be reasonably symmetric , _ e. coli _ has 12 proteins with charges @xmath25 but none with charges @xmath26 . excluding proteins with very large charges , @xmath27 , the root mean square charge equals @xmath28 .
a number of other organisms , both other bacteria and eukaryotes such as yeast , have had the charge distribution on their proteomes determined by runcong and mitaku @xcite and by the author @xcite .
almost all of them have a roughly gaussian distribution centered approximately at zero , like the distribution in fig . [ qcoli ] .
the exceptions are some extremophiles .
extremophiles are organisms that live in extreme environments , for example _ halobacterium sp .
_ lives in environments with very high levels of salt @xcite .
the cytosol of _ halobacterium sp . _ contains much higher levels of potassium ions than do other organisms so perhaps it is not a surprise that the distribution of charges on its proteins is different .
we have fitted the gaussian function @xmath29 to the data for the number of proteins as a function of their charge .
it is drawn as the solid curve in fig .
[ qcoli ] .
the fit parameters are mean charge @xmath30 and standard deviation @xmath31 .
1739 is @xmath32 and so the distribution is normalised so that its integral gives the total number of proteins . within a couple of standard deviations of the mean the gaussian function fits the data well but it underestimates the numbers of proteins with charges such that @xmath33 is several times the standard deviation .
we also note that there is a correlation between the net charge @xmath9 on a protein and its size , measured by the number of amino acids @xmath34 .
figure [ scat_coli ] is a scatter plot of charge @xmath9 and number of amino acids @xmath34 for the proteins of _
e. coli_. although at any particular size @xmath34 there is a wide distribution of charges , on average the more highly charged proteins are larger than average .
we expect the volume of a protein to scale with @xmath34 .
consider a dilute solution of a single one of the proteins of _
e. coli_. apart from water , the only other constituents are a 1:1 salt at a concentration @xmath35 and a buffer which controls the ph while making a negligible contribution to the ionic strength . here
we will always assume the ph=7 but other ph s can be considered if the net charges on the proteins can be calculated .
also , the counterions of the protein are assumed the same as either the anions or cations of the salt , depending on the sign of @xmath9 .
the interactions between the protein molecules in the salt solution can be characterised by means of the protein s second virial coefficient @xmath36 : a function of temperature , ph and salt concentration .
proteins are complex molecules and we are unable to calculate from first principles the absolute value of @xmath36 for any of the 4358 proteins possessed by _
e. coli_. however , predicting the change in the second virial coefficient when the salt concentration varies is a much easier problem , _ if _ we assume that changing the salt concentration changes only the direct electrostatic interaction between the net charges of a protein .
this is a strong assumption but studies of the simple protein lysozyme have shown that the variation of its second virial coefficient can be described using a simple model which only includes its net charge @xcite .
here we will follow warren @xcite and apply his analysis of lysozyme to the complete set of proteins of _
e. coli_. we will discuss which proteins are likely to be less well described by this theory than is lysozyme . a protein molecule of charge @xmath9
is surrounded by its counterions and as the concentration of the protein is increased so is the counterion density .
this increase in the counterion density decreases the translational entropy of the counterions and this contributes a positive amount to the second virial coefficient .
see warren @xcite and references therein for details .
@xmath36 has the form @xcite @xmath37 where @xmath38 is an assumed constant term due to excluded volume interactions and other interactions which are insensitive to salt concentration .
the second term is from the counterions and the salt .
it is quadratic in the charge and so of course is zero for uncharged proteins and is independent of the sign of the net charge on a protein .
as stated above we are unable to calculate the absolute value of @xmath36 ; the values of @xmath38 of the proteins are unknown .
however , we can calculate the difference in @xmath36 when the salt concentration is changed from @xmath39 to @xmath40 .
it is @xmath41 this is easy to calculate for any protein and in fig .
[ db2 ] we have plotted the number of proteins @xmath20 as a function of the change in their second virial coefficient , @xmath42 , when the salt concentration is decreased from 1 m to @xmath0 m .
the results are given in units of nm@xmath43 .
for comparison the volume of a typical bacterial protein is about 60nm@xmath43 and so if a protein were to interact solely via a hard repulsion it would have a second virial coefficient of about 4 times its volume or about 240nm@xmath43 .
results for proteins with @xmath44 are shown .
proteins with larger titratable charges are likely to have an effective charge lower than @xmath9 , see refs . and references therein . from the linear poisson - boltzmann equation , the potential ( divided by @xmath17 ) at the surface of a spherical particle with charge @xmath9 and
radius @xmath45 is @xmath46 .
@xmath47 is the bjerrum length , and and @xmath48 is the debye screening length , given by @xmath49 . for the dielectric constant of water 80 times that in vacuum and at room temperature , @xmath50 nm . globular proteins are approximately spherical and typically have radii around 2 to 4 nm .
taking a protein with a radius of 3 nm , in salt at a concentration @xmath51 m , we have that for @xmath52 , the potential at the surface is about @xmath53 .
larger charges correspond to larger surface potentials and these large potentials bind oppositely charged ions to the surface reducing the effective charge . on average
, this effect will be diminished to a certain extent by the fact that the most highly charged proteins are larger than average .
see fig .
[ scat_coli ] , where it is clear that the charge and size of a protein are correlated .
recent simulations by lobaskin _
_ @xcite of spheres with radius 2 nm and charge @xmath54 in the absence of salt found an effective charge of a little under @xmath55 .
thus we restrict ourselves to proteins with charges of magnitude less than or equal to 30 .
4300 of the 4358 proteins , or almost 99% , have charges in this range .
the mean change in @xmath36 of these 4300 proteins when the salt concentration is decreased from 1 to @xmath0 m is 139nm@xmath43 and the standard deviation is 218nm@xmath43 . a couple of caveats .
the first is that the effect of salt on protein solutions is known to depend not only on whether the salt is a 1:1 salt , a 1:2 salt etc . but also to the nature of ions , whether it is mg@xmath56 or ca@xmath56 for example @xcite .
our generic theory applies only where there are no specific interactions between the salt and the protein .
there is good agreement between experiment and theory for lysozyme plus nacl @xcite and so we may hope that it applies to nacl and many proteins but it clearly misses potentially important effects for other salts where there are specific protein - salt interactions .
the second is that proteins are not simple charged spheres , for example some have large dipole moments .
dipoles exert net attractions which are screened and hence weakened by added salt .
thus proteins with a small charges but large dipole moments are poorly described by the current theory : if the dipole interactions are dominant then the second virial coefficient may even increase when the salt concentration is increased .
@xcite discuss this point . note that although we can estimate the charge on a protein from its amino - acid sequence we can not estimate its dipole moment without knowing its three - dimensional structure , and so the sequence data from genomics is not adequate to determine dipole moments .
here we have shown how data from genomics can be used to estimate the charges on the proteins of an organism .
we then used these charges to estimate the changes in the second virial coefficients of 4300 ( 99% ) of the proteins of _ e. coli _ when the salt concentration is changed . note that _
e. coli _ can survive and multiply in external environments with a very wide range of salt concentrations ; cayley _ et al . _
@xcite studied the growth of _
e. coli _ in environments with salt concentrations ranging from very low to @xmath57 molar , corresponding to potassium ion concentrations inside the cell of @xmath1 to @xmath2 mm @xcite .
thus , studying the change in interactions of proteins with salt concentration is of direct relevance to the _ in vivo _ behaviour of proteins . within molecular biology
there is a clear shift of emphasis away from studying the proteins of an organism one or a few at a time , and towards determining the structure and function of large sets of proteins , in particular proteomes .
the systematic study of these large sets of proteins is often called proteomics .
this work is a first attempt to keep up with this shift by performing a simple theoretical calculation of a solution phase physical property for a complete proteome , rather than for one or a handful of proteins as is usually done .
it may be termed physical proteomics .
.the charged macromolecular species in the cytosol of _ e. coli_. the data is from neidhardt @xcite .
@xmath9 is the charge on a macromolecule . for proteins
the range given is the mean plus and minus twice the standard deviation .
@xmath20 is the total number of molecules of a species per cell . the volume of an _
e. coli _ cell is about @xmath58m@xmath43 so 1 molecule per cell corresponds to a concentration of about @xmath59 molar .
a prokaryote ribosome consists of about 4500 bases of rna plus protein .
the ribosomal proteins are mostly quite stongly positively charged and so will decrease the net negative charge . as a rough estimate we settle on a net charge of @xmath60 .
the charge on a trna molecule is around @xmath61 , @xmath3 from each of its bases @xcite .
the charge on a mrna molecule is on average around @xmath62 @xcite .
the charge on dna is equal to twice the number of base pairs , 4,639,221 for _ e. coli k-12 _ @xcite .
the charge density from the proteins assumes that the proteins have the mean charge of @xmath63 that they would have if their density and charge were uncorrelated .
[ t2 ] [ cols="^,^,^,^",options="header " , ] future work could consider mixtures of proteins , ultimately aiming to understand the cytosol of a living cell , which is a mixture of of order @xmath64 different types of proteins as well as dna , rna , ions like atp and potassium , etc .. this is of course very complex but _ if _ in the cytosol the proteins of _ e. coli _ are present in amounts which are uncorrelated with their net charge , the mean charge of the proteins will be close to the mean of the distribution of fig .
[ qcoli ] .
this is quite small .
neidhardt @xcite has taken an inventory of the species inside _ e. coli _ , and the results for charged macromolecules are shown in table [ t2 ] .
the charged macromolecules in a cell are protein , dna and the various forms of rna : transfer rna ( trna ) , messenger rna ( mrna ) and the rna in ribosomes ( rrna ) .
see refs . for an introduction to the proteins , dna and rna .
although for every molecule of trna molecule there are 10 of protein , for every ribosome there are 100 molecules of protein , and for every mrna there are 1000 molecules of protein , the contributions of the trna , ribosomes , dna and proteins , to the overall charge density of the macromolecules are very roughly comparable . the ribosomes contribute the largest amount . the macromolecules
are negatively charged and this negative charge is balanced by potassium ions @xcite .
thus the cytosol resembles a solution of a negatively charged polyelectrolyte , except that there is not one , relatively simple , macromolecular species , but thousands of rather complex and diverse species of macromolecules .
it is a pleasure to acknowledge discussions with j. cuesta , d. frenkel and p. warren .
a brief introduction to the biological nomenclature : an organism s dna contains many genes , each of which codes for a protein .
the complete set of genes is called the organism s genome and we will refer to the complete set of proteins as its proteome .
some authors use the word proteome somewhat differently , they use it to denote the set of proteins present in the cytosol of an organism at a particular time . see for example @xcite for a more detailed definition of a genome .
the complete proteome of _
e. coli k-12 _ ,
i.e. , the amino - acid sequences of all its proteins , can be downloaded from databases such as that at the european bioinformatics institute ( http://www.ebi.ac.uk/proteome ) .
_ e. coli k-12 _ was sequenced by blattner _
_ @xcite .
f. c. neidhardt , chemical composition of _
e. coli _ , in _ _ e .
coli _ and _ s. typhimurium _ : cellular and molecular biology _ edited by f. .c .
neidhardt j. l. ingraham , k. b. low , b. magasanik and h. e. umbarger ( american society for microbiology , washington d. c. , 1987 ) . | bacteria typically have a few thousand different proteins .
the number of proteins with a given charge is a roughly gaussian function of charge centered near zero , and with a width around ten ( in units of the charge on the proton ) .
we have used the charges on e. coli s proteins to estimate the changes in the second virial coefficients of all its proteins as the concentration of a 1:1 salt is increased .
the second virial coefficient has dimensions of volume and we find that on average it decreases by about twice the average volume of a protein when the salt concentration is increased from @xmath0 to 1 molar .
the standard deviation of the decrease is of the same order .
the consequences of this for the complex mixture of proteins inside an e. coli cell , are briefly discussed . |
ising model in a quenched random magnetic field ( rfim ) has attracted a considerable interest in the last three decades .
the model which is actually based on the local fields acting on the lattice sites which are taken to be random according to a given probability distribution was introduced for the first time by larkin @xcite for superconductors and later generalized by imry and ma @xcite .
a lower critical dimension @xmath1 of the rfim has remained an unsolved mystery for many years and now it is well established that a transition should exist in three and higher dimensions for finite temperature and randomness , which means that @xmath2 @xcite .
a great many experimental works have paid attention to the equilibrium properties of rfim and quite noteworthy results have been obtained .
for instance , it has been shown that diluted anti - ferromagnets such as @xmath3 @xcite , @xmath4 @xcite , and @xmath5 @xcite in a uniform magnetic field just correspond to a ferromagnet in a random uniaxial magnetic field @xcite . from the theoretical point of view
, equilibrium rfim has been studied by a wide variety of techniques such as mean - field theory ( mft ) @xcite , effective - field theory ( eft ) @xcite , monte carlo ( mc ) simulations @xcite , and the series expansion ( se ) method @xcite .
based on these theoretical works , it is well known that different random - field distributions may lead to different phase diagrams and the presence of quenched randomness constitutes an important role in material science , since it may induce some important macroscopic effects on the thermal and magnetic properties of real materials .
when a ferromagnetic material is subject to a periodically varying time dependent magnetic field ( kinetic ising model ) , the system may not respond to the external magnetic field instantaneously which causes interesting behaviors due to the competing time scales of the relaxation behavior of the system and periodic external magnetic field . at high temperatures and for the high amplitudes of the periodic magnetic field ,
the system is able to follow the external field with some delay while this is not the case for low temperatures and small magnetic field amplitudes .
this spontaneous symmetry breaking indicates the presence of a dynamic phase transition ( dpt ) @xcite which shows itself in the dynamic order parameter ( dop ) which is defined as the time average of the magnetization over a full period of the oscillating field .
dpt properties of kinetic ising model have been firstly observed theoretically by tom and oliveira within the framework of mft @xcite . since then
, much attention has been devoted to investigate the dynamic nature of the phase transitions by means of several theoretical and experimental works .
on the theoretical side , non - equilibrium phase transition properties of kinetic ising model have been widely investigated by various techniques @xcite . besides , on the experimental picture , a dpt occurring for the high frequency magnetic fields was studied by jiang et al .
@xcite using the surface magneto - optical kerr - effect technique for epitaxially grown ultrathin co films on a cu ( 001 ) surface . for a [ co(@xmath6)/pt(@xmath7 ) ] multi - layer system with strong perpendicular anisotropy , an example of dpt has been observed by robb et al .
they found that the experimental non - equilibrium phase diagrams strongly resemble the dynamic behavior predicted from theoretical calculations of a kinetic ising model .
it is clear from these works , there exists a strong evidence of qualitative consistency between theoretical and experimental studies . on the other hand ,
non - equilibrium stationary states and dynamic phase transition properties of rfim have not been well understood yet , and there exists a limited number of works in the literature @xcite . for instance , hausmann et al .
@xcite considered the behavior of an ising ferromagnet under the influence of a fast switching , random external field . according to the analytical results based on mft for the stationary state of the system
, they observed a novel type of first order phase transition which has also been verified by their extensive mc simulations .
paula et al .
@xcite determined the stationary states of rfim by using mft and constructed the phase diagrams from the stationary states of the magnetization as a function of temperature and field amplitude .
they found that the continuous phase transitions coincide with the equilibrium ones @xcite , while the first - order transitions occur at fields larger than the corresponding values at equilibrium .
in addition , they also observed that the difference between the fields at the limit of stability of the ordered phase and that of the equilibrium is maximum at zero temperature and vanishes at the tricritical point .
furthermore , acharyya @xcite studied the non - equilibrium dynamic phase transition , in the two - dimensional kinetic ising model in the presence of a randomly varying magnetic field both by mft and mc simulations and discerned that in contrast to the results found in ref .
@xcite , the nature of the transition is always continuous . in a recent work , crokidakis @xcite performed mc simulations on cubic lattices for a non - equilibrium ising model that stochastically evolves under the simultaneous operation of several spin - flip mechanisms where the local magnetic fields change sign randomly with time due to competing kinetics . from the numerical results , it has been predicted that there exist first - order transitions at low temperatures and large disorder strengths , which correspond to the existence of a non - equilibrium tricritical point at finite temperature .
very recently , costabile et al .
@xcite have studied the dynamical phase transitions of the kinetic ising model in the presence of a random magnetic field by using eft with correlations where the eft dynamic equation has been given for the simple cubic lattice @xmath8 , and the dynamic order parameter has been calculated .
it has been observed that the system presents ferromagnetic and paramagnetic states for low and high temperatures , respectively .
apart from this , they have predicted a non - equilibrium tricritical point in a phase diagram in the temperature versus applied field amplitude plane .
they have also compared the results with the equilibrium phase diagram @xcite , where only the first - order line is different . in the theoretical works mentioned above
, the random field effects have been taken into account either by a given probability distribution function ( random in space ) namely a bimodal distribution , or by generating a new configuration of random fields uniformly at each time step ( random in time ) . in the present paper
, we have studied dynamic phase transitions and stationary states of rfim driven by a periodically varying time dependent magnetic field on a simple cubic lattice .
amplitude of the applied magnetic field is sampled from both bimodal and trimodal probability distributions . for this purpose , we organized the paper as follows : in sec .
[ formulation ] we briefly present the formulations .
the results and discussions are summarized in sec .
[ results ] , and finally sec . [ conclusion ] contains our conclusions .
we consider a three dimensional ising ferromagnet @xmath9 defined on a simple cubic lattice with a time dependent external magnetic field .
the time dependent hamiltonian describing our model is @xmath10 where the first term is a summation over the nearest - neighbor spins with @xmath11 and @xmath12 is a time - dependent external oscillating magnetic field which is given by @xmath13 where @xmath14 is the amplitude of the external magnetic field acting on the site @xmath15 , and @xmath16 denotes the angular frequency of the oscillating external field .
the amplitude of the field is distributed according to a given probability distribution function .
the present study deals with a trimodal field distribution which has a form @xmath17.\ ] ] in order to describe the dynamical evolution of the system , we follow a glauber type stochastic process @xcite .
the dynamical equation of motion can be obtained by using the master equation as follows @xmath18\right\rangle,\ ] ] where @xmath19 is the transition rate per unit time , @xmath20 is the local field acting on the lattice site @xmath15 , and @xmath21 and @xmath22 denote the boltzmann constant and temperature , respectively . in eq .
( [ eq4 ] ) , we set @xmath19 at unity .
if we apply the differential operator technique @xcite in eq .
( [ eq4 ] ) by taking into account the random configurational averages we get @xmath23 where @xmath24 represents the average magnetization , @xmath25 is a differential operator , @xmath26 is the coordination number of the lattice , and the inner @xmath27 and the outer @xmath28 brackets represent the thermal and configurational averages , respectively .
the function @xmath29 in eq .
( [ eq5 ] ) is then defined by @xmath30.\ ] ] when the right - hand side of eq .
( [ eq5 ] ) is expanded , the multispin correlation functions appear . the simplest approximation , and
one of the most frequently adopted is to decouple these correlations according to @xmath31 for @xmath32 @xcite . if we expand the right - hand side of eq .
( [ eq5 ] ) within the help of eq .
( [ eq7 ] ) then we obtain the dynamical equation of motion as follows @xmath33 the coefficients in eq .
( [ eq8 ] ) are defined as @xmath34f(x)|_{x=0},\quad j=0,1, ... ,q.\ ] ] these coefficients can be calculated by employing the mathematical relation @xmath35 .
( [ eq8 ] ) can be regarded as a kind of initial value problem and the solution can be easily found benefiting from the initial value of the average order parameter @xmath36 by using the fourth order runge - kutta method ( rk-4 ) . for selected values of the hamiltonian parameters and temperature , the time dependence of magnetization converges to a finite value after some iterations i.e. the solutions have property @xmath37 for arbitrary initial value of the magnetization ( @xmath36 ) .
thus , by obtaining this convergent region after some transient steps ( which depends on hamiltonian parameters and the temperature ) the dop which is the time average of the magnetization over a full cycle of the oscillating magnetic field can be calculated from @xmath38 where @xmath39 is a stable and periodical function which can be one of the two types , according to whether it has the following property or not @xcite @xmath40 a solution satisfying eq .
( [ eq11 ] ) is called symmetric solution which corresponds to a paramagnetic ( p ) phase where the magnetization oscillates around zero whereas the solution which does not ensure eq . ( [ eq11 ] )
is called nonsymmetric solution , and it corresponds to ferromagnetic ( f ) phase where the magnetization oscillates around a non - zero value . in these two cases , the observed behavior of the magnetization is regardless of the choice of the initial value of magnetization . on the other hand ,
in contrast to the equilibrium rfim , there exist coexistence regions ( f+p phases ) in the phase diagrams in temperature versus field amplitude plane where the stationary state of non - equilibrium rfim problem depends on the initial value @xmath36 of the time dependent magnetization .
furthermore , it is not possible to obtain the free energy for kinetic models in the presence of time dependent external fields .
hence , in order to determine the type of the dynamic phase transition ( first or second order ) , it is convenient to check the temperature dependence of dop .
namely , if the dop decreases continuously to zero in the vicinity of critical temperature , this transition is classified as of the second order whereas if it vanishes discontinuously then the transition is assumed to be of the first order .
in this section , we discuss how the random - fields effect the phase diagrams of the kinetic ising model . also , in order to clarify the type of the dynamic phase transitions in the system , we give the temperature dependence of the dynamic order parameter .
plane for oscillation frequency values @xmath41 ( left panel ) and @xmath42 ( right panel ) , for comparison with refs .
solid ( dashed ) lines correspond to second ( first ) order phase transitions and solid symbols represent dynamic tricritcal points.,title="fig:",width=377 ] + in order to provide a testing ground for our calculations , we have primarily studied the dynamic phase diagrams of the kinetic ising model under an oscillating magnetic field where the amplitude of the externally applied field was taken as a uniformly constant value .
this model defines non - equilibrium properties of the pure system and has been examined previously within the framework of eft @xcite . in these works ,
the authors investigated the phase diagrams of the system in a @xmath43 plane where @xmath26 is the coordination number of the lattice . from these works
, we see that , for the oscillation frequency values @xmath41 and @xmath44 , the location of dynamic tricritical ( dtc ) point was identified imprecisely . hence , in order to compare our results with those found in refs .
@xcite , we depict the phase diagrams in fig .
[ fig1 ] in the same plane .
it is clear from fig .
[ fig1 ] that our numerical values of dtc point for @xmath41 and @xmath44 agree well with @xcite whereas they are quite different from those obtained in ref .
@xcite which is probably due to the fact that in ref .
@xcite , insufficient number of data points was used to construct the phase diagrams of the system . the distribution function given in eq .
( [ eq3 ] ) corresponds to a bimodal field distribution for @xmath45 where the amplitude of the oscillating field can be either @xmath46 or @xmath47 with equal probability . in this case
, the system can be thought as a spin system under the influence of two oscillating external field sources . in refs .
@xcite a similar model has been studied within mft and eft , respectively where the authors did not consider any oscillating external magnetic field .
main conclusion of those studies was that the dynamic second order phase transition lines in @xmath48 plane coincide with the equilibrium counterparts @xcite whereas the maximum and minimum differences between the dynamic and equilibrium first order phase transition lines were observed at the zero temperature and at the tricritical point , respectively .
however , as seen in fig .
[ fig2]a , if the external magnetic field has an oscillatory character with an amplitude which is applied at random then a completely different situation arises .
namely , the second order phase transition fields of non - equilibrium rfim with oscillating random external fields are greater than those obtained for static rfim @xcite . on the other hand , according to the dynamic phase diagram shown in fig .
[ fig2]a , the system exhibits f phase for high temperature and low field amplitude values .
in addition , as the oscillation frequency @xmath16 increases then dtc point depresses and the f+p phase region gets narrower . in fig .
[ fig2]b , variations of dtc point coordinates @xmath49 and @xmath50 with respect to the frequency @xmath51 are plotted . as seen in this figure , @xmath49 grows whereas @xmath50 decays with increasing @xmath16 and these values saturate at @xmath52 and @xmath53 .
this means that the coexistence region in the phase diagrams gets narrower but does not disappear with increasing @xmath16 . in other words , for sufficiently high frequencies , both
the location of dtc point and the area of f+p region become independent of oscillation frequency of the external field .
we also note that f region is always independent of frequency for the whole range of @xmath16 . in fig .
[ fig3 ] , we show the temperature dependence of the dynamic order parameter @xmath54 as a function of the oscillation frequency @xmath16 corresponding to the phase diagrams depicted in fig .
[ fig2 ] . as we mentioned before ,
increasing or decreasing frequency values does not affect the area of the f region .
namely , for a fixed @xmath55 value , the system always undergoes a dynamic phase transition on the f - p phase boundary line , and the critical temperature value is independent of @xmath16 .
this situation is independent of the selection of the initial magnetization for @xmath56 whereas it is valid for @xmath57 only if one starts with an initial condition @xmath58 .
as is clearly seen in fig .
[ fig3 ] , all curves coincide with each other for @xmath59 .
the inset shows the average magnetization @xmath39 as a function of time @xmath60 at @xmath61 and for @xmath62 and @xmath63 where we see that @xmath39 curve oscillates with smaller amplitude as @xmath16 increases , but the average value , i.e. dop @xmath64 over a complete cycle of the magnetic field does not change . on the other hand , as shown in the right panel of fig .
[ fig3 ] , for sufficiently strong amplitudes of the external field , such as @xmath65 , dop versus temperature curves exhibit different characteristics with increasing frequency , although the critical temperature does not change , and interestingly , for sufficiently high frequency values such as @xmath66 dop curves resemble those of the pure kinetic ising model driven by a periodic external field with the high oscillation frequency . + as a function of frequency @xmath16 for @xmath59 ( left panel ) and 4.0 ( right panel ) .
insets in both panels show the time variation of the average magnetization @xmath39 at @xmath61 ( left ) and @xmath67 ( right).,title="fig:",width=377 ] + plane for a bimodal distribution with some selected values of frequency @xmath16 .
the solid ( dashed ) lines correspond to second ( first ) order phase transitions and solid symbols represent tricritical points.,title="fig:",width=529 ] + in fig . [ fig4 ] , dynamic phase diagrams of the system
are depicted for a wide range of oscillation frequency values . in this figure
, we observe that dtc point depresses for a while and the area of the coexistence region in the phase diagrams slightly gets narrower and then remains unchanged with increasing frequency values .
the low frequency phase diagrams of the system are found to be completely different from those of the pure kinetic ising model @xcite whereas for the high frequency regime , dynamic phase diagrams of the system are exactly identical to those of the pure case .
these observations originate from the symmetry of the random field distribution .
namely , for a bimodal field distribution , we have two oscillating magnetic field sources .
initially , half of the lattice sites are under the influence of a periodically oscillating field ( source-1 ) with an amplitude @xmath55 , and the remaining spins on the other half of the lattice are influenced by an oscillating external field with an amplitude @xmath68 ( source-2 ) .
as the time progresses , due to the sinusoidal form of the periodic fields , magnetic field sources acting on the lattice sites change their sign instantaneously at the end of the one half of the oscillation period .
consequently , the presence of a phase difference between the source-1 and source-2 stimulates some unusual effects on the system .
furthermore , for sufficiently high frequencies of the oscillating external fields , relaxation time of the system becomes greater than the oscillation period of the external field . as a result of this , source-1 and source-2 fields receive identical responses from the system which perceive the field sources as a single oscillatory field source , hence the dynamic phase diagrams of the system resemble the high frequency phase diagrams of the corresponding pure model .
we may also note that f+p regions in the @xmath69 phase diagrams always exist for the whole range of frequency values . in fig .
[ fig5 ] , we represent some examples of typical dop versus temperature profiles corresponding to the phase diagrams shown in fig . [ fig4 ] with @xmath42 .
namely , the system always undergoes a second - order dynamic phase transition for @xmath70 . in this case , dynamic nature of the phase transition and the stationary state of the system are independent of the initial magnetization @xmath36 . on the other hand , two successive first order dynamic phase transitions ( i.e.
, a first - order reentrant phenomena ) occur at the value @xmath71 with the initial condition @xmath72 .
in addition , for @xmath73 with @xmath74 , dop curve exhibit a discontinuous jump at a phase transition temperature which puts forward the existence of a first - order transition on the system .
for the latter two examples , dynamic nature of the phase transition and the stationary state of the system strictly depends on the initial magnetization . as a function of the temperature for @xmath42 with some selected values of @xmath55 .
the dashed lines correspond to first order phase transitions.,title="fig:",width=377 ] + for a trimodal field distribution of magnetic fields defined in eq .
( [ eq3 ] ) , we plot the dynamic phase diagrams of the system in a @xmath48 plane in fig .
[ fig6 ] for @xmath41 and with some selected values of disorder parameter @xmath75 . as shown in fig .
[ fig6]a , the system exhibits dtc behavior for a relatively narrow range of @xmath75 .
the temperature coordinate @xmath50 of dtc point decreases as @xmath75 increases whereas the field amplitude part @xmath49 increases , and consequently dtc point depresses to zero at a certain value of @xmath75 .
after the destruction of the dtc point , all dynamic phase transition processes are found to be of the second order .
we also note that at the zero temperature , there exist a critical value @xmath76 below which the system exhibits a dynamic phase transition at a critical magnetic field .
this value of @xmath77 is independent of @xmath16 .
moreover , according to fig .
[ fig6]b , as @xmath75 increases then the area of f+p region in the dynamic phase diagrams gets narrower and after a specific value of @xmath75 , we can not observe any coexistence region in @xmath48 plane .
+ as a function of the amplitude of the oscillating field @xmath55 for a trimodal distribution corresponding to the dynamic phase diagrams depicted in fig .
[ fig6 ] with some selected values of @xmath75 for @xmath58 , @xmath41 and for the reduced temperatures ( a ) @xmath78 , ( b ) @xmath79 .
, title="fig:",width=377 ] + variation of dop @xmath64 as a function of the amplitude of the oscillating field @xmath55 for a trimodal distribution corresponding to the dynamic phase diagrams depicted in fig .
[ fig6 ] is plotted in fig .
[ fig7 ] with some selected values of @xmath75 . in this figure , initial value of the average magnetization
is selected as @xmath72 . according to fig .
[ fig7 ] , dop curves undergo first and second order dynamic phase transitions at low and high temperatures , respectively . at low temperatures , critical amplitude value @xmath80
at which a dynamic phase transition occurs depends on the initial magnetization @xmath36 .
namely , for @xmath72 critical @xmath80 value is greater than that obtained for @xmath74 which is due to the presence of f+p phase in the system at a fixed value of @xmath75 ( see fig . [
fig6]b ) . finally , for a fixed oscillation frequency @xmath41 , let us investigate the time dependence of the average magnetization @xmath39 corresponding to the phase diagrams shown in fig .
[ fig6]b . in fig .
[ fig8]a , we plot the time series of @xmath39 curves for @xmath45 , @xmath81 , and with two values of field amplitude @xmath55 . from the upper panel in fig .
[ fig8]a , we see that the dynamic nature of the stationary state of the system is independent of the initial magnetization @xmath36 for @xmath70 .
namely , for this set of system parameters , average magnetization of the system oscillates around a nonzero value ( f phase ) after some transient time .
however , for @xmath82 ( while the other parameters are the same ) stationary state of the system ( f or p ) depends on the initial magnetization @xmath36 which indicates that the dynamic phase diagrams exhibit a coexistence region ( f+p phase ) in @xmath48 plane . on the other hand , as seen in fig .
[ fig8]b which is plotted for @xmath83 , we can not observe any coexistent phase region in the system .
namely , for @xmath84 with @xmath81 , @xmath39 curves oscillate around a finite nonzero value , whereas for @xmath85 and @xmath86 the curves oscillates around zero which means that the system is in p phase .
hence , we see that the stationary state of the system for a trimodal distribution of the oscillating field amplitude may be independent of @xmath36 , depending on the value of distribution parameter @xmath75 . +
in conclusion , we have investigated the kinetic behavior of a spin-1/2 ising model on a simple cubic lattice @xmath8 under the influence of a quenched random magnetic field which oscillates periodically in time by means of an effective - field theory based on a standard decoupling approximation and the time evolution of the system has been presented by utilizing a glauber type stochastic process . for bimodal and trimodal field distributions , we have studied the global dynamic phase diagrams of the system in a @xmath48 plane and variation of the dynamic order parameter with respect to the temperature and random amplitude of the oscillating field , as well as the time series of the average magnetization curves .
we found that the dynamic behavior of the random field ising model with a periodically oscillating magnetic field exhibit quite different characteristics in comparison with the static rfim problem .
our numerical analysis clearly indicates that such field distributions may lead to a number of interesting phenomena .
for example , we have found that for a bimodal distribution with sufficiently high frequency values , dop curves and non - equilibrium phase diagrams of the system resemble those of the pure kinetic ising model driven by a periodic external field with the high oscillation frequency . on the other hand ,
we have also observed that f+p regions in @xmath48 phase diagrams always exist for the whole range of frequency values for a bimodal distribution of oscillatory field amplitude whereas for a trimodal distribution , area of f+p region gets narrower with increasing distribution parameter @xmath75 , and if we increase @xmath75 further then we can not observe any coexistence region in @xmath48 plane .
eft method takes the standard mean field predictions one step forward by taking into account the single site correlations which means that the thermal fluctuations are partially considered within the framework of eft .
although all of the observations reported in this work shows that eft can be successfully applied to such nonequilibrium systems in the presence of quenched disorder , the true nature of the physical facts underlying the observations displayed in the system ( especially the origin of the coexistence phase ) may be further understood with an improved version of the present eft formalism which can be achieved by attempting to consider the multi site correlations which originate when expanding the spin identities .
we believe that this attempt could provide a treatment beyond the present approximation .
one of the authors ( y.y . ) would like to thank the scientific and technological research council of turkey ( tbitak ) for partial financial support .
the numerical calculations reported in this paper were performed at tbitak ulakbim , high performance and grid computing center ( tr - grid e - infrastructure ) .
a. i. larkin , sov .
jetp * 31 * , 784 ( 1970 ) .
y. imry and s. k. ma , phys . rev .
lett . * 35 * , 1399 ( 1975 ) .
g. grinstein and s. k. ma , phys .
. lett . * 49 * , 685 ( 1982 ) .
j. f. fernandez , g. grinstein , y. imry , and s. kirkpatrick , phys .
lett . * 51 * , 203 ( 1983 ) .
j. z. imbrie , phys .
* 53 * , 1747 ( 1984 ) .
j. bricmont and a. kupiainen , phys .
lett . * 59 * , 1829 ( 1987 ) .
i. b. ferreira , a. r. king , v. jaccarino , j. l. cardy , and h. j. guggenheim , phys .
b * 28 * , 5192 ( 1983 ) .
h. yoshizawa , r. a. cowley , g. shirane , r. j. birgeneau , h. j. guggenheim , and h. ikeda , phys .
48 * , 438 ( 1982 ) .
t. schneider and e. pytte , phys . rev .
b * 15 * , 1519 ( 1977 ) .
d. andelman , phys .
b * 27 * , 3079 ( 1983 ) .
a. aharony , phys .
b * 18 * , 3318 ( 1978 ) .
d. c. mattis , phys .
* 55 * , 3009 ( 1985 ) .
h. e. borges and p. r. silva , physica a * 144 * , 561 ( 1987 ) .
e. f. sarmento and t. kaneyoshi , phys .
b * 39 * , 9555 ( 1989 ) .
r. m. sebastianes and w. figueiredo , phys .
b * 46 * , 969 ( 1992 ) .
t. kaneyoshi , physica a * 139 * , 455 ( 1985 ) .
. akinci , y. yksel and h. polat , phys .
e * 83 * , 061103 ( 2011 ) .
d. p. landau , h. h. lee , and w. kao , j. appl .
phys . * 49 * , 1356 ( 1978 ) .
j. machta , m . e. j. newman , and l. b. chayes , phys . rev .
e * 62 * , 8782 ( 2000 ) .
n. g. fytas , a. malakis , and k. eftaxias , j. stat .
theory exp .
( 2008 ) , 03015 .
n. g. fytas and a. malakis , eur .
j. b * 61 * , 111 ( 2008 ) .
m. gofman , j. adler , a. aharony , a. b. harris , and m. schwartz , phys .
b * 53 * , 6362 ( 1996 ) . b. k. chakrabarti and m. acharyya , rev .
phys . * 71 * , 847 ( 1999 ) , and the references therein .
t. tom and m. j. de oliveira , phys .
a * 41 * , 4251 ( 1990 ) .
w. s. lo and r. a. pelcovits , phys .
a * 42 * , 7471 ( 1990 ) .
m. acharyya and b. k. chakrabarti , phys .
b * 52 * , 6550 ( 1995 ) . m. acharyya , physica a * 235 * , 469 ( 1997 ) .
m. acharyya , phys .
e * 56 * , 2407 ( 1997 ) . m. acharyya , phys .
e * 58 * , 179 ( 1998 ) .
m. acharyya , phys .
e * 59 * , 218 ( 1999 ) .
x. shi , g. wei and l. li , phys .
a * 372 * , 5922 ( 2008 ) b. deviren , o. canko and m. keskin , chin .
b * 19 * , 050518 ( 2010 ) .
q. jiang , h. n. yang and g. c. wang , phys .
b * 52 * , 14911 ( 1995 ) .
d. t. robb , y. h. xu , o. hellwig , j. mccord , a. berger , m. a. novotny and p. a. rikvold , phys .
b * 78 * , 134422 ( 2008 ) .
j. hausmann and p. rujn , phys .
lett . * 79 * , 3339 ( 1997 ) .
g. l. s. paula and w. figueiredo , eur .
j. b. * 1 * , 519 ( 1998 ) .
m. acharyya , phys .
e * 58 * , 174 ( 1998 ) .
n. crokidakis , phys . rev .
e * 81 * , 041138 ( 2010 ) .
e. costabile and j. r. de sousa , phys .
e * 85 * , 011121 ( 2012 ) .
r. j. glauber , j. math .
* 4 * , 294 ( 1963 ) .
r. honmura and t. kaneyoshi , j. phys .
c * 12 * , 3979 ( 1979 ) . t. kaneyoshi , acta phys . pol . a * 83 * , 703 ( 1993 ) .
i. tamura and t. kaneyoshi , prog .
. phys . * 66 * , 1892 ( 1981 ) . | nonequilibrium behavior and dynamic phase transition properties of a kinetic ising model under the influence of periodically oscillating random - fields have been analyzed within the framework of effective field theory ( eft ) based on a decoupling approximation ( da ) .
dynamic equation of motion has been solved for a simple cubic lattice ( @xmath0 ) by utilizing a glauber type stochastic process .
amplitude of the sinusoidally oscillating magnetic field is randomly distributed on the lattice sites according to bimodal and trimodal distribution functions . for a bimodal type of amplitude distribution , it is found in the high frequency regime that the dynamic phase diagrams of the system in temperature versus field amplitude plane resemble the corresponding phase diagrams of pure kinetic ising model .
our numerical results indicate that for a bimodal distribution , both in the low and high frequency regimes , the dynamic phase diagrams always exhibit a coexistence region in which the stationary state ( ferro or para ) of the system is completely dependent on the initial conditions whereas for a trimodal distribution , coexistence region disappears depending on the values of system parameters .
pacs numbers : : 03.65.vf , 64.60.ht , 75.10.nr , 75.30.kz |
in the last few years , important developments in canonical quantum gravity have taken place , culminating with the recent formulation of two mathematically consistent @xcite , non - trivial canonical quantizations of general relativity in terms of ashtekar new variables and spin networks .
instrumental in these developments have been the underlying advances in mathematical techniques for dealing with infinite dimensional nonlinear spaces , like the theory of cylindrical functions and associated measures @xcite , and the introduction of spin networks to eliminate the over - completeness of the wilson loop basis @xcite ( for a recent summary , see the review article by rovelli @xcite ) . in spite of the advances ,
most of the results obtained from these theories up to now concern statements made _ at a kinematical level _
( without imposing the hamiltonian constraint ) .
in order to start discussing if these theories contain the correct semiclassical physics and are therefore physically viable theories of quantum gravity , we need methods to introduce semiclassical states ( encouraging recent kinematical results can be seen in @xcite ) but also to probe their dynamics . in particular a current open issue is if the semiclassical physics can be discussed at a kinematical level only or requires the imposition of the correct dynamics . probing the dynamics has always been a problem in canonical general relativity @xcite . in any gauge theory ,
the only physically relevant quantities are those that are invariant under the gauge symmetries of the theory , or in the canonical language , that have vanishing poisson brackets with the constraints that represent the gauge symmetries . at a quantum mechanical level , the physical states are those that are annihilated by the constraints .
in general relativity we do not have any example of a quantity that classically has vanishing poisson brackets with the constraints , at least in the case of a compact manifold .
in fact , it is strongly suspected that such quantities will never be constructed in closed form @xcite .
finding quantum states that are annihilated by the constraints is also a challenge .
most researchers in this area do not consider these problems a fundamental obstruction , since in order to do meaningful physics one only needs approximate expressions for the observables , as was advocated long ago by bergmann and newman @xcite .
finding an approximation method that does not destroy the non - perturbative nature of the canonical treatment is , however , a challenge .
we have recently proposed one such method @xcite .
it is based on considering general relativity coupled to a cosmological constant and taking the @xmath0 limit @xcite .
the construction goes as follows : for general relativity with a cosmological constant , the hamiltonian constraint reads , @xmath1 where @xmath2 is a smearing function , @xmath3 is the hamiltonian constraint without a cosmological constant and @xmath4 is the determinant of the spatial metric .
if one now considers the limit @xmath0 , and re - scales the constraint by @xmath5 , one is left with a theory , which we will call `` zeroth order '' theory for which the hamiltonian constraint is , @xmath6 that is , the hamiltonian constraint is just the square root of the determinant of the spatial metric .
in addition to this the theory has the ordinary diffeomorphism constraint and if one uses ashtekar variables there will also be a gauss law constraint ( both are independent of @xmath7 ) . imposing classically the hamiltonian constraint of the zeroth order theory , one immediately sees that it corresponds to metrics of identically vanishing determinant , that is , degenerate metrics .
it is easy to construct quantities that have vanishing poisson bracket with the hamiltonian constraint of the zeroth order theory .
for instance , one can consider any function depending only on the three metric ( or if one is using ashtekar variables @xmath8 , the densitized triads @xmath9 ) and not on its canonically conjugate momenta . to have a genuine observable the quantity should also have vanishing poisson bracket with the diffeomorphism constraint .
there are standard ways of achieving this coupling the theory to matter @xcite .
a novel point is that in our approach , since the extra matter couplings in the hamiltonian are higher order in @xmath10 these techniques yield genuine observables for the zeroth order theory , whereas normally they just construct quantities that have vanishing poisson bracket only with the diffeormophism constraint but not with the hamiltonian constraint .
since we will not need this technique for the examples we consider in this paper , we do not include further details here ( see a discussion in @xcite ) .
if one assumes that the observables of the theory are power series in the inverse cosmological constant , @xmath11 and one requests that these observables have vanishing poisson brackets with the hamiltonian of the theory ( more precisely , one would like that the poisson bracket be proportional to the constraints of the theory , for simplicity we will just demand that they vanish , but it is immediate to extend the construction to the case in which they vanish on - shell)and we expand such requirement in powers of @xmath10 one gets @xmath12 the first equation determines @xmath13 and the second one leads to a linear partial differential equation for @xmath14 . the construction can be readily continued to higher orders . in all cases
one obtains a linear partial differential equation , albeit with a more and more complex inhomogeneous term .
it should be noted that one can obtain many observables starting with different solutions to the first equation .
the linear partial differential equations are not hard to solve , given that the coefficients of the derivatives are functions of the triads , whereas the derivatives are with respect to the connections .
given the simplicity of it , the system is always integrable ( although the solutions might be pathological , as we see in the examples later in the paper ) and therefore yields all the observables of the theory .
this amazing simplicity also has a quantum counterpart , at the time of finding states that are annihilated by the hamiltonian constraint .
if we now focus on the formulation of the quantum constraints in terms of the spin network representation of quantum gravity , for instance as discussed by thiemann @xcite , one starts by considering our zeroth order hamiltonian , which is closely related to the well - understood volume operator .
it is easy to find eigenstates for this operator , in particular ones with vanishing eigenvalues .
when one considers the first order corrections , because the zeroth and first order hamiltonians have a well defined algebraic action in the space of spin network states , one gets a completely manageable ( albeit complicated ) system of equations .
we will not go into details of the quantum theory of full general relativity in this paper , since it would make the paper much less readable as we would have to introduce quite a bit of ( readily available but quite detailed ) technology of spin networks .
we will only present , for completeness , a brief of what the issues in the full quantum theory are in the last section of the paper .
let us however describe in general lines how the quantum perturbative approach would be for a generic system with a _
hamiltonian constraint ( we will discuss the case of more than one hamiltonian as a model of what can happen in a field theory in a subsequent section ) .
consider a system where the quantum hamiltonian constraint is of the form @xmath15 where @xmath16 is a parameter that we consider small .
we then consider quantum states that admit a power series expansion in powers of @xmath16 , @xmath17 what one wants to do now is to solve the theory perturbatively .
this requires some comments . what does it mean to solve perturbatively a quantum mechanical constrained system
the approach we will take is the following : consider the eigenvalue equation for the hamiltonian constraint , @xmath18 and consider its solution up to a certain order in @xmath16 , finding the zero - energy eigenstate within such a theory .
that is , what we end up doing is bound - state perturbation theory ( provided the spectrum of @xmath19 is discrete ) and finding one set of particular states , those with zero eigenvalue .
a good example to bear in mind about the method is to consider an atom in a ( weak ) magnetic field and think that one is looking for states with a given energy .
if one goes to a certain order in perturbation theory in terms of the magnetic field , one will be able to achieve the given energy level only for certain values of the magnetic field .
that is , in this approach the perturbative parameter is quantized .
it is also clear that the perturbed states that achieve a certain energy level for a given value of the perturbative parameter do not necessarily stem from perturbing the zeroth order theory level of that particular energy .
it is important to notice this in the gravitational case , where one looks for states with vanishing eigenvalue .
the perturbed states of interest will in general not have @xmath20 .
so we have cast the problem of solving quantum mechanically a constrained system as finding perturbatively an eigenstate for a given hamiltonian in a quantum mechanical system .
this problem is well defined and , as we will see in the examples , solvable .
for the full general relativity case , the zeroth order hamiltonian is quite close to the volume operator , so we do have a discrete spectrum to start with .
however , because it is a field theory , one really has infinitely many hamiltonians and certain consistency requirements have to be met .
therefore the quantum mechanical analogy we made may turn out to be too nave .
the last example we consider in this paper is geared towards clarifying that point .
why should one care about a regime in which @xmath7 is large in planck units , given that the current observables value in such units is @xmath21 ? .
our attitude is the following : we have now at least two candidates for theories of quantum gravity .
we need to probe them in any regime we can get a handle on in order to help decide if the theories are plausible or not from a physical stand point .
if we can do this , we can then see in detail what regime of @xmath7 we can really access .
perhaps it is a bit better than what one initially expects and the results can be of interest to early universe cosmology . or perhaps one is lucky enough to encounter a situation similar to other power series expansions ( like the strong coupling limit in the lattice or the large @xmath22 expansion in qcd ) where the results are robust and useful way outside the navely expected regime .
we will see some hints of this in one of the examples of this paper .
the above observations make the whole approach worthwhile exploring further .
however , the surprising way in which it appears to deal in a simple way with questions that in the full theory seem unapproachable raises a certain level of skepticism .
several pointed questions can be asked that make the whole approach look questionable . among them , \a ) does the approximation have any hope of working for values of @xmath7 smaller than the planck scale ?
\b ) general relativity is a complex enough dynamical system , that some pathologies might be expected .
for instance , solutions might exhibit some level of chaos .
it could also happen that other pathologies appear , like the full theory having less observables than the zeroth order one .
does our method always produce @xmath23 observables ( where @xmath24 is the number of degrees of freedom and @xmath25 the number of constraints ) ? with a method that reduces everything to linear ode s , is there not a risk of `` papering over '' these subtleties .
\c ) what sort of approximation is one getting at a quantum level ? what is meant by an approximate state , or an approximate solution to a quantum constraint ?
\d ) when one applies the quantum perturbative approach in a field theory , one effectively is dealing with infinitely many hamiltonian constraints .
is the method still viable or one gets an inconsistent set of conditions on the perturbed states , leading to an empty theory ?
\e ) the zeroth order hamiltonian constraints we consider seem to admit many more quantum states than the full hamiltonian constraints .
does the method produce spurious states for the full theory or correctly notices the further limitations that should appear at higher orders ?
this paper attempts to provide ( at least partial ) answers to these questions by probing the perturbative method in model systems .
in section ii we will consider the application of the technique to bianchi models .
this will allow us to probe questions a ) and b ) . from bianchi models
we will learn that , surprisingly , when one evaluates the approximate power series that we get for the observables on solutions to the equations of motion , the cosmological constant drops off from the expressions .
evaluated on the solutions to the equations of motion , the expression for the observables become a sum of terms that approximate better and better a constant of motion without reference to the cosmological constant . for a bianchi
i recollapsing universe we will explicitly demonstrate that with only a few terms one finds an expression that maintains an almost constant ( less than @xmath26 variation ) value across @xmath27 of the life of the universe .
we will also see that in the case of bianchi ix one can easily construct observables , but it is likely that they will remain good approximations only for short periods of time when one enters the chaotic domain .
we will also see that one does not obtain more observables than the required ones . one can obtain many _ apparently different _
power series with the method but they just represent reparametrizations of the basic observables .
we will also learn how to deal with possible singularities in the expressions obtained for the observables as functions of the phase space . in section iii we will discuss the heavy asymmetric top in three dimensions .
this is a hamiltonian system that can be viewed as a pathological constrained system if one considers it for a fixed value of the energy .
it is pathological in that it has less observables than four .
we will see that our method _ only apparently _ generates too many observables . on close inspection
we will find that several of the expressions are really not good candidates for observables because they are expressions that have support only on limited regions of phase space , that are in general not preserved by evolution .
in the end , the method produces the correct number of observables for the model . in section iv we
explore a model consisting of two harmonic oscillators with constant energy difference .
this system has received quite a bit of attention , for instance see @xcite . unless the ratio of the frequencies of the two oscillators is a rational number , the system is pathological in that it only admits one quantum state .
we will show that our perturbative approach indeed notices this fact and produces correct approximations for the quantum states in all cases . in section
v we explore a model of coupled harmonic oscillators with two `` hamiltonian '' constraints .
general relativity being a field theory one has infinitely many hamiltonian constraints ( one per spatial point ) .
therefore the other models we consider in the paper seem a bit short in mimicking this feature , since they are mechanical systems with only one constraint . in this section
we increase modestly this feature by considering a model with two hamiltonians .
we will see that the method deals with no problem with such a system .
we will end with a short reflection on the implications of the results of this paper and what challenges remain for the application of the approximate scheme in the context of full classical and quantum general relativity .
as a first example of the application of the technique , we will discuss the issue of finding observables for bianchi models .
bianchi models are homogeneous cosmologies , and as a consequence the einstein equations become ordinary differential equations .
this allows to achieve , in some cases , quite a bit of progress towards their solution .
following @xcite , we represent the bianchi models using as variables the ( diagonal ) components of the triads @xmath28 and the canonically conjugate momenta @xmath29 .
all class - a bianchi models admit a simple hamiltonian formulation in terms of these variables , but we will concentrate our attention in two models , the bianchi i model and the bianchi ix model .
the former is the simplest model and the latter the one with the richest dynamics .
bianchi i models start from an initial singularity and expand anisotropically . depending on the sign of the cosmological constant they may expand forever or recollapse .
bianchi ix models approach the singularity through an infinite sequence of oscillations of increasing complexity . in terms of the ashtekar variables
the bianchi i and ix models with a cosmological constant are described by a hamiltonian constraint , @xmath30 where @xmath31 corresponds to the bianchi ix model and @xmath32 to the bianchi i model .
the gauss law and diffeomorphism constraints are identically satisfied . to apply our construction we need to start by choosing an observable for the zeroth order theory . in general such quantities
are given by @xmath33 .
if one sets out to find four independent observables one could start , for instance , with the following independent choices for zeroth order observables , @xmath34 let us now concentrate on the study of the bianchi i cosmology . in this case
@xmath35 are both exact observables for the full model with a cosmological constant .
starting with @xmath36 one finds the following first order correction , @xmath37 and in addition to this one has the solution to the homogeneous equation , which is given by an arbitrary function that commutes with @xmath38 , that is , similar to @xmath13 .
the expression for the bianchi ix model is very similar to the one we just introduced , it has an extra term @xmath39 . in order to investigate the behavior of the approximate observable we are considering
we will evaluate it on an exact solution of the equations of motion , the metric of reference @xcite , @xmath40\ ] ] where , @xmath41 and where @xmath42 and where @xmath43 .
this solution corresponds to a bianchi i universe that expands out of a big bang at @xmath44 but after a while the cosmological constant leads it to recollapse at a time @xmath45 .
the ashtekar variables for this metric ( particularized to @xmath46 ) read , @xmath47{2}\sqrt [ 3]{\tan(1/2\,\omega\,t ) } \left ( \sin(\omega\,t)\right ) ^{2/3}}{\omega}}\\ a_1&= & 1/6\,{\frac { \sqrt [ 3]{2}\left ( 2\,\sin(\omega\,t ) \left ( \cos(1/2\ , \omega\,t)\right ) ^{2}-3\,\sin(\omega\,t)-4\,\sin(1/2\,\omega\,t)\cos ( \omega\,t)\cos(1/2\,\omega\,t)\right ) } { \sqrt [ 3]{\sin(\omega\,t ) } \left ( \sin(1/2\,\omega\,t)\right ) ^{2/3}\left ( \cos(1/2\,\omega\,t )
\right ) ^{2}}}\\ a_2&= & a_3=1/3\,{\frac { \sin(1/2\,\omega\,t)\omega } { \sqrt [ 3]{\cos(1/2\,\omega\,t)}}}.\end{aligned}\ ] ] as we can see , for the solution in question @xmath48 and @xmath49 vanish at @xmath44 .
if we look at the expression for the first order correction we found , this implies that it diverges for @xmath44 .
this is an example of what we stated earlier , namely that the expressions for the observables we found can sometimes be singular in certain points of phase space .
one can construct a first order correction that is well behaved at @xmath44 by making use of the free function @xmath50 that solves the homogeneous equation . if one adds to the first order correction the solution to the homogeneous equation given by @xmath51
, it becomes , @xmath52 and this quantity is obviously well behaved at @xmath44 .
if one now proceeds to compute the second order correction , one finds a similar situation with respect to the divergence at @xmath44 .
this can be corrected by adding the function @xmath53 and one obtains for the correction , @xmath54 and a similar subtraction yields the third order correction , @xmath55 if we now evaluate the above expressions for the exact solution we discussed above one gets ( @xmath56 ) , @xmath57 \left[\cos^{8/3}({x\over 2 } ) \left(-1+\cos^{2}(x)\right ) \right]^{-1}\nonumber\\ { o^{(3)}_1 \over \lambda^3 } & = & { \frac { 1}{810 } } \left [ -568\cos^{6}({x\over 2})-15 + 15\cos^{2}(x ) -194\cos^{4}({x\over 2 } ) + 488\cos^{6}({x\over 2})\cos^{2}(x ) + 354\cos^{4}({x\over 2})\cos^2(x)\right.\nonumber\\ & & \left.-47\cos^{2}({x\over 2})\cos^2(x ) -33\cos^2({x\over 2 } ) -16\cos^3({x\over 2})\sin(x)\cos(x ) \sin({x\over 2 } ) -72\cos({x\over 2})\sin(x)\cos(x)\sin({x\over 2})\right.\nonumber\\ & & \left.+88\cos^5({x\over 2})\sin(x)\cos(x ) \sin({x\over 2})\right ] \left[\cos^{14/3}({x\over 2 } ) \left(-1+\cos^{2}(x)\right ) \right]^{-1}.\end{aligned}\ ] ]
the first observation is that the above expressions for the correction have lost the dependence on @xmath7 stemming from the perturbative approach ( there is a trivially re - scalable dependence on @xmath7 through @xmath56 ) .
therefore , if the above expressions converge , the convergence will be independent of @xmath7 .
this can be seen in the following plot , in which we show the observables as functions of @xmath58 .
we see that in spite of the lack of dependence in @xmath7 , the curves do converge , the third order one maintaining an almost constant value for the longest period of time ( more than half the lifetime of the universe ) .
the various corrections are not well behaved near the big crunch .
it should be remembered that the above calculations were particularized for @xmath46 .
the value of the observable will in general be @xmath59-dependent . one way of understanding
this is that the `` physical '' information content of the observable is to give a measure of the @xmath59 parameter , which is associated to the anisotropy of the model .
for instance , the exact observable @xmath60 takes the value @xmath61 .
since the solution to the equations of motion we are considering is parameterized by a single constant , all observables become given functions of that constant . as was shown in the above calculations , the bianchi ix model appears as only slightly more complex , and similarly any class - a bianchi model
can be treated in the same way .
because of the local nature of our calculations ( in time ) , it is not surprising to find observables in the bianchi ix case . in that cosmology , because of the chaotic oscillations the dynamics undergoes , it is unlikely one will be able to find expressions for the observables that are global , but expressions like the ones we find should face no difficulty . as we saw , the zeroth order observable one starts from has a huge degree of ambiguity .
there is also a lot of ambiguity in each perturbative step , due to the presence of the solution of the homogeneous equation .
it is evident that certain choices will lead to more useful ( or in some cases shorter or even polynomial ) expressions , as we exhibited when we manipulated the homogeneous solutions to avoid expressions that are singular at certain points in phase space .
summarizing , we learn the following lessons from the bianchi example we have just discussed .
first of all , we see that the approximation for the observables we get , when evaluated on a solution of the equations of motion , becomes independent of the cosmological constant .
this suggests that the perturbative approach will remain good even when the constant is small .
in fact , we see in the example , that the observable is well approximated for a significant fraction of the lifetime of the universe in question .
we also see that the presence of chaos in the bianchi ix model does not preclude us from finding approximate expressions for the observable , though it is likely that it will impose further restrictions on the domain of validity ( in terms of evolution time ) of the approximate observable .
it is possible to test these hypotheses with numerical evolutions of the bianchi ix cosmology .
as another example of the application of the perturbative hamiltonian ideas we are proposing , we analyze a model which in spite of its simplicity , adds pathological behaviors that are interesting to analyze under the proposed scheme .
the model is an asymmetric heavy top with constant energy .
this system only has one observable , the momentum canonically conjugate to the angle around the vertical axis .
we will assume that the asymmetry in the top is small and treat this model as a perturbation of a symmetric heavy top , which has two observables .
will our technique notice that one loses an observable when one loses the symmetry ?
consider a heavy symmetric top with moments of inertia @xmath62 and @xmath63 .
now suppose one sticks on two small masses at opposite sides of the surface of the top , forming a line perpendicular to and intersecting the top s third axis at its center of mass .
then the moments of inertia along directions @xmath64 and @xmath65 will be slightly different . let us quantify this as @xmath66 , with @xmath16 a small perturbation ( we use the lowercase to avoid confusion with the cosmological constant , although the parameter plays the same perturbative role in this problem as the inverse cosmological constant in full general relativity ) .
the lagrangian for such a system will be given , in terms of the traditional euler angles ( see for instance @xcite ) , @xmath67 where @xmath68 is the length along the top third axis from the top base up to the center of mass . to introduce a canonical formulation ,
we compute the canonical momenta , @xmath69 and similar equations for the other momenta , which yields an invertible linear system of equations from where one can compute the velocities as functions of the canonical momenta .
one can then perform a legendre transform and obtain a hamiltonian @xmath70 . for small values of the perturbative parameter , one can expand , @xmath71 where the general form of the hamiltonians is , @xmath72 with @xmath73
. we would like to treat the system as an example of constrained system .
therefore we request that the energy have a given value @xmath74 , that is , @xmath75 before treating the case of the top with constant energy , let us first recall how one formulates the usual asymmetric top , but as a parameterized system .
this allows a more natural contact with the constrained systems we will handle later .
the parameterized top is obtained by introduce a time @xmath76 and its canonically conjugate momentum @xmath77 , such that , @xmath78 where @xmath79 . the observables are defined by , @xmath80 which yields , @xmath81 this system is solved by finding the solution to the following set of differential equations , @xmath82 the first of these equations can be integrated immediately , @xmath83 .
the other two equations are just hamilton s equations of motion , which integrated yield , @xmath84 if one substitutes @xmath85 by @xmath86 in the above expressions , one obtains the six constants of motion , @xmath87 which are observables of the parameterized system .
a crucial element in the solution of this system , that allows to obtain the six constants of motion explicitly is that the solution of the equation for @xmath85 in terms of @xmath76 is straightforward .
we will see that this poses problems in the next example .
let us now study the case of interest , the ( slightly ) asymmetric heavy top , with hamiltonian constraint ( [ 31 ] ) and assuming observables of the form @xmath88 .
let us start with the equations to zeroth order in @xmath16 , @xmath89 with @xmath90 which leads to the pde , @xmath91 which again can be solved by the substitution technique , @xmath92 at zeroth order the system has two trivial constants of motion , @xmath93 one can in principle integrate the remaining equations to yield , @xmath94 as before , one can ( formally ) invert the solutions to obtain the constants of motion as functions of phase space by solving for @xmath85 and substituting back in the following equations : @xmath95 we therefore appear to have found again six constants of motion as in the case we studied before .
however , the above equations involve periodic functions with arguments with incommensurate periods .
one will only determine @xmath85 up to multiples of the period and when substituting in the other equations will end up with multivalued functions .
if one chooses a particular branch , the resulting observables only take values on a portion of the accessible phase space .
they are therefore not acceptable as observables for the system .
this is an important observation , since we will see that this is the mechanism which prevents our technique from generating observables for systems that do not have them .
let us now consider the first order corrections .
the equation to be solved are , @xmath96 and if we start from @xmath97 , we are left with , @xmath98 now the pde is non - homogeneous .
we can still solve it with our parameterization technique via , @xmath99 as before , we will need to solve for @xmath85 and replace it in the equations to obtain expressions for the initial values as functions of phase - space .
we do run , however , into the same difficulty as before , we can not obtain a uniquely defined expression for @xmath85 and therefore we can not construct a first order correction for the observable chosen .
the procedure therefore fails to work . what happens if we choose @xmath100 instead ?
in that case @xmath101 and therefore the equation for @xmath102 indeed has a solution .
we will therefore recover , order by order , an expression for the _ single _ observable the system admits .
the general conclusion seems to be that , although the linear pde s of our procedure in principle can always be solved , requiring that the solution be a well defined function on phase space ( and not only on a portion of it ) severely restricts the observables that can be obtained by the technique , and apparently prevents the technique from generating spurious observables in systems with less observables than degrees of freedom .
let us now consider examples of the application of the method at a quantum level .
hajek and others @xcite have considered a model of constrained system , consisting of two harmonic oscillators with a fixed energy difference .
the possible eigenstates of the system therefore consist ( if one uses the standard inner product for the oscillators ) in situations in which oscillators have the appropriate quanta of energy to satisfy the constraint that the difference of their energies be constant .
it is clear that this can only happen if the ratio of the frequencies of the oscillators is rational .
otherwise , the system would admit at most one quantum state .
this is the pathology of this system : if the ratio of the frequencies is not rational , the quantum system only has one state satisfying the energy constraint .
if it is rational , the system admits infinitely many states .
one could therefore consider starting with a zeroth order model consisting of two coupled oscillators with rational ratio of frequencies and perturb the system by adding the hamiltonian of two oscillators with an irrational ratio .
the zeroth order system has infinitely many quantum states and the perturbed system has only one state .
can our method detect and correctly handle this situation ? this is what we attempt to probe in this section .
the hamiltonian constraint for the system in question simply reads , @xmath103 where we have chosen the energy difference equal to one .
the quantization of this model , viewed as a constrained system , has order ambiguities . consider the transformation @xmath104 the constraint reads , @xmath105 and introducing the complex variables , @xmath106 the constraint takes the form , @xmath107 we can then proceed to the quantization , @xmath108{}&=&i \delta_{jk}\\ \alpha_j\rightarrow \hat{a}_j,&\quad & \alpha^*_j\rightarrow \hat{a}^\dagger_j,\end{aligned}\ ] ] and the quantum constraint can be written as @xmath109 , with , @xmath110 with @xmath111 representing the ordering ambiguities . to obtain the quantum states that are annihilated by this constraint ,
let us consider separately the spectra of both @xmath112 s .
since they are harmonic oscillators , we have , @xmath113 with @xmath114 arbitrary ordering - dependent constants . in order to solve the constraint , the values of @xmath115 must satisfy , @xmath116 to produce solutions
, one can choose the constants @xmath117 .
for instance , for @xmath118 , we get @xmath119 . in order to have more than one solution
, one needs to have , @xmath120 or , @xmath121 that is , the ratio of the frequencies of the oscillators must be a rational number . if @xmath122 then @xmath123 , @xmath124 with @xmath125 an integer and the wavefunctions of the physical space are @xmath126
. we will now construct a model that is a perturbation of the above considered one but is still exactly solvable .
we can then solve the model exactly and perturbatively and compare the results .
let us consider the system , @xmath127 this system is equivalent to the original hamiltonian constraint if we define @xmath128
. we can therefore readily solve it . for system to have more than one state
, we have , @xmath129 or , @xmath130 therefore @xmath131 and @xmath132 with @xmath125 integer , and remembering that @xmath133 , in order to have a solution the perturbative parameter has to have a given value , @xmath134 in order to compare with the perturbative calculation we will perform in the next paragraphs , let us consider the condition for having more than one solution for this model , @xmath135 which in the case @xmath136 reads , to first order , @xmath137 or , @xmath138 in order to have @xmath139 , the ratio @xmath140 must be a rational number close to @xmath141 .
we can therefore write it as , @xmath142 with large @xmath125 . expanding , we therefore get , @xmath143\sim 2 { \ell \over n r } -2 { m\over m r } + o(r^{-2}).\ ] ] summarizing , the perturbed model can have infinitely many or a single quantum state , depending on the value of @xmath16 .
let us now see if treating the model perturbatively we recover the same result .
we consider the zeroth and first order hamiltonians as , @xmath144 we start by considering the eigenvalue problem up to first order in @xmath16 , @xmath145 let us start by considering the resulting zeroth order equation , @xmath146 the spectrum of @xmath38 is , @xmath147 here we need to distinguish if @xmath148 is rational or not . if it is rational , then the spectrum is degenerate and we need to apply ordinary degenerate bound state perturbation theory as described in any quantum mechanics textbook .
if it is not rational , the spectrum is not degenerate . as we discussed in the introduction ,
let us concentrate on the case in which the frequencies are related by @xmath149 .
the degeneracy of the spectrum can be directly seen in that it takes the same values for @xmath150 , @xmath151 with given @xmath152 and arbitrary values of @xmath125 , that is , @xmath153 the eigenstates of @xmath19 are , @xmath154 in the number representation for the harmonic oscillators .
we therefore find that the solution to zeroth order is given by ( at the moment ) an arbitrary combination of the given eigenstates , @xmath155 let us now consider the equation to first order in @xmath16 , @xmath156 since @xmath157 to have , @xmath158 we now take the equation ( [ firstlambda ] ) and project it onto the the eigenspace with eigenvalue @xmath159 . while doing this , the contribution from @xmath160 acting on this space cancels with the term @xmath161 acting on the same space .
we are therefore left with , @xmath162 and recall that @xmath163 and in turn , @xmath164 since the matrix elements of @xmath165 are between states of energy @xmath159 , @xmath166 and @xmath167 do not contribute and the hamiltonian is diagonal .
therefore of all the terms in the superposition defining @xmath168 we are left with @xmath169 , which normalizing yields , @xmath170 and energy , @xmath171 therefore the first order corrected energy is , @xmath172 and as before we need to choose @xmath173 in such a way that @xmath174 is a solution of the theory up to the order we are considering , i.e. , @xmath175 .
if we want other solutions , we need to make the energy eigenvalue vanish , up to the order we are considering this implies , @xmath176 for small values of @xmath16 we should recover the results of the exact calculation . we notice that @xmath16 decreases with increasing @xmath125 . in that limit
we therefore have that , @xmath177 and this agrees with the expansion in small @xmath16 of our exact calculation .
it should be noted that the above calculation shows that starting from a given state in the zeroth order theory and choosing a perturbative parameter , we get a first order solution to the constraint .
it is immediate to see however , that this can be accomplished by starting from an infinite number of sets in the zeroth order theory .
simply consider a state obtained by multiplying @xmath152 and @xmath125 times an integer .
the same value of lambda in equation ( [ valuelambda ] ) ensures that the first order state is a solution .
the calculation can be completed by considering the projection of ( [ firstlambda ] ) on the subspace of states orthogonal to the states of energy @xmath159 .
such projection determines the correction to the state , @xmath178 .
this can be straightforwardly done .
the calculation can be repeated for irrational quotients of frequencies . in that case one
is doing ordinary ( non - degenerate ) bound state perturbation theory .
the calculations resemble very much the ones we did above , so we will not detail them here .
some comments are nevertheless in order .
the zeroth order theory has only one state when the frequency ratio is irrational and many states when it is rational .
in the perturbed theory , in both cases we can find values of @xmath16 such that the theory has many states .
that is , the perturbed model we constructed is such that depending on the value of the perturbative parameter one has either one or many states . if we do not choose the correct value of lambda ( as determined by equation ( [ valuelambda ] ) ) , then the model only admits the state @xmath174 as solution .
in particular , we have correctly addressed the situation posed in the introduction to this section : if the zeroth order model has rational frequencies ( and therefore infinite solutions ) but the peturbed model corresponds to a value of @xmath16 that only admits one solution , we will not satisfy the energy constraint to first order and we are therefore left with the vacuum as the only solution ( in the factor ordering chosen ) . therefore the method seems to handle well the drastic reduction in number of states implied by the perturbation .
it is suggestive that the model considered has a quantized value of @xmath16 in the full treatment if one wishes to have infinitely many states . that behavior is reproduced correctly in the perturbative approach
however , the quantization of the perturbative parameter is a feature of the perturbative method as long as the zeroth order hamiltonian has a discrete spectrum ( which is the only case in which we can apply the method ) .
it is interesting to notice that in the other models considered by hajicek , the spectrum of the hamiltonian is continuous and therefore we can not apply our method , and the full models do not have any particular requirements on the frequencies of the oscillators .
another lesson from this example is that the lack of states in the perturbative treatment strongly suggests that the full model also lacks quantum states , unless there is some sort of non - analytical behavior in terms of the perturbative parameter .
in the case of quantum gravity , one is dealing with a field theory , therefore one has an infinite number of constraints . in the spin network representation ,
if one considers networks with a finite number of vertices , the number of equations resulting from the constraints might be finite , but in realistic situations of semi - classical interest it will still be large . therefore the quantum models we have considered up to now , which are quantum mechanical systems with one constraint , may not capture some of the aspects present in the gravitational case .
it might be that dealing with a larger than one number of constraints implies further relations between the perturbative coefficients that make the system incompatible .
it is somewhat unlikely that this will happen , since after all , assuming the full theory is solvable , if the solution admits a power series expansion in @xmath7 , one should find a solution that is acceptable perturbatively to any order in perturbation theory .
nevertheless , it might be useful to consider a simplified model with more than one constraint to make sure the technique works .
it is in particular interesting to see how , if one has several constraints to satisfy , one can achieve it with the same value of the perturbative parameter for all the equations .
this is what we attempt in this section .
consider three different harmonic oscillators , @xmath179 and define the two `` hamiltonian constraints '' , @xmath180 the two constraints have vanishing poisson brackets among themselves . we will have to request that the perturbed hamiltonians also have vanishing poisson brackets for all values of @xmath16 and this will impose restrictions on our perturbative approach . as we discussed in the previous section , the eigenstates of the quantum version of the constraints are simply given in the number representation by @xmath181 , @xmath182 it is possible to choose the @xmath183 in such a way that @xmath184 has zero eigenvalue .
the system admits other solutions only if @xmath185 which leads to a system of equations for the @xmath186 , @xmath187 that has as general solution , @xmath188 with @xmath68 an arbitrary integer .
therefore the physical states of the theory ( states annihilated by the constraints ) are , @xmath189 we will now consider a perturbative term that still allows the system to be solved exactly , @xmath190 so , @xmath191 the modification introduced by the perturbation is equivalent to changing @xmath192 by @xmath193 . if we now consider the construction of quantum states @xmath181 that are annihilated by both hamiltonian constraints , and seek for values of @xmath186 that differ only slightly from the unperturbed ones , which is achieved by choosing @xmath194 , @xmath195 we find that , @xmath196 the joint solution of these equations requires that @xmath197 , and assuming that @xmath16 is small , we find that , @xmath198 which implies that @xmath199 we recall the form of the unperturbed hamiltonian and the first order perturbations , @xmath200 with @xmath201 and @xmath202 , and we wish to solve for quantum states , @xmath203 the spectrum of @xmath38 is given by ,
@xmath204+\hbar\left(k_1+k_2-k_3\right),\\ \epsilon^{(0)}_2 & = & \hbar \omega_1 \left [ n_1 -{m\over n } n_2 + 2{p\over q } n_3\right]+\hbar\left(k_1-k_2 + 2k_3\right).\end{aligned}\ ] ] once more , we have a degenerate spectrum , with , @xmath205 so the energies are , @xmath206 + \hbar\left(k_1+k_2-k_3\right)\\ \epsilon^{(0)}_2(p , q , s)&= & \hbar \omega_1 \left[p-{m\over n}q+2{p\over q}s\right]+ \hbar\left(k_1-k_2 + 2k_3\right),\end{aligned}\ ] ] and the eigenstates are , @xmath207 going to next order in perturbation theory , we again have to require ( as in the example we discussed in the previous section ) that the first order correction to the state satisfy , @xmath208 .
we now compute the first order correction to the eigenvalues , @xmath209 or , explicitly , @xmath210 as before , we choose the @xmath183 s in such a way that @xmath184 is a solution , @xmath211 we now require that the states have vanishing eigenvalues , @xmath212 for these equations to have a solution with a common value of @xmath16 , we need to choose carefully the states @xmath213 , i.e , we must have @xmath214 , and the solution for @xmath16 is ,
@xmath215 so we see we completely reproduce the expansion of the exact solution we found in the previous subsection . as in the previous example , there are many eigenstates one can start from in the zeroth order theory that yield a solution for the same @xmath16 .
the formula ( [ forla ] ) is invariant if one multiplies @xmath216 times an arbitrary integer .
we therefore see that indeed one needs extra conditions in order to have both constraints vanish on a given value of the perturbative parameter .
but the conditions just imply that one starts with different zeroth order eigenvalues for the states in both constraints .
the situation is schematized in the figure .
finally , if one sought to compute the first order correction of the quantum state , one get two equations for it , one per each constraint .
if one studies them in detail , one finds that they admit the same solution .
we do not have a detailed discussion of the full general relativity quantum theory ready , but nevertheless one can sketch how the procedure would be .
the calculations in this section should only be taken as a guide , and will be highly formal in nature .
the details are very important in determining the final quantum theory , therefore we can only refer to these calculations as a `` sketch '' of what can be achieved .
we will see simultaneous solutions to the diffeomorphism and hamiltonian constrain . in order to achieve this
we start with a basis consistent of equivalence classes of spin nets under subgroup of diffeomorphisms that keeps the vertices of the spin net fixed , @xmath217 } > = \sum d_{[\vec{v } ] } | s>,\ ] ] where the @xmath218}$ ] are the diffeomorphisms that leave the vertices @xmath219 of the spin network fixed .
this is a space very similar to the `` habitat '' considered by lewandowski and marolf @xcite .
the action of both the diffeomorphism constraint and thiemann s hamiltonian is well defined in such a space .
we can therefore proceed to study the problem of eigenvalues and eigenvectors to first order in perturbation theory , @xmath220 considering the equation at zeroth order we get , @xmath221 the solution of this equation is given by the eigenvalues and eigenvectors of the determinant of the metric . here
we have to deal with the fact that this is a field theory .
one can smear the determinant of the metric with a function @xmath222 .
the resulting operator is closely related to the well - understood @xcite volume operator , we call it `` smeared volume '' and denote it as @xmath223 .
one will essentially get the volume associated with each vertex , which is a finite quantity , multiplied times the value of the smearing function at the vertex .
eigenstates will be reasonably easy to find .
they will correspond to combinations of spin networks that yield the same total for the sum of the volumes of each vertex times the smearing function valued at the vertex .
that quantity , in turn will be the eigenvalue @xmath224 .
we can schematically write these states as @xmath225 where we have a sum of states such that the all have the same value of @xmath223 for a given @xmath226 .
there are many states for each value of @xmath227 , to illustrate this , consider a spin network with a given number of vertices and a given value of @xmath227 and replace two lines in it by the same two lines with some knotting in between them .
the value of the eigenvalue will not change , but it will be a different spin network .
let us now consider the first order corrections . as we did in the quantum mechanical examples , we will project the perturbative equation on the space of states @xmath228 , @xmath229 the first and last term in both members cancels out , since @xmath230 .
one is therefore left with , @xmath231 one can use this equation to explicitly compute @xmath232 , for a given proposal for the hamiltonian constraint .
one can now proceed to make the `` energy '' vanish at this order of perturbation theory , therefore determining the ( quantized ) value of the cosmological constant and the possible values of the energy of the zeroth order state @xmath233 . to obtain the correction to the state , we need to project the above equation on states that are orthogonal to the eigenstates with eigenvalue @xmath227 .
let us call those states @xmath234 , that is , states with different values of the volume at the vertices .
we therefore assign the volume at the vertices as a measure of orthogonality .
we therefore write , @xmath235 now , since @xmath236 , we can write , @xmath237 and this equation would determine the first order correction to the quantum state
. this would be only strictly true if the state of volume @xmath223 were non - degenerate .
the degeneracy implies that there might be non - vanishing projections @xmath238 for vectors that belong to the space of states with eigenvalue @xmath227 .
it is well known in degenerate perturbation theory that higher order calculations determine these components .
the dependence on @xmath226 of the last equation might appear as surprising , since the first order quantum state should be @xmath226 independent .
here we can draw on our experience on the system with two constraints studied in last section , which shows that a single correction appears no matter which hamiltonian one uses to compute the correction . in the gravity case this would correspond to getting the same solution for different @xmath226 s . another way to see that the equation is @xmath226-independent is to notice that one can characterize the action of the hamiltonian independently at each vertex and in such calculations the role of @xmath226 is that of a constant overall factor that drops out of the equations and one is left , as in the case of the oscillators , with a system of equations .
we have therefore constructed states such that , @xmath239 such that , @xmath240 one is however , interested in states that are genuinely invariant under diffeomorphisms , whereas the above states are only invariant under diffeomorphisms that leave the vertices of the spin network fixed .
the kind of states we look for are , @xmath241 such that @xmath242 we shall show that with the above defined states this is indeed the case .
this follows from the poisson algebra of diffeomorphism and hamiltonian constraints , @xmath243 therefore , @xmath244 which is what we wished to show .
the details of these calculations can only be filled in with a given prescription for a hamiltonian constraint and a definite space of states to operate upon .
this will require much more detailed work than is appropriate for this paper .
one can hint at what would happen , for instance , if one uses thiemann s hamiltonian .
let us start with trivalent spin networks .
in such a context @xmath223 vanishes identically and the perturbative method has nothing to add : one is left without higher order corrections and the solution to the problem is reduced to determining which states are annihilated by the hamiltonian constraints .
to get something non - trivial , one needs four - valent vertices .
calculations on four valent vertices are readily possible , but cumbersome . in this context
one could sharpen the determination of eigenstates , etc .
very schematically , one would see that the correction to the states that one gets depend on the action of the hamiltonian constraint at each vertex .
the states appear different from the ones obtained via `` group averaging '' that involve acting repeatedly with the hamiltonian .
however , since the action of the hamiltonian is `` local '' ( i.e. it does not add connections among vertices , but `` dresses up '' each vertex ) , the kind of states one gets have similar locality properties to those obtained with the `` group averaging '' procedure . in fact , if one works at higher order , at each order one acts with a hamiltonian , so one can see that by going to higher orders , one is recovering something similar to the `` group averaging '' .
an interesting point in trying to draw an analogy between quantum gravity and the simple models discussed in this paper is if the perturbation hamiltonian has a discrete spectrum or not .
we know that the zeroth order hamiltonian ( the smeared volume ) has a discrete spectrum . at the moment
, it is knot known if either the thiemann or the vassiliev hamiltonians have a discrete spectrum .
however , we have shown in this paper that the perturbative method works well in the case in which the perturbative hamiltonian has a continuous spectrum ( the case of the coupled oscillators ) .
it is well known that the perturbative approach works well when the perturbing hamiltonian has a discrete spectrum ( for instance if one considers an atom in a magnetic field with constant energy ) .
it is clear that this is just the beginning of a discussion of the four dimensional quantum gravity case .
it is interesting to notice , however , how one can quickly gain intuition as to the kind of states one would get and how they depend on the action of the hamiltonian without doing the explicit computations . since the latter must involve four valent vertices and
therefore are extensive in nature , having a quick way to intuitively try out proposals for the hamiltonian is a great asset .
it is obvious that the complexities that one expects in quantum gravity ( in particular the infinite dimensional nature of the problem ) can not be really mimicked by finite dimensional systems .
we believe , however , that the finite dimensional models we analyze in this paper help dissipate some of the most elementary skepticisms about the possibilities of the approach we are proposing .
namely : 1 ) that the method seems to simplistic to deal with chaotic , possibly pathological hamiltonian systems as one expects full general relativity to contain in certain regimes ; 2 ) that the method is only circumscribed to a range of values of the cosmological constant of no physical relevance ; 3 ) that the application in the quantum domain of the technique is possibly inconsistent .
applied to the full quantum theory , our method appears to reduce the problem of finding quantum states to a set of well defined , albeit complicated spin network calculations .
moreover , it appears to yield a quick intuitive handle on the properties of the quantum states of possible proposals for hamiltonian constraints .
we are currently exploring further the application of the approximation technique in full quantum gravity , as sketched in the last section .
this paper was inspired by karel kucha s suggestions that we test our proposal in several model examples , some of which he pointed out ; he also made valuable criticisms to the manuscript .
we are grateful to abhay ashtekar and charles torre for discussions .
this work was supported in part by the national science foundation under grant int-9811610 , and research funds of the pennsylvania state university .
we acknowledge support of pedeciba .
we wish to thank thomas thiemann and the albert einstein institute of the max planck society for hospitality during the completion of this work .
t. thiemann , class .
grav . * 15 * , 839 ; * 15 * , 875 ; * 15 * , 1207 ; * 15 * , 1249 ; * 15 * , 1281 ; * 15 * , 1463 ( 1998 ) . c. di bartolo , r. gambini , j. griego , j. pullin , phys .
lett . * 84 * , 2314 ( 2000 ) ; gr - qc/9911009 , gr - qc/9911010 , class .
( to appear ) .
a. ashtekar and j. lewandowski , j. math .
phys . * 36 * , 2170 ( 1995 ) ; j. geom .
* 17 * , 191 ( 1995 ) ; a. ashtekar , j. lewandowski , d. marolf , j. mourao and t. thiemann , j. math .
phys . * 36 * , 6456 ( 1995 ) . c. rovelli and l. smolin , nucl .
* b442 * , 593 ( 1995 ) . c. rovelli , in http://www.livingreviews.org t. thiemann , hep - th/0005233 ; t. thiemann , o. winkler , hep - th/005237 ; hep - th/005234 ; hep - th/005235 ( 2000 ) .
the subject has a long history .
see for instance k. kucha , gr - qc/9304012 , in `` general relativity and gravitation 1992 : proceedings of the thirteenth international conference on general relativity and gravitation held at cordoba , argentina , 28 june-4 july 1992 '' , r.j .
gleiser , c.n .
kozameh , o.m .
moreschi , editors , institute of physics publishing , bristol ( 1993 ) , and references therein .
j. n. goldberg , j. lewandowski and c. stornaiolo , commun .
phys . * 148 * , 377 ( 1992 ) . c. g. torre , phys .
* d48 * , 2373 ( 1993 ) ; class .
* 8 * , 1895 ( 1991 ) ; c. g. torre and i. m. anderson , phys .
lett . * 70 * , 3525 ( 1993 ) .
p. g. bergmann , e. t. newman , rev .
* 29 * , 443 ( 1957 ) .
r. gambini , j. pullin , making classical and quantum canonical general relativity computable through a power series in the inverse cosmological constant `` , gr - qc/0008031 ( 2000 ) .
rescaling the constraint implies rescaling the lapse . in the @xmath0 limit this means a vanishing lapse .
our theory is therefore similar to the husain kucha theory , but with a cosmological constant . in their case
the hamiltonian constraint disappears , in ours the surviving term at zeroth order is given by the smeared volume @xmath223 ; v. husain , k. kucha , phys .
rev . * d42 * , 4070 ( 1990 ) . c. rovelli , nucl .
* b405 * , 797 ( 1993 ) ; v. husain , phys .
rev . * d47 * , 5394 ( 1993 ) .
l. smolin , phys .
rev . * d49 * , 4028 ( 1994 ) .
p. hajek , class .
* 7 * , 861 ( 1990 ) ; * 7 * , 871 ( 1990 ) ; a. ashtekar and r. s. tate , j. math .
phys . * 35 * , 6434 ( 1994 ) a. ashtekar , j. pullin , ann .
israel phys . soc . 9 , 65 - 77 ( 1990 ) ' ' developments in general relativity , astrophysics and quantum theory : a jubilee volume in honour of nathan rosen " f. cooperstock , l. p. horwitz and j. rosen . s. hervik , `` the bianchi type i minisuperspace model , '' gr - qc/0003084
. i. goldstein , `` classical mechanics '' , addison wesley , reading , ma ( 1980 ) .
j. lewandowski , d. marolf , int .
d7 * , 299 ( 1998 ) .
a. ashtekar and j. lewandowski , adv .
* 1 * , 388 ( 1998 ) ; t. thiemann , j. math .
* 39 * , 3347 ( 1998 ) . | we have recently introduced an approach for studying perturbatively classical and quantum canonical general relativity . the perturbative technique appears to preserve many of the attractive features of the non - perturbative quantization approach based on ashtekar s new variables and spin networks . with this approach
one can find perturbatively classical observables ( quantities that have vanishing poisson brackets with the constraints ) and quantum states ( states that are annihilated by the quantum constraints ) .
the relative ease with which the technique appears to deal with these traditionally hard problems opens several questions about how relevant the results produced can possibly be . among the questions
is the issue of how useful are results for large values of the cosmological constant and how the approach can deal with several pathologies that are expected to be present in the canonical approach to quantum gravity . with the aim of clarifying these points , and to make our construction as explicit as possible , we study its application in several simple models .
we consider bianchi cosmologies , the asymmetric top , the coupled harmonic oscillators with constant energy density and a simple quantum mechanical system with two hamiltonian constraints .
we find that the technique satisfactorily deals with the pathologies of these models and offers promise for finding ( at least some ) results even for small values of the cosmological constant .
finally , we briefly sketch how the method would operate in the full four dimensional quantum general relativity case . |
the total stellar mass is a fundamental parameter for any galaxy .
not only do many other galaxy properties vary with stellar mass , but an accurate measure of stellar mass is required to probe the dark matter content ( i.e. the total mass minus the stellar mass ) in a galaxy . however , measuring the total stellar mass is problematic , even once the total luminosity has been accurately measured .
for example , a common approach is to measure the total luminosity of a galaxy at near - ir wavelengths for which the light mostly comes from old stars that dominate the mass , and the effects of dust are much reduced compared to optical wavelengths .
a typical approach is to use the full - sky ground - based near - ir imaging of the 2mass survey ( jarrett et al .
however it has been reported that the 2mass reduction pipeline systematically underestimates the total luminosity and size of large , nearby galaxies due to a truncation of their surface brightness profiles ( schombert & smith 2012 ; scott et al .
2013 ) .
an alternative approach is to use the 3.6@xmath0 m band of the spitzer space telescope ( werner et al .
2004 ) or the 3.4@xmath0 m band of the wise space telescope ( wright et al .
such wavelengths are particularly well suited to measure the stellar masses of galaxies . for example , norris et al .
( 2014 ) concluded that photometry from wise can _ ... provide extremely simple , yet robust stellar mass tracers for dust free older stellar populations ...
"_. this is because the 3.43.6@xmath0 m light from galaxies is dominated by the light from old stars , and it is less effected by variations in the star formation history than shorter wavelengths .
although intermediate - aged stars , hot dust and polycyclic aromatic hydrocarbons may contribute to the emission at 3.6@xmath0 m , these sources are negligible for most early - type galaxies which are dominated by old stellar populations ( meidt et al .
2012 ; querejeta et al .
2015 ) . here
we use 3.6@xmath0 m imaging from the spitzer space telescope .
the 3.6@xmath0 m mass - to - light ratio ( m / l@xmath4 ) has virtually no dependence on metallicity , and only a very small dependence on age for old stellar ages .
we use the latest single - burst stellar population models ( rck et al .
2015 ) which are based on empirical mid - infrared stellar spectra ( cushing et al . 2005 ; rayner et al . 2009 ) .
these models cover a range of metallicity , ages and imf slopes .
they are shown to reproduce well the mid - infrared colours of early - type galaxies . for a kroupa imf ,
these models give m / l@xmath4 @xmath5 0.8 for a stellar population mean age of 9 gyr , with a variation between different isochrones , i.e. from basti and padova , of @xmath50.05 . for a metallicity range of [ fe / h ] = 0.4 to solar ( i.e. typical of the mean values for massive early - type galaxies )
, the variation is insignificant at @xmath50.02 .
we note that the flexible stellar population synthesis ( fsps ; conroy & gunn 2010 ) models with agb circumstellar dust included ( villaume et al . 2015 ) also give m / l@xmath4 @xmath50.8 for a 9 gyr old , moderately metal - rich population .
meidt et al .
( 2014 ) adopted a constant value of m / l@xmath4 = 0.6 for their s4 g sample , although their sample was dominated by late - type galaxies with younger mean ages on average .
stellar mass - to - light ratios have a strong dependence on the initial mass function ( imf ) .
the m / l@xmath4 values quoted above refer to a kropua imf .
salpeter and other imfs tend to have higher m / l@xmath4 values by a factor of @xmath51.5 - 3 ( rck et al .
2015 ) , which would lead to larger stellar masses for a given 3.6@xmath0 m luminosity .
recent work indicates that the imf for early - type galaxies is skewed to low mass stars ( see e.g. ferre - mateu et al .
2013 ; martin - navarro et al .
2015 ; mcconnell et al .
currently , it is not yet clear what is causing the imf variations nor whether these variations are confined to galaxy central regions , high metallicity regions or spheroids .
here we adopt a kroupa imf for our global m / l@xmath4 but caution that the stellar masses for massive elliptical galaxies may need revising upwards .
since m / l@xmath4 varies with stellar age , an age - appropriate ratio should be employed .
here we adopt an age dependent mass - to - light ratio from the rck et al .
( 2015 ) models using mean stellar ages from the literature .
we assume a kroupa imf .
the sluggs survey targets 25 nearby massive early - type galaxies in different environments and 3 so - called bonus galaxies ( brodie et al .
we study the kinematics and metallicity of both the galaxy itself , and its system of globular clusters , to large galactocentric radii .
the sample galaxies are chosen to cover a range of key parameters including stellar mass and physical size .
until now , the approach in the sluggs survey to measure stellar mass has been to obtain the extinction - corrected k - band ( 2.2@xmath0 m ) magnitude from the 2mass extended galaxy catalog and apply the correction of scott et al .
( 2013 ) for missing light .
we then applied a constant m / l@xmath6 = 1 irrespective of stellar metallicity or age .
a value of unity is simplistic , but a reasonable approximation for a very old stellar population with a kroupa imf ( bruzual & charlot 2003 ) .
the effective radii ( r@xmath7 ) of the sluggs galaxies are listed in brodie et al .
( 2014 , b14 ) , which are based on cappellari et al .
cappellari et al . used sizes from both optical and near - ir imaging .
they noted that the near - ir sizes from the 2mass survey ( jarrett et al .
2003 ) for the largest , most massive galaxies appear to be systematically underestimated and they scaled - up their near - ir sizes to match the optical sizes on average .
a recent study by van den bosch ( 2016 ) also found the 2mass survey to underestimate the sizes ( and total fluxes ) of nearby galaxies .
accurate galaxy effective radii are important in order to compare galaxies on a similar relative scale .
for example the sluggs survey and other integral field spectroscopy studies , derive kinematic profiles as a function of effective radii and measure properties such as specific angular momentum within 1 r@xmath7 ( arnold et al .
2014 ; alabi et al .
2015 ; foster et al .
thus the photometric effective radii need to be accurate in order to correctly compare kinematic properties between different galaxies .
effective radii , combined with accurate stellar masses , are needed to probe the dark matter fraction within a given multiple of r@xmath7 .
the spitzer space telescope imaging presented here offers an opportunity to revisit the sizes and masses of the sluggs early - type galaxies . in the next sections we present the 3.6@xmath0 m data from the spitzer space telescope and our methodology for deriving total magnitudes , stellar masses , and effective radii for 27 sluggs early - type galaxies .
these new measurements are compared with literature values . in an appendix
we list the measurements for six additional nearby early - type galaxies , which we include for the interested reader .
here we use images from the irac instrument of the spitzer space telescope , which has a pixel scale of 1.22 arcsec over a 5.2 @xmath8 5.2 arcmin@xmath9 field - of - view .
we have downloaded the latest ( july 2016 ) available 3.6@xmath0 m basic calibrated data frames from the spitzer heritage archive .
these astronomical observation requests ( aors ) are detailed in appendix a ( for sluggs galaxies ) and c ( for non - sluggs galaxies ) .
these data have been corrected for scattered light , dark current , flat - fielded and flux calibrated .
the mopex package is used to assemble the long - exposure ( @xmath101 s ) frames into an image mosaic showing the field of view around each target galaxy .
an example of the final mosaic for ngc 1407 is shown is fig . 1 .
we follow a similar reduction procedure to that of savorgnan & graham ( 2016 , sg16 ) .
thus when multiple pointings are available , an overlap correction is applied to create a uniform background level . using iraf
we determine the sky background level and rms at multiple points on the outskirts of each mosaiced image .
the sky values are averaged to give a final value for the sky background of each image , which is subtracted from the image .
finally , bright stars and other unwanted objects are masked out of the mosaic .
for further details see sg16
. unfortunately , the 3.6@xmath0 m data for the low - mass sluggs galaxy ngc 4474 are not useful for measuring the total light ( and hence mass ) or galaxy size . in this case , the galaxy is only partially visible as it is near the edge of the available spitzer pointing .
we therefore adopt its stellar mass ( log m@xmath2 = 10.23 ) from its 2mass k - band magnitude and its effective radius ( r@xmath7 = 1.5 kpc ) from brodie et al .
to measure the total light from each galaxy we model the galaxy in 2d using the iraf task _ ellipse _ and obtain the total magnitude of the galaxy model .
the galaxy centre was initially allowed to vary , but if it varied by more than 1 pixel , we fixed it to the average central value .
for a few galaxies ( i.e. ngc 4486 , 4594 , 5866 and 7457 ) we could not obtain a good galaxy model with a radially varying position angle ( pa ) and so in these cases we fixed the pa to a representative value based on the radial trend .
the model extends in galactocentric radius until the integrated magnitude at that radius is less than 0.02 mag different from the previous ( penultimate ) radius .
thus we effectively adopt the asymptotic total magnitude of the model galaxy .
we estimate our combined photometric and systematic uncertainty to be @xmath11 0.05 mag .
we do not correct our 3.6@xmath0 m magnitudes for galactic extinction ( which are less than 0.01 mag . ) .
the total 3.6@xmath0 m magnitudes that we measure in the vega system , and other basic properties of the sluggs galaxies , are given in table 1 .
@cccccccccc galaxy & type & core & dist . & age & m@xmath4 & m@xmath2 & r@xmath7 & @xmath12 & n + @xmath13 $ ] & & & [ mpc ] & [ gyr ] & [ mag ] & [ m@xmath3 ] & [ arcsec ] & [ mag / arcsec@xmath15 & + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) + + 1052 & e3 - 4/s0 & 1 & 19.4 & 13.0 & 7.12 & 11.02 & 21.9 & 17.16 & 3.4 + 2549 & s0 & 3 & 12.3 & 8.9 & 7.75 & 10.28 & 14.7 & 16.88 & 3.1 + 3379 & e0 - 1 & 1 & 10.3 & 13.7 & 5.92 & 10.96 & 54.9 & 18.02 & 5.7 + 3665 & s0 & & 33.1 & 13.2 & 7.12 & 11.48 & 50.5 & 18.95 & 5.4 + 3998 & s0 & 2 & 13.7 & 13.7 & 7.04 & 10.76 & 19.1 & 16.92 & 4.0 + 4551 & e & 3 & 16.1 & 13.2 & 8.67 & 10.24 & 13.8 & 17.45 & 2.1 + notes : columns are ( 1 ) galaxy name , ( 2 ) hubble type , ( 3 ) 1= core , 2 = intermediate , 3 = cusp central light profile , ( 4 ) distance , ( 5 ) mean stellar age from mcdermid et al .
( 2015 ) , except for ngc 1052 from milone et al .
( 2007 ) , ( 6 ) 3.6 micron apparent mag . , ( 7 ) stellar mass , ( 8) effective radius , ( 9 ) @xmath12 , ( 10 ) sersic n. sersic fit to 3.6@xmath0 m surface brightness profile and residuals as a function of circular equivalent radius for non - sluggs galaxies .
upper _ panel shows the data points ( with excluded data points shown by open circles ) and the best fit sersic profile in red .
parameters for the sersic fit are given in the top right .
the dashed line shows 3@xmath8 the rms of the sky background level .
@xmath16 gives the rms of the residuals in mag [email protected] _ lower _ panel shows the residuals of the sersic model fit minus the surface brightness data .
, scaledwidth=40.0% ] sersic fit to 3.6@xmath0 m surface brightness profile and residuals as a function of circular equivalent radius for non - sluggs galaxies .
upper _ panel shows the data points ( with excluded data points shown by open circles ) and the best fit sersic profile in red .
parameters for the sersic fit are given in the top right .
the dashed line shows 3@xmath8 the rms of the sky background level .
@xmath16 gives the rms of the residuals in mag [email protected] _ lower _ panel shows the residuals of the sersic model fit minus the surface brightness data .
, scaledwidth=40.0% ] | galaxy starlight at 3.6@xmath0 m is an excellent tracer of stellar mass .
here we use the latest 3.6@xmath0 m imaging from the spitzer space telescope to measure the total stellar mass and effective radii in a homogeneous way for a sample of galaxies from the sluggs survey .
these galaxies are representative of nearby early - type galaxies in the stellar mass range of 10 @xmath1 log m@xmath2/m@xmath3 @xmath1 11.7 , and our methodology can be applied to other samples of early - type galaxies .
we model each galaxy in 2d and estimate its total asymptotic magnitude from a 1d curve - of - growth .
magnitudes are converted into stellar masses using a 3.6@xmath0 m mass - to - light ratio from the latest stellar population models of rck et al . , assuming a kroupa imf .
we apply a ratio based on each galaxy s mean mass - weighted stellar age within one effective radius ( the mass - to - light ratio is insensitive to galaxy metallicity for the generally old stellar ages and high metallicities found in massive early - type galaxies ) .
our 3.6@xmath0 m stellar masses agree well with masses derived from 2.2@xmath0 m data . from the 1d surface brightness profile
we fit a single sersic law , excluding the very central regions .
we measure the effective radius , sersic n parameter and effective surface brightness for each galaxy .
we find that galaxy sizes derived from shallow optical imaging and the 2mass survey tend to underestimate the true size of the largest , most massive galaxies in our sample .
we adopt the 3.6@xmath0 m stellar masses and effective radii for the sluggs survey galaxies . [ firstpage ]
galaxies : masses galaxies : evolution galaxies : individual |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.